Compare commits
1 commit
master
...
gg/cublas-
Author | SHA1 | Date | |
---|---|---|---|
|
a40f6110f0 |
1219 changed files with 105124 additions and 330201 deletions
161
.clang-format
161
.clang-format
|
@ -1,161 +0,0 @@
|
|||
---
|
||||
Language: Cpp
|
||||
AlignAfterOpenBracket: Align
|
||||
AlignArrayOfStructures: Left
|
||||
AlignConsecutiveAssignments: AcrossComments
|
||||
AlignConsecutiveBitFields: AcrossComments
|
||||
AlignConsecutiveDeclarations: AcrossComments
|
||||
AlignConsecutiveMacros: AcrossComments
|
||||
# AlignConsecutiveShortCaseStatements: AcrossComments
|
||||
AlignEscapedNewlines: Left # LeftWithLastLine
|
||||
AlignOperands: Align
|
||||
AlignTrailingComments:
|
||||
Kind: Always
|
||||
OverEmptyLines: 1
|
||||
AllowAllArgumentsOnNextLine: true
|
||||
AllowAllParametersOfDeclarationOnNextLine: false
|
||||
# AllowBreakBeforeNoexceptSpecifier: OnlyWithParen
|
||||
AllowShortBlocksOnASingleLine: Never
|
||||
AllowShortCaseLabelsOnASingleLine: false
|
||||
AllowShortFunctionsOnASingleLine: Inline
|
||||
AllowShortIfStatementsOnASingleLine: Never
|
||||
AllowShortLambdasOnASingleLine: Inline
|
||||
AllowShortLoopsOnASingleLine: false
|
||||
AlwaysBreakBeforeMultilineStrings: true
|
||||
BinPackArguments: true
|
||||
BinPackParameters: true # OnePerLine
|
||||
BitFieldColonSpacing: Both
|
||||
BreakBeforeBraces: Custom # Attach
|
||||
BraceWrapping:
|
||||
AfterCaseLabel: true
|
||||
AfterClass: false
|
||||
AfterControlStatement: false
|
||||
AfterEnum: false
|
||||
AfterFunction: false
|
||||
AfterNamespace: false
|
||||
AfterObjCDeclaration: false
|
||||
AfterStruct: false
|
||||
AfterUnion: false
|
||||
AfterExternBlock: false
|
||||
BeforeCatch: false
|
||||
BeforeElse: false
|
||||
BeforeLambdaBody: false
|
||||
BeforeWhile: false
|
||||
IndentBraces: false
|
||||
SplitEmptyFunction: false
|
||||
SplitEmptyRecord: false
|
||||
SplitEmptyNamespace: false
|
||||
# BreakAdjacentStringLiterals: true
|
||||
BreakAfterAttributes: Never
|
||||
BreakBeforeBinaryOperators: None
|
||||
BreakBeforeInlineASMColon: OnlyMultiline
|
||||
BreakBeforeTernaryOperators: false
|
||||
# BreakBinaryOperations: Never
|
||||
BreakConstructorInitializers: AfterColon
|
||||
# BreakFunctionDefinitionParameters: false
|
||||
BreakInheritanceList: AfterComma
|
||||
BreakStringLiterals: true
|
||||
# BreakTemplateDeclarations: Yes
|
||||
ColumnLimit: 120
|
||||
CommentPragmas: '^ IWYU pragma:'
|
||||
CompactNamespaces: false
|
||||
ConstructorInitializerIndentWidth: 4
|
||||
ContinuationIndentWidth: 4
|
||||
Cpp11BracedListStyle: false
|
||||
DerivePointerAlignment: false
|
||||
DisableFormat: false
|
||||
EmptyLineBeforeAccessModifier: Leave
|
||||
EmptyLineAfterAccessModifier: Never
|
||||
ExperimentalAutoDetectBinPacking: false
|
||||
FixNamespaceComments: true
|
||||
IncludeBlocks: Regroup
|
||||
IncludeCategories:
|
||||
- Regex: '^<.*\.h>'
|
||||
Priority: 1
|
||||
SortPriority: 0
|
||||
- Regex: '^<.*'
|
||||
Priority: 2
|
||||
SortPriority: 0
|
||||
- Regex: '.*'
|
||||
Priority: 3
|
||||
SortPriority: 0
|
||||
IncludeIsMainRegex: '([-_](test|unittest))?$'
|
||||
IncludeIsMainSourceRegex: ''
|
||||
IndentAccessModifiers: false
|
||||
IndentCaseBlocks: true
|
||||
IndentCaseLabels: true
|
||||
IndentExternBlock: NoIndent
|
||||
IndentGotoLabels: false
|
||||
IndentPPDirectives: AfterHash
|
||||
IndentWidth: 4
|
||||
IndentWrappedFunctionNames: false
|
||||
InsertBraces: true # NOTE: may lead to incorrect formatting
|
||||
InsertNewlineAtEOF: true
|
||||
JavaScriptQuotes: Leave
|
||||
JavaScriptWrapImports: true
|
||||
KeepEmptyLinesAtTheStartOfBlocks: false
|
||||
LambdaBodyIndentation: Signature
|
||||
LineEnding: LF
|
||||
MacroBlockBegin: ''
|
||||
MacroBlockEnd: ''
|
||||
MaxEmptyLinesToKeep: 1
|
||||
NamespaceIndentation: None
|
||||
ObjCBinPackProtocolList: Auto
|
||||
ObjCBlockIndentWidth: 4
|
||||
ObjCSpaceAfterProperty: true
|
||||
ObjCSpaceBeforeProtocolList: true
|
||||
PPIndentWidth: -1
|
||||
PackConstructorInitializers: CurrentLine
|
||||
PenaltyBreakAssignment: 2
|
||||
PenaltyBreakBeforeFirstCallParameter: 1
|
||||
PenaltyBreakComment: 300
|
||||
PenaltyBreakFirstLessLess: 120
|
||||
PenaltyBreakString: 1000
|
||||
PenaltyBreakTemplateDeclaration: 10
|
||||
PenaltyExcessCharacter: 1000000
|
||||
PenaltyReturnTypeOnItsOwnLine: 200
|
||||
PointerAlignment: Middle
|
||||
QualifierAlignment: Left
|
||||
#QualifierOrder: ['static', 'inline', 'friend', 'constexpr', 'const', 'volatile', 'type', 'restrict']
|
||||
RawStringFormats:
|
||||
- Language: Cpp
|
||||
Delimiters:
|
||||
- cc
|
||||
- CC
|
||||
- cpp
|
||||
- Cpp
|
||||
- CPP
|
||||
- 'c++'
|
||||
- 'C++'
|
||||
CanonicalDelimiter: ''
|
||||
ReferenceAlignment: Middle
|
||||
ReflowComments: false # IndentOnly
|
||||
SeparateDefinitionBlocks: Always
|
||||
SortIncludes: CaseInsensitive
|
||||
SortUsingDeclarations: LexicographicNumeric
|
||||
SpaceAfterCStyleCast: true
|
||||
SpaceAfterLogicalNot: false
|
||||
SpaceAfterTemplateKeyword: true
|
||||
SpaceBeforeAssignmentOperators: true
|
||||
SpaceBeforeCpp11BracedList: false
|
||||
SpaceBeforeCtorInitializerColon: true
|
||||
SpaceBeforeInheritanceColon: true
|
||||
SpaceBeforeParens: ControlStatements
|
||||
SpaceBeforeRangeBasedForLoopColon: true
|
||||
SpaceInEmptyBlock: false
|
||||
SpaceInEmptyParentheses: false
|
||||
SpacesBeforeTrailingComments: 2
|
||||
SpacesInAngles: Never
|
||||
SpacesInContainerLiterals: true
|
||||
SpacesInLineCommentPrefix:
|
||||
Minimum: 1
|
||||
Maximum: -1
|
||||
SpacesInParentheses: false
|
||||
SpacesInSquareBrackets: false
|
||||
SpaceBeforeSquareBrackets: false
|
||||
Standard: c++17
|
||||
TabWidth: 4
|
||||
UseTab: Never
|
||||
WhitespaceSensitiveMacros: ['STRINGIZE']
|
||||
...
|
||||
|
|
@ -12,15 +12,12 @@ Checks: >
|
|||
-readability-implicit-bool-conversion,
|
||||
-readability-magic-numbers,
|
||||
-readability-uppercase-literal-suffix,
|
||||
-readability-simplify-boolean-expr,
|
||||
clang-analyzer-*,
|
||||
-clang-analyzer-security.insecureAPI.DeprecatedOrUnsafeBufferHandling,
|
||||
performance-*,
|
||||
portability-*,
|
||||
-portability-simd-intrinsics,
|
||||
misc-*,
|
||||
-misc-const-correctness,
|
||||
-misc-non-private-member-variables-in-classes,
|
||||
-misc-no-recursion,
|
||||
-misc-use-anonymous-namespace,
|
||||
FormatStyle: none
|
||||
|
|
|
@ -15,7 +15,7 @@ node('x86_runner1'){ // Running on x86 runner containing latest vecto
|
|||
stage('Running llama.cpp'){
|
||||
sh'''#!/bin/bash
|
||||
module load gnu-bin2/0.1 # loading latest versions of vector qemu and vector gcc
|
||||
qemu-riscv64 -L /softwares/gnu-bin2/sysroot -cpu rv64,v=true,vlen=256,elen=64,vext_spec=v1.0 ./llama-cli -m /home/alitariq/codellama-7b.Q4_K_M.gguf -p "Anything" -n 9 > llama_log.txt # Running llama.cpp on vector qemu-riscv64
|
||||
qemu-riscv64 -L /softwares/gnu-bin2/sysroot -cpu rv64,v=true,vlen=256,elen=64,vext_spec=v1.0 ./main -m /home/alitariq/codellama-7b.Q4_K_M.gguf -p "Anything" -n 9 > llama_log.txt # Running llama.cpp on vector qemu-riscv64
|
||||
cat llama_log.txt # Printing results
|
||||
'''
|
||||
}
|
||||
|
|
|
@ -1,92 +0,0 @@
|
|||
ARG UBUNTU_VERSION=22.04
|
||||
|
||||
FROM ubuntu:$UBUNTU_VERSION AS build
|
||||
|
||||
ARG TARGETARCH
|
||||
|
||||
ARG GGML_CPU_ARM_ARCH=armv8-a
|
||||
|
||||
RUN apt-get update && \
|
||||
apt-get install -y build-essential git cmake libcurl4-openssl-dev
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
COPY . .
|
||||
|
||||
RUN if [ "$TARGETARCH" = "amd64" ]; then \
|
||||
cmake -S . -B build -DCMAKE_BUILD_TYPE=Release -DLLAMA_CURL=ON -DGGML_NATIVE=OFF -DGGML_BACKEND_DL=ON -DGGML_CPU_ALL_VARIANTS=ON; \
|
||||
elif [ "$TARGETARCH" = "arm64" ]; then \
|
||||
cmake -S . -B build -DCMAKE_BUILD_TYPE=Release -DLLAMA_CURL=ON -DGGML_NATIVE=OFF -DGGML_CPU_ARM_ARCH=${GGML_CPU_ARM_ARCH}; \
|
||||
else \
|
||||
echo "Unsupported architecture"; \
|
||||
exit 1; \
|
||||
fi && \
|
||||
cmake --build build -j $(nproc)
|
||||
|
||||
RUN mkdir -p /app/lib && \
|
||||
find build -name "*.so" -exec cp {} /app/lib \;
|
||||
|
||||
RUN mkdir -p /app/full \
|
||||
&& cp build/bin/* /app/full \
|
||||
&& cp *.py /app/full \
|
||||
&& cp -r gguf-py /app/full \
|
||||
&& cp -r requirements /app/full \
|
||||
&& cp requirements.txt /app/full \
|
||||
&& cp .devops/tools.sh /app/full/tools.sh
|
||||
|
||||
## Base image
|
||||
FROM ubuntu:$UBUNTU_VERSION AS base
|
||||
|
||||
RUN apt-get update \
|
||||
&& apt-get install -y libgomp1 curl\
|
||||
&& apt autoremove -y \
|
||||
&& apt clean -y \
|
||||
&& rm -rf /tmp/* /var/tmp/* \
|
||||
&& find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
|
||||
&& find /var/cache -type f -delete
|
||||
|
||||
COPY --from=build /app/lib/ /app
|
||||
|
||||
### Full
|
||||
FROM base AS full
|
||||
|
||||
COPY --from=build /app/full /app
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
RUN apt-get update \
|
||||
&& apt-get install -y \
|
||||
git \
|
||||
python3 \
|
||||
python3-pip \
|
||||
&& pip install --upgrade pip setuptools wheel \
|
||||
&& pip install -r requirements.txt \
|
||||
&& apt autoremove -y \
|
||||
&& apt clean -y \
|
||||
&& rm -rf /tmp/* /var/tmp/* \
|
||||
&& find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
|
||||
&& find /var/cache -type f -delete
|
||||
|
||||
ENTRYPOINT ["/app/tools.sh"]
|
||||
|
||||
### Light, CLI only
|
||||
FROM base AS light
|
||||
|
||||
COPY --from=build /app/full/llama-cli /app
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
ENTRYPOINT [ "/app/llama-cli" ]
|
||||
|
||||
### Server, Server only
|
||||
FROM base AS server
|
||||
|
||||
ENV LLAMA_ARG_HOST=0.0.0.0
|
||||
|
||||
COPY --from=build /app/full/llama-server /app
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ]
|
||||
|
||||
ENTRYPOINT [ "/app/llama-server" ]
|
|
@ -1,94 +0,0 @@
|
|||
ARG UBUNTU_VERSION=22.04
|
||||
# This needs to generally match the container host's environment.
|
||||
ARG CUDA_VERSION=12.6.0
|
||||
# Target the CUDA build image
|
||||
ARG BASE_CUDA_DEV_CONTAINER=nvidia/cuda:${CUDA_VERSION}-devel-ubuntu${UBUNTU_VERSION}
|
||||
|
||||
ARG BASE_CUDA_RUN_CONTAINER=nvidia/cuda:${CUDA_VERSION}-runtime-ubuntu${UBUNTU_VERSION}
|
||||
|
||||
FROM ${BASE_CUDA_DEV_CONTAINER} AS build
|
||||
|
||||
# CUDA architecture to build for (defaults to all supported archs)
|
||||
ARG CUDA_DOCKER_ARCH=default
|
||||
|
||||
RUN apt-get update && \
|
||||
apt-get install -y build-essential cmake python3 python3-pip git libcurl4-openssl-dev libgomp1
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
COPY . .
|
||||
|
||||
RUN if [ "${CUDA_DOCKER_ARCH}" != "default" ]; then \
|
||||
export CMAKE_ARGS="-DCMAKE_CUDA_ARCHITECTURES=${CUDA_DOCKER_ARCH}"; \
|
||||
fi && \
|
||||
cmake -B build -DGGML_NATIVE=OFF -DGGML_CUDA=ON -DLLAMA_CURL=ON ${CMAKE_ARGS} -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined . && \
|
||||
cmake --build build --config Release -j$(nproc)
|
||||
|
||||
RUN mkdir -p /app/lib && \
|
||||
find build -name "*.so" -exec cp {} /app/lib \;
|
||||
|
||||
RUN mkdir -p /app/full \
|
||||
&& cp build/bin/* /app/full \
|
||||
&& cp *.py /app/full \
|
||||
&& cp -r gguf-py /app/full \
|
||||
&& cp -r requirements /app/full \
|
||||
&& cp requirements.txt /app/full \
|
||||
&& cp .devops/tools.sh /app/full/tools.sh
|
||||
|
||||
## Base image
|
||||
FROM ${BASE_CUDA_RUN_CONTAINER} AS base
|
||||
|
||||
RUN apt-get update \
|
||||
&& apt-get install -y libgomp1 curl\
|
||||
&& apt autoremove -y \
|
||||
&& apt clean -y \
|
||||
&& rm -rf /tmp/* /var/tmp/* \
|
||||
&& find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
|
||||
&& find /var/cache -type f -delete
|
||||
|
||||
COPY --from=build /app/lib/ /app
|
||||
|
||||
### Full
|
||||
FROM base AS full
|
||||
|
||||
COPY --from=build /app/full /app
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
RUN apt-get update \
|
||||
&& apt-get install -y \
|
||||
git \
|
||||
python3 \
|
||||
python3-pip \
|
||||
&& pip install --upgrade pip setuptools wheel \
|
||||
&& pip install -r requirements.txt \
|
||||
&& apt autoremove -y \
|
||||
&& apt clean -y \
|
||||
&& rm -rf /tmp/* /var/tmp/* \
|
||||
&& find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
|
||||
&& find /var/cache -type f -delete
|
||||
|
||||
|
||||
ENTRYPOINT ["/app/tools.sh"]
|
||||
|
||||
### Light, CLI only
|
||||
FROM base AS light
|
||||
|
||||
COPY --from=build /app/full/llama-cli /app
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
ENTRYPOINT [ "/app/llama-cli" ]
|
||||
|
||||
### Server, Server only
|
||||
FROM base AS server
|
||||
|
||||
ENV LLAMA_ARG_HOST=0.0.0.0
|
||||
|
||||
COPY --from=build /app/full/llama-server /app
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ]
|
||||
|
||||
ENTRYPOINT [ "/app/llama-server" ]
|
33
.devops/full-cuda.Dockerfile
Normal file
33
.devops/full-cuda.Dockerfile
Normal file
|
@ -0,0 +1,33 @@
|
|||
ARG UBUNTU_VERSION=22.04
|
||||
|
||||
# This needs to generally match the container host's environment.
|
||||
ARG CUDA_VERSION=11.7.1
|
||||
|
||||
# Target the CUDA build image
|
||||
ARG BASE_CUDA_DEV_CONTAINER=nvidia/cuda:${CUDA_VERSION}-devel-ubuntu${UBUNTU_VERSION}
|
||||
|
||||
FROM ${BASE_CUDA_DEV_CONTAINER} as build
|
||||
|
||||
# Unless otherwise specified, we make a fat build.
|
||||
ARG CUDA_DOCKER_ARCH=all
|
||||
|
||||
RUN apt-get update && \
|
||||
apt-get install -y build-essential python3 python3-pip git
|
||||
|
||||
COPY requirements.txt requirements.txt
|
||||
|
||||
RUN pip install --upgrade pip setuptools wheel \
|
||||
&& pip install -r requirements.txt
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
COPY . .
|
||||
|
||||
# Set nvcc architecture
|
||||
ENV CUDA_DOCKER_ARCH=${CUDA_DOCKER_ARCH}
|
||||
# Enable cuBLAS
|
||||
ENV LLAMA_CUBLAS=1
|
||||
|
||||
RUN make
|
||||
|
||||
ENTRYPOINT ["/app/.devops/tools.sh"]
|
44
.devops/full-rocm.Dockerfile
Normal file
44
.devops/full-rocm.Dockerfile
Normal file
|
@ -0,0 +1,44 @@
|
|||
ARG UBUNTU_VERSION=22.04
|
||||
|
||||
# This needs to generally match the container host's environment.
|
||||
ARG ROCM_VERSION=5.6
|
||||
|
||||
# Target the CUDA build image
|
||||
ARG BASE_ROCM_DEV_CONTAINER=rocm/dev-ubuntu-${UBUNTU_VERSION}:${ROCM_VERSION}-complete
|
||||
|
||||
FROM ${BASE_ROCM_DEV_CONTAINER} as build
|
||||
|
||||
# Unless otherwise specified, we make a fat build.
|
||||
# List from https://github.com/ggerganov/llama.cpp/pull/1087#issuecomment-1682807878
|
||||
# This is mostly tied to rocBLAS supported archs.
|
||||
ARG ROCM_DOCKER_ARCH=\
|
||||
gfx803 \
|
||||
gfx900 \
|
||||
gfx906 \
|
||||
gfx908 \
|
||||
gfx90a \
|
||||
gfx1010 \
|
||||
gfx1030 \
|
||||
gfx1100 \
|
||||
gfx1101 \
|
||||
gfx1102
|
||||
|
||||
COPY requirements.txt requirements.txt
|
||||
|
||||
RUN pip install --upgrade pip setuptools wheel \
|
||||
&& pip install -r requirements.txt
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
COPY . .
|
||||
|
||||
# Set nvcc architecture
|
||||
ENV GPU_TARGETS=${ROCM_DOCKER_ARCH}
|
||||
# Enable ROCm
|
||||
ENV LLAMA_HIPBLAS=1
|
||||
ENV CC=/opt/rocm/llvm/bin/clang
|
||||
ENV CXX=/opt/rocm/llvm/bin/clang++
|
||||
|
||||
RUN make
|
||||
|
||||
ENTRYPOINT ["/app/.devops/tools.sh"]
|
21
.devops/full.Dockerfile
Normal file
21
.devops/full.Dockerfile
Normal file
|
@ -0,0 +1,21 @@
|
|||
ARG UBUNTU_VERSION=22.04
|
||||
|
||||
FROM ubuntu:$UBUNTU_VERSION as build
|
||||
|
||||
RUN apt-get update && \
|
||||
apt-get install -y build-essential python3 python3-pip git
|
||||
|
||||
COPY requirements.txt requirements.txt
|
||||
|
||||
RUN pip install --upgrade pip setuptools wheel \
|
||||
&& pip install -r requirements.txt
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
COPY . .
|
||||
|
||||
RUN make
|
||||
|
||||
ENV LC_ALL=C.utf8
|
||||
|
||||
ENTRYPOINT ["/app/.devops/tools.sh"]
|
|
@ -1,91 +0,0 @@
|
|||
ARG ONEAPI_VERSION=2025.0.0-0-devel-ubuntu22.04
|
||||
|
||||
## Build Image
|
||||
|
||||
FROM intel/oneapi-basekit:$ONEAPI_VERSION AS build
|
||||
|
||||
ARG GGML_SYCL_F16=OFF
|
||||
RUN apt-get update && \
|
||||
apt-get install -y git libcurl4-openssl-dev
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
COPY . .
|
||||
|
||||
RUN if [ "${GGML_SYCL_F16}" = "ON" ]; then \
|
||||
echo "GGML_SYCL_F16 is set" \
|
||||
&& export OPT_SYCL_F16="-DGGML_SYCL_F16=ON"; \
|
||||
fi && \
|
||||
echo "Building with dynamic libs" && \
|
||||
cmake -B build -DGGML_NATIVE=OFF -DGGML_SYCL=ON -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx -DLLAMA_CURL=ON ${OPT_SYCL_F16} && \
|
||||
cmake --build build --config Release -j$(nproc)
|
||||
|
||||
RUN mkdir -p /app/lib && \
|
||||
find build -name "*.so" -exec cp {} /app/lib \;
|
||||
|
||||
RUN mkdir -p /app/full \
|
||||
&& cp build/bin/* /app/full \
|
||||
&& cp *.py /app/full \
|
||||
&& cp -r gguf-py /app/full \
|
||||
&& cp -r requirements /app/full \
|
||||
&& cp requirements.txt /app/full \
|
||||
&& cp .devops/tools.sh /app/full/tools.sh
|
||||
|
||||
FROM intel/oneapi-basekit:$ONEAPI_VERSION AS base
|
||||
|
||||
RUN apt-get update \
|
||||
&& apt-get install -y libgomp1 curl\
|
||||
&& apt autoremove -y \
|
||||
&& apt clean -y \
|
||||
&& rm -rf /tmp/* /var/tmp/* \
|
||||
&& find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
|
||||
&& find /var/cache -type f -delete
|
||||
|
||||
### Full
|
||||
FROM base AS full
|
||||
|
||||
COPY --from=build /app/lib/ /app
|
||||
COPY --from=build /app/full /app
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
RUN apt-get update \
|
||||
&& apt-get install -y \
|
||||
git \
|
||||
python3 \
|
||||
python3-pip \
|
||||
&& pip install --upgrade pip setuptools wheel \
|
||||
&& pip install -r requirements.txt \
|
||||
&& apt autoremove -y \
|
||||
&& apt clean -y \
|
||||
&& rm -rf /tmp/* /var/tmp/* \
|
||||
&& find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
|
||||
&& find /var/cache -type f -delete
|
||||
|
||||
|
||||
ENTRYPOINT ["/app/tools.sh"]
|
||||
|
||||
### Light, CLI only
|
||||
FROM base AS light
|
||||
|
||||
COPY --from=build /app/lib/ /app
|
||||
COPY --from=build /app/full/llama-cli /app
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
ENTRYPOINT [ "/app/llama-cli" ]
|
||||
|
||||
### Server, Server only
|
||||
FROM base AS server
|
||||
|
||||
ENV LLAMA_ARG_HOST=0.0.0.0
|
||||
|
||||
COPY --from=build /app/lib/ /app
|
||||
COPY --from=build /app/full/llama-server /app
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ]
|
||||
|
||||
ENTRYPOINT [ "/app/llama-server" ]
|
||||
|
|
@ -1,44 +0,0 @@
|
|||
ARG ASCEND_VERSION=8.0.rc2.alpha003-910b-openeuler22.03-py3.8
|
||||
|
||||
FROM ascendai/cann:$ASCEND_VERSION AS build
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
COPY . .
|
||||
|
||||
RUN yum install -y gcc g++ cmake make
|
||||
ENV ASCEND_TOOLKIT_HOME=/usr/local/Ascend/ascend-toolkit/latest
|
||||
ENV LIBRARY_PATH=${ASCEND_TOOLKIT_HOME}/lib64:$LIBRARY_PATH
|
||||
ENV LD_LIBRARY_PATH=${ASCEND_TOOLKIT_HOME}/lib64:${ASCEND_TOOLKIT_HOME}/lib64/plugin/opskernel:${ASCEND_TOOLKIT_HOME}/lib64/plugin/nnengine:${ASCEND_TOOLKIT_HOME}/opp/built-in/op_impl/ai_core/tbe/op_tiling:${LD_LIBRARY_PATH}
|
||||
ENV PYTHONPATH=${ASCEND_TOOLKIT_HOME}/python/site-packages:${ASCEND_TOOLKIT_HOME}/opp/built-in/op_impl/ai_core/tbe:${PYTHONPATH}
|
||||
ENV PATH=${ASCEND_TOOLKIT_HOME}/bin:${ASCEND_TOOLKIT_HOME}/compiler/ccec_compiler/bin:${PATH}
|
||||
ENV ASCEND_AICPU_PATH=${ASCEND_TOOLKIT_HOME}
|
||||
ENV ASCEND_OPP_PATH=${ASCEND_TOOLKIT_HOME}/opp
|
||||
ENV TOOLCHAIN_HOME=${ASCEND_TOOLKIT_HOME}/toolkit
|
||||
ENV ASCEND_HOME_PATH=${ASCEND_TOOLKIT_HOME}
|
||||
|
||||
# find libascend_hal.so, because the drive hasn`t been mounted.
|
||||
ENV LD_LIBRARY_PATH=${ASCEND_TOOLKIT_HOME}/runtime/lib64/stub:$LD_LIBRARY_PATH
|
||||
|
||||
RUN echo "Building with static libs" && \
|
||||
source /usr/local/Ascend/ascend-toolkit/set_env.sh --force && \
|
||||
cmake -B build -DGGML_NATIVE=OFF -DGGML_CANN=ON -DBUILD_SHARED_LIBS=OFF && \
|
||||
cmake --build build --config Release --target llama-cli
|
||||
|
||||
# TODO: use image with NNRT
|
||||
FROM ascendai/cann:$ASCEND_VERSION AS runtime
|
||||
COPY --from=build /app/build/bin/llama-cli /llama-cli
|
||||
|
||||
ENV LC_ALL=C.utf8
|
||||
|
||||
ENV ASCEND_TOOLKIT_HOME=/usr/local/Ascend/ascend-toolkit/latest
|
||||
ENV LIBRARY_PATH=${ASCEND_TOOLKIT_HOME}/lib64:$LIBRARY_PATH
|
||||
ENV LD_LIBRARY_PATH=${ASCEND_TOOLKIT_HOME}/lib64:${ASCEND_TOOLKIT_HOME}/lib64/plugin/opskernel:${ASCEND_TOOLKIT_HOME}/lib64/plugin/nnengine:${ASCEND_TOOLKIT_HOME}/opp/built-in/op_impl/ai_core/tbe/op_tiling:${LD_LIBRARY_PATH}
|
||||
ENV PYTHONPATH=${ASCEND_TOOLKIT_HOME}/python/site-packages:${ASCEND_TOOLKIT_HOME}/opp/built-in/op_impl/ai_core/tbe:${PYTHONPATH}
|
||||
ENV PATH=${ASCEND_TOOLKIT_HOME}/bin:${ASCEND_TOOLKIT_HOME}/compiler/ccec_compiler/bin:${PATH}
|
||||
ENV ASCEND_AICPU_PATH=${ASCEND_TOOLKIT_HOME}
|
||||
ENV ASCEND_OPP_PATH=${ASCEND_TOOLKIT_HOME}/opp
|
||||
ENV TOOLCHAIN_HOME=${ASCEND_TOOLKIT_HOME}/toolkit
|
||||
ENV ASCEND_HOME_PATH=${ASCEND_TOOLKIT_HOME}
|
||||
|
||||
ENTRYPOINT ["/llama-cli" ]
|
84
.devops/llama-cpp-clblast.srpm.spec
Normal file
84
.devops/llama-cpp-clblast.srpm.spec
Normal file
|
@ -0,0 +1,84 @@
|
|||
# SRPM for building from source and packaging an RPM for RPM-based distros.
|
||||
# https://fedoraproject.org/wiki/How_to_create_an_RPM_package
|
||||
# Built and maintained by John Boero - boeroboy@gmail.com
|
||||
# In honor of Seth Vidal https://www.redhat.com/it/blog/thank-you-seth-vidal
|
||||
|
||||
# Notes for llama.cpp:
|
||||
# 1. Tags are currently based on hash - which will not sort asciibetically.
|
||||
# We need to declare standard versioning if people want to sort latest releases.
|
||||
# 2. Builds for CUDA/OpenCL support are separate, with different depenedencies.
|
||||
# 3. NVidia's developer repo must be enabled with nvcc, cublas, clblas, etc installed.
|
||||
# Example: https://developer.download.nvidia.com/compute/cuda/repos/fedora37/x86_64/cuda-fedora37.repo
|
||||
# 4. OpenCL/CLBLAST support simply requires the ICD loader and basic opencl libraries.
|
||||
# It is up to the user to install the correct vendor-specific support.
|
||||
|
||||
Name: llama.cpp-clblast
|
||||
Version: %( date "+%%Y%%m%%d" )
|
||||
Release: 1%{?dist}
|
||||
Summary: OpenCL Inference of LLaMA model in C/C++
|
||||
License: MIT
|
||||
Source0: https://github.com/ggerganov/llama.cpp/archive/refs/heads/master.tar.gz
|
||||
BuildRequires: coreutils make gcc-c++ git mesa-libOpenCL-devel clblast-devel
|
||||
Requires: clblast
|
||||
URL: https://github.com/ggerganov/llama.cpp
|
||||
|
||||
%define debug_package %{nil}
|
||||
%define source_date_epoch_from_changelog 0
|
||||
|
||||
%description
|
||||
CPU inference for Meta's Lllama2 models using default options.
|
||||
|
||||
%prep
|
||||
%setup -n llama.cpp-master
|
||||
|
||||
%build
|
||||
make -j LLAMA_CLBLAST=1
|
||||
|
||||
%install
|
||||
mkdir -p %{buildroot}%{_bindir}/
|
||||
cp -p main %{buildroot}%{_bindir}/llamaclblast
|
||||
cp -p server %{buildroot}%{_bindir}/llamaclblastserver
|
||||
cp -p simple %{buildroot}%{_bindir}/llamaclblastsimple
|
||||
|
||||
mkdir -p %{buildroot}/usr/lib/systemd/system
|
||||
%{__cat} <<EOF > %{buildroot}/usr/lib/systemd/system/llamaclblast.service
|
||||
[Unit]
|
||||
Description=Llama.cpp server, CPU only (no GPU support in this build).
|
||||
After=syslog.target network.target local-fs.target remote-fs.target nss-lookup.target
|
||||
|
||||
[Service]
|
||||
Type=simple
|
||||
EnvironmentFile=/etc/sysconfig/llama
|
||||
ExecStart=/usr/bin/llamaclblastserver $LLAMA_ARGS
|
||||
ExecReload=/bin/kill -s HUP $MAINPID
|
||||
Restart=never
|
||||
|
||||
[Install]
|
||||
WantedBy=default.target
|
||||
EOF
|
||||
|
||||
mkdir -p %{buildroot}/etc/sysconfig
|
||||
%{__cat} <<EOF > %{buildroot}/etc/sysconfig/llama
|
||||
LLAMA_ARGS="-m /opt/llama2/ggml-model-f32.bin"
|
||||
EOF
|
||||
|
||||
%clean
|
||||
rm -rf %{buildroot}
|
||||
rm -rf %{_builddir}/*
|
||||
|
||||
%files
|
||||
%{_bindir}/llamaclblast
|
||||
%{_bindir}/llamaclblastserver
|
||||
%{_bindir}/llamaclblastsimple
|
||||
/usr/lib/systemd/system/llamaclblast.service
|
||||
%config /etc/sysconfig/llama
|
||||
|
||||
|
||||
%pre
|
||||
|
||||
%post
|
||||
|
||||
%preun
|
||||
%postun
|
||||
|
||||
%changelog
|
|
@ -1,5 +1,5 @@
|
|||
# SRPM for building from source and packaging an RPM for RPM-based distros.
|
||||
# https://docs.fedoraproject.org/en-US/quick-docs/creating-rpm-packages
|
||||
# https://fedoraproject.org/wiki/How_to_create_an_RPM_package
|
||||
# Built and maintained by John Boero - boeroboy@gmail.com
|
||||
# In honor of Seth Vidal https://www.redhat.com/it/blog/thank-you-seth-vidal
|
||||
|
||||
|
@ -12,7 +12,7 @@
|
|||
# 4. OpenCL/CLBLAST support simply requires the ICD loader and basic opencl libraries.
|
||||
# It is up to the user to install the correct vendor-specific support.
|
||||
|
||||
Name: llama.cpp-cuda
|
||||
Name: llama.cpp-cublas
|
||||
Version: %( date "+%%Y%%m%%d" )
|
||||
Release: 1%{?dist}
|
||||
Summary: CPU Inference of LLaMA model in pure C/C++ (no CUDA/OpenCL)
|
||||
|
@ -32,16 +32,16 @@ CPU inference for Meta's Lllama2 models using default options.
|
|||
%setup -n llama.cpp-master
|
||||
|
||||
%build
|
||||
make -j GGML_CUDA=1
|
||||
make -j LLAMA_CUBLAS=1
|
||||
|
||||
%install
|
||||
mkdir -p %{buildroot}%{_bindir}/
|
||||
cp -p llama-cli %{buildroot}%{_bindir}/llama-cuda-cli
|
||||
cp -p llama-server %{buildroot}%{_bindir}/llama-cuda-server
|
||||
cp -p llama-simple %{buildroot}%{_bindir}/llama-cuda-simple
|
||||
cp -p main %{buildroot}%{_bindir}/llamacppcublas
|
||||
cp -p server %{buildroot}%{_bindir}/llamacppcublasserver
|
||||
cp -p simple %{buildroot}%{_bindir}/llamacppcublassimple
|
||||
|
||||
mkdir -p %{buildroot}/usr/lib/systemd/system
|
||||
%{__cat} <<EOF > %{buildroot}/usr/lib/systemd/system/llamacuda.service
|
||||
%{__cat} <<EOF > %{buildroot}/usr/lib/systemd/system/llamacublas.service
|
||||
[Unit]
|
||||
Description=Llama.cpp server, CPU only (no GPU support in this build).
|
||||
After=syslog.target network.target local-fs.target remote-fs.target nss-lookup.target
|
||||
|
@ -49,7 +49,7 @@ After=syslog.target network.target local-fs.target remote-fs.target nss-lookup.t
|
|||
[Service]
|
||||
Type=simple
|
||||
EnvironmentFile=/etc/sysconfig/llama
|
||||
ExecStart=/usr/bin/llama-cuda-server $LLAMA_ARGS
|
||||
ExecStart=/usr/bin/llamacppcublasserver $LLAMA_ARGS
|
||||
ExecReload=/bin/kill -s HUP $MAINPID
|
||||
Restart=never
|
||||
|
||||
|
@ -67,10 +67,10 @@ rm -rf %{buildroot}
|
|||
rm -rf %{_builddir}/*
|
||||
|
||||
%files
|
||||
%{_bindir}/llama-cuda-cli
|
||||
%{_bindir}/llama-cuda-server
|
||||
%{_bindir}/llama-cuda-simple
|
||||
/usr/lib/systemd/system/llamacuda.service
|
||||
%{_bindir}/llamacppcublas
|
||||
%{_bindir}/llamacppcublasserver
|
||||
%{_bindir}/llamacppcublassimple
|
||||
/usr/lib/systemd/system/llamacublas.service
|
||||
%config /etc/sysconfig/llama
|
||||
|
||||
%pre
|
|
@ -1,5 +1,5 @@
|
|||
# SRPM for building from source and packaging an RPM for RPM-based distros.
|
||||
# https://docs.fedoraproject.org/en-US/quick-docs/creating-rpm-packages
|
||||
# https://fedoraproject.org/wiki/How_to_create_an_RPM_package
|
||||
# Built and maintained by John Boero - boeroboy@gmail.com
|
||||
# In honor of Seth Vidal https://www.redhat.com/it/blog/thank-you-seth-vidal
|
||||
|
||||
|
@ -38,9 +38,9 @@ make -j
|
|||
|
||||
%install
|
||||
mkdir -p %{buildroot}%{_bindir}/
|
||||
cp -p llama-cli %{buildroot}%{_bindir}/llama-cli
|
||||
cp -p llama-server %{buildroot}%{_bindir}/llama-server
|
||||
cp -p llama-simple %{buildroot}%{_bindir}/llama-simple
|
||||
cp -p main %{buildroot}%{_bindir}/llama
|
||||
cp -p server %{buildroot}%{_bindir}/llamaserver
|
||||
cp -p simple %{buildroot}%{_bindir}/llamasimple
|
||||
|
||||
mkdir -p %{buildroot}/usr/lib/systemd/system
|
||||
%{__cat} <<EOF > %{buildroot}/usr/lib/systemd/system/llama.service
|
||||
|
@ -51,7 +51,7 @@ After=syslog.target network.target local-fs.target remote-fs.target nss-lookup.t
|
|||
[Service]
|
||||
Type=simple
|
||||
EnvironmentFile=/etc/sysconfig/llama
|
||||
ExecStart=/usr/bin/llama-server $LLAMA_ARGS
|
||||
ExecStart=/usr/bin/llamaserver $LLAMA_ARGS
|
||||
ExecReload=/bin/kill -s HUP $MAINPID
|
||||
Restart=never
|
||||
|
||||
|
@ -69,9 +69,9 @@ rm -rf %{buildroot}
|
|||
rm -rf %{_builddir}/*
|
||||
|
||||
%files
|
||||
%{_bindir}/llama-cli
|
||||
%{_bindir}/llama-server
|
||||
%{_bindir}/llama-simple
|
||||
%{_bindir}/llama
|
||||
%{_bindir}/llamaserver
|
||||
%{_bindir}/llamasimple
|
||||
/usr/lib/systemd/system/llama.service
|
||||
%config /etc/sysconfig/llama
|
||||
|
||||
|
|
32
.devops/main-cuda.Dockerfile
Normal file
32
.devops/main-cuda.Dockerfile
Normal file
|
@ -0,0 +1,32 @@
|
|||
ARG UBUNTU_VERSION=22.04
|
||||
# This needs to generally match the container host's environment.
|
||||
ARG CUDA_VERSION=11.7.1
|
||||
# Target the CUDA build image
|
||||
ARG BASE_CUDA_DEV_CONTAINER=nvidia/cuda:${CUDA_VERSION}-devel-ubuntu${UBUNTU_VERSION}
|
||||
# Target the CUDA runtime image
|
||||
ARG BASE_CUDA_RUN_CONTAINER=nvidia/cuda:${CUDA_VERSION}-runtime-ubuntu${UBUNTU_VERSION}
|
||||
|
||||
FROM ${BASE_CUDA_DEV_CONTAINER} as build
|
||||
|
||||
# Unless otherwise specified, we make a fat build.
|
||||
ARG CUDA_DOCKER_ARCH=all
|
||||
|
||||
RUN apt-get update && \
|
||||
apt-get install -y build-essential git
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
COPY . .
|
||||
|
||||
# Set nvcc architecture
|
||||
ENV CUDA_DOCKER_ARCH=${CUDA_DOCKER_ARCH}
|
||||
# Enable cuBLAS
|
||||
ENV LLAMA_CUBLAS=1
|
||||
|
||||
RUN make
|
||||
|
||||
FROM ${BASE_CUDA_RUN_CONTAINER} as runtime
|
||||
|
||||
COPY --from=build /app/main /main
|
||||
|
||||
ENTRYPOINT [ "/main" ]
|
44
.devops/main-rocm.Dockerfile
Normal file
44
.devops/main-rocm.Dockerfile
Normal file
|
@ -0,0 +1,44 @@
|
|||
ARG UBUNTU_VERSION=22.04
|
||||
|
||||
# This needs to generally match the container host's environment.
|
||||
ARG ROCM_VERSION=5.6
|
||||
|
||||
# Target the CUDA build image
|
||||
ARG BASE_ROCM_DEV_CONTAINER=rocm/dev-ubuntu-${UBUNTU_VERSION}:${ROCM_VERSION}-complete
|
||||
|
||||
FROM ${BASE_ROCM_DEV_CONTAINER} as build
|
||||
|
||||
# Unless otherwise specified, we make a fat build.
|
||||
# List from https://github.com/ggerganov/llama.cpp/pull/1087#issuecomment-1682807878
|
||||
# This is mostly tied to rocBLAS supported archs.
|
||||
ARG ROCM_DOCKER_ARCH=\
|
||||
gfx803 \
|
||||
gfx900 \
|
||||
gfx906 \
|
||||
gfx908 \
|
||||
gfx90a \
|
||||
gfx1010 \
|
||||
gfx1030 \
|
||||
gfx1100 \
|
||||
gfx1101 \
|
||||
gfx1102
|
||||
|
||||
COPY requirements.txt requirements.txt
|
||||
|
||||
RUN pip install --upgrade pip setuptools wheel \
|
||||
&& pip install -r requirements.txt
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
COPY . .
|
||||
|
||||
# Set nvcc architecture
|
||||
ENV GPU_TARGETS=${ROCM_DOCKER_ARCH}
|
||||
# Enable ROCm
|
||||
ENV LLAMA_HIPBLAS=1
|
||||
ENV CC=/opt/rocm/llvm/bin/clang
|
||||
ENV CXX=/opt/rocm/llvm/bin/clang++
|
||||
|
||||
RUN make
|
||||
|
||||
ENTRYPOINT [ "/app/main" ]
|
20
.devops/main.Dockerfile
Normal file
20
.devops/main.Dockerfile
Normal file
|
@ -0,0 +1,20 @@
|
|||
ARG UBUNTU_VERSION=22.04
|
||||
|
||||
FROM ubuntu:$UBUNTU_VERSION as build
|
||||
|
||||
RUN apt-get update && \
|
||||
apt-get install -y build-essential git
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
COPY . .
|
||||
|
||||
RUN make
|
||||
|
||||
FROM ubuntu:$UBUNTU_VERSION as runtime
|
||||
|
||||
COPY --from=build /app/main /main
|
||||
|
||||
ENV LC_ALL=C.utf8
|
||||
|
||||
ENTRYPOINT [ "/main" ]
|
|
@ -1,108 +0,0 @@
|
|||
ARG UBUNTU_VERSION=22.04
|
||||
# This needs to generally match the container host's environment.
|
||||
ARG MUSA_VERSION=rc3.1.0
|
||||
# Target the MUSA build image
|
||||
ARG BASE_MUSA_DEV_CONTAINER=mthreads/musa:${MUSA_VERSION}-devel-ubuntu${UBUNTU_VERSION}
|
||||
|
||||
ARG BASE_MUSA_RUN_CONTAINER=mthreads/musa:${MUSA_VERSION}-runtime-ubuntu${UBUNTU_VERSION}
|
||||
|
||||
FROM ${BASE_MUSA_DEV_CONTAINER} AS build
|
||||
|
||||
# MUSA architecture to build for (defaults to all supported archs)
|
||||
ARG MUSA_DOCKER_ARCH=default
|
||||
|
||||
RUN apt-get update && \
|
||||
apt-get install -y \
|
||||
build-essential \
|
||||
cmake \
|
||||
python3 \
|
||||
python3-pip \
|
||||
git \
|
||||
libcurl4-openssl-dev \
|
||||
libgomp1
|
||||
|
||||
COPY requirements.txt requirements.txt
|
||||
COPY requirements requirements
|
||||
|
||||
RUN pip install --upgrade pip setuptools wheel \
|
||||
&& pip install -r requirements.txt
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
COPY . .
|
||||
|
||||
# Use the default MUSA archs if not specified
|
||||
RUN if [ "${MUSA_DOCKER_ARCH}" != "default" ]; then \
|
||||
export CMAKE_ARGS="-DMUSA_ARCHITECTURES=${MUSA_DOCKER_ARCH}"; \
|
||||
fi && \
|
||||
cmake -B build -DGGML_NATIVE=OFF -DGGML_MUSA=ON -DLLAMA_CURL=ON ${CMAKE_ARGS} -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined . && \
|
||||
cmake --build build --config Release -j$(nproc)
|
||||
|
||||
RUN mkdir -p /app/lib && \
|
||||
find build -name "*.so" -exec cp {} /app/lib \;
|
||||
|
||||
RUN mkdir -p /app/full \
|
||||
&& cp build/bin/* /app/full \
|
||||
&& cp *.py /app/full \
|
||||
&& cp -r gguf-py /app/full \
|
||||
&& cp -r requirements /app/full \
|
||||
&& cp requirements.txt /app/full \
|
||||
&& cp .devops/tools.sh /app/full/tools.sh
|
||||
|
||||
## Base image
|
||||
FROM ${BASE_MUSA_RUN_CONTAINER} AS base
|
||||
|
||||
RUN apt-get update \
|
||||
&& apt-get install -y libgomp1 curl\
|
||||
&& apt autoremove -y \
|
||||
&& apt clean -y \
|
||||
&& rm -rf /tmp/* /var/tmp/* \
|
||||
&& find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
|
||||
&& find /var/cache -type f -delete
|
||||
|
||||
COPY --from=build /app/lib/ /app
|
||||
|
||||
### Full
|
||||
FROM base AS full
|
||||
|
||||
COPY --from=build /app/full /app
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
RUN apt-get update \
|
||||
&& apt-get install -y \
|
||||
git \
|
||||
python3 \
|
||||
python3-pip \
|
||||
&& pip install --upgrade pip setuptools wheel \
|
||||
&& pip install -r requirements.txt \
|
||||
&& apt autoremove -y \
|
||||
&& apt clean -y \
|
||||
&& rm -rf /tmp/* /var/tmp/* \
|
||||
&& find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
|
||||
&& find /var/cache -type f -delete
|
||||
|
||||
|
||||
ENTRYPOINT ["/app/tools.sh"]
|
||||
|
||||
### Light, CLI only
|
||||
FROM base AS light
|
||||
|
||||
COPY --from=build /app/full/llama-cli /app
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
ENTRYPOINT [ "/app/llama-cli" ]
|
||||
|
||||
### Server, Server only
|
||||
FROM base AS server
|
||||
|
||||
ENV LLAMA_ARG_HOST=0.0.0.0
|
||||
|
||||
COPY --from=build /app/full/llama-server /app
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ]
|
||||
|
||||
ENTRYPOINT [ "/app/llama-server" ]
|
|
@ -1,21 +0,0 @@
|
|||
{
|
||||
perSystem =
|
||||
{ config, lib, ... }:
|
||||
{
|
||||
apps =
|
||||
let
|
||||
inherit (config.packages) default;
|
||||
binaries = [
|
||||
"llama-cli"
|
||||
"llama-embedding"
|
||||
"llama-server"
|
||||
"llama-quantize"
|
||||
];
|
||||
mkApp = name: {
|
||||
type = "app";
|
||||
program = "${default}/bin/${name}";
|
||||
};
|
||||
in
|
||||
lib.genAttrs binaries mkApp;
|
||||
};
|
||||
}
|
|
@ -1,52 +0,0 @@
|
|||
{ inputs, ... }:
|
||||
|
||||
{
|
||||
perSystem =
|
||||
{
|
||||
config,
|
||||
lib,
|
||||
system,
|
||||
...
|
||||
}:
|
||||
{
|
||||
devShells =
|
||||
let
|
||||
pkgs = import inputs.nixpkgs { inherit system; };
|
||||
stdenv = pkgs.stdenv;
|
||||
scripts = config.packages.python-scripts;
|
||||
in
|
||||
lib.pipe (config.packages) [
|
||||
(lib.concatMapAttrs (
|
||||
name: package: {
|
||||
${name} = pkgs.mkShell {
|
||||
name = "${name}";
|
||||
inputsFrom = [ package ];
|
||||
shellHook = ''
|
||||
echo "Entering ${name} devShell"
|
||||
'';
|
||||
};
|
||||
"${name}-extra" =
|
||||
if (name == "python-scripts") then
|
||||
null
|
||||
else
|
||||
pkgs.mkShell {
|
||||
name = "${name}-extra";
|
||||
inputsFrom = [
|
||||
package
|
||||
scripts
|
||||
];
|
||||
# Extra packages that *may* be used by some scripts
|
||||
packages = [
|
||||
pkgs.python3Packages.tiktoken
|
||||
];
|
||||
shellHook = ''
|
||||
echo "Entering ${name} devShell"
|
||||
addToSearchPath "LD_LIBRARY_PATH" "${lib.getLib stdenv.cc.cc}/lib"
|
||||
'';
|
||||
};
|
||||
}
|
||||
))
|
||||
(lib.filterAttrs (name: value: value != null))
|
||||
];
|
||||
};
|
||||
}
|
|
@ -1,37 +0,0 @@
|
|||
{
|
||||
lib,
|
||||
dockerTools,
|
||||
buildEnv,
|
||||
llama-cpp,
|
||||
interactive ? true,
|
||||
coreutils,
|
||||
}:
|
||||
|
||||
# A tar that can be fed into `docker load`:
|
||||
#
|
||||
# $ nix build .#llamaPackages.docker
|
||||
# $ docker load < result
|
||||
|
||||
# For details and variations cf.
|
||||
# - https://nixos.org/manual/nixpkgs/unstable/#ssec-pkgs-dockerTools-buildLayeredImage
|
||||
# - https://discourse.nixos.org/t/a-faster-dockertools-buildimage-prototype/16922
|
||||
# - https://nixery.dev/
|
||||
|
||||
# Approximate (compressed) sizes, at the time of writing, are:
|
||||
#
|
||||
# .#llamaPackages.docker: 125M;
|
||||
# .#llamaPackagesCuda.docker: 537M;
|
||||
# .#legacyPackages.aarch64-linux.llamaPackagesXavier.docker: 415M.
|
||||
|
||||
dockerTools.buildLayeredImage {
|
||||
name = llama-cpp.pname;
|
||||
tag = "latest";
|
||||
|
||||
contents =
|
||||
[ llama-cpp ]
|
||||
++ lib.optionals interactive [
|
||||
coreutils
|
||||
dockerTools.binSh
|
||||
dockerTools.caCertificates
|
||||
];
|
||||
}
|
|
@ -1,39 +0,0 @@
|
|||
{ inputs, ... }:
|
||||
{
|
||||
perSystem =
|
||||
{
|
||||
config,
|
||||
system,
|
||||
lib,
|
||||
pkgsCuda,
|
||||
...
|
||||
}:
|
||||
{
|
||||
legacyPackages =
|
||||
let
|
||||
caps.llamaPackagesXavier = "7.2";
|
||||
caps.llamaPackagesOrin = "8.7";
|
||||
caps.llamaPackagesTX2 = "6.2";
|
||||
caps.llamaPackagesNano = "5.3";
|
||||
|
||||
pkgsFor =
|
||||
cap:
|
||||
import inputs.nixpkgs {
|
||||
inherit system;
|
||||
config = {
|
||||
cudaSupport = true;
|
||||
cudaCapabilities = [ cap ];
|
||||
cudaEnableForwardCompat = false;
|
||||
inherit (pkgsCuda.config) allowUnfreePredicate;
|
||||
};
|
||||
};
|
||||
in
|
||||
builtins.mapAttrs (name: cap: (pkgsFor cap).callPackage ./scope.nix { }) caps;
|
||||
|
||||
packages = lib.optionalAttrs (system == "aarch64-linux") {
|
||||
jetson-xavier = config.legacyPackages.llamaPackagesXavier.llama-cpp;
|
||||
jetson-orin = config.legacyPackages.llamaPackagesOrin.llama-cpp;
|
||||
jetson-nano = config.legacyPackages.llamaPackagesNano.llama-cpp;
|
||||
};
|
||||
};
|
||||
}
|
|
@ -1,45 +0,0 @@
|
|||
{ inputs, ... }:
|
||||
{
|
||||
# The _module.args definitions are passed on to modules as arguments. E.g.
|
||||
# the module `{ pkgs ... }: { /* config */ }` implicitly uses
|
||||
# `_module.args.pkgs` (defined in this case by flake-parts).
|
||||
perSystem =
|
||||
{ system, ... }:
|
||||
{
|
||||
_module.args = {
|
||||
# Note: bringing up https://zimbatm.com/notes/1000-instances-of-nixpkgs
|
||||
# again, the below creates several nixpkgs instances which the
|
||||
# flake-centric CLI will be forced to evaluate e.g. on `nix flake show`.
|
||||
#
|
||||
# This is currently "slow" and "expensive", on a certain scale.
|
||||
# This also isn't "right" in that this hinders dependency injection at
|
||||
# the level of flake inputs. This might get removed in the foreseeable
|
||||
# future.
|
||||
#
|
||||
# Note that you can use these expressions without Nix
|
||||
# (`pkgs.callPackage ./devops/nix/scope.nix { }` is the entry point).
|
||||
|
||||
pkgsCuda = import inputs.nixpkgs {
|
||||
inherit system;
|
||||
# Ensure dependencies use CUDA consistently (e.g. that openmpi, ucc,
|
||||
# and ucx are built with CUDA support)
|
||||
config.cudaSupport = true;
|
||||
config.allowUnfreePredicate =
|
||||
p:
|
||||
builtins.all (
|
||||
license:
|
||||
license.free
|
||||
|| builtins.elem license.shortName [
|
||||
"CUDA EULA"
|
||||
"cuDNN EULA"
|
||||
]
|
||||
) (p.meta.licenses or [ p.meta.license ]);
|
||||
};
|
||||
# Ensure dependencies use ROCm consistently
|
||||
pkgsRocm = import inputs.nixpkgs {
|
||||
inherit system;
|
||||
config.rocmSupport = true;
|
||||
};
|
||||
};
|
||||
};
|
||||
}
|
|
@ -1,36 +0,0 @@
|
|||
{
|
||||
lib,
|
||||
llamaVersion,
|
||||
numpy,
|
||||
tqdm,
|
||||
sentencepiece,
|
||||
pyyaml,
|
||||
poetry-core,
|
||||
buildPythonPackage,
|
||||
pytestCheckHook,
|
||||
}:
|
||||
|
||||
buildPythonPackage {
|
||||
pname = "gguf";
|
||||
version = llamaVersion;
|
||||
pyproject = true;
|
||||
nativeBuildInputs = [ poetry-core ];
|
||||
propagatedBuildInputs = [
|
||||
numpy
|
||||
tqdm
|
||||
sentencepiece
|
||||
pyyaml
|
||||
];
|
||||
src = lib.cleanSource ../../gguf-py;
|
||||
pythonImportsCheck = [
|
||||
"numpy"
|
||||
"gguf"
|
||||
];
|
||||
nativeCheckInputs = [ pytestCheckHook ];
|
||||
doCheck = true;
|
||||
meta = with lib; {
|
||||
description = "Python package for writing binary files in the GGUF format";
|
||||
license = licenses.mit;
|
||||
maintainers = [ maintainers.ditsuke ];
|
||||
};
|
||||
}
|
|
@ -1,247 +0,0 @@
|
|||
{
|
||||
lib,
|
||||
glibc,
|
||||
config,
|
||||
stdenv,
|
||||
runCommand,
|
||||
cmake,
|
||||
ninja,
|
||||
pkg-config,
|
||||
git,
|
||||
mpi,
|
||||
blas,
|
||||
cudaPackages,
|
||||
autoAddDriverRunpath,
|
||||
darwin,
|
||||
rocmPackages,
|
||||
vulkan-headers,
|
||||
vulkan-loader,
|
||||
curl,
|
||||
shaderc,
|
||||
useBlas ?
|
||||
builtins.all (x: !x) [
|
||||
useCuda
|
||||
useMetalKit
|
||||
useRocm
|
||||
useVulkan
|
||||
]
|
||||
&& blas.meta.available,
|
||||
useCuda ? config.cudaSupport,
|
||||
useMetalKit ? stdenv.isAarch64 && stdenv.isDarwin,
|
||||
# Increases the runtime closure size by ~700M
|
||||
useMpi ? false,
|
||||
useRocm ? config.rocmSupport,
|
||||
rocmGpuTargets ? builtins.concatStringsSep ";" rocmPackages.clr.gpuTargets,
|
||||
enableCurl ? true,
|
||||
useVulkan ? false,
|
||||
llamaVersion ? "0.0.0", # Arbitrary version, substituted by the flake
|
||||
|
||||
# It's necessary to consistently use backendStdenv when building with CUDA support,
|
||||
# otherwise we get libstdc++ errors downstream.
|
||||
effectiveStdenv ? if useCuda then cudaPackages.backendStdenv else stdenv,
|
||||
enableStatic ? effectiveStdenv.hostPlatform.isStatic,
|
||||
precompileMetalShaders ? false,
|
||||
}:
|
||||
|
||||
let
|
||||
inherit (lib)
|
||||
cmakeBool
|
||||
cmakeFeature
|
||||
optionals
|
||||
strings
|
||||
;
|
||||
|
||||
stdenv = throw "Use effectiveStdenv instead";
|
||||
|
||||
suffices =
|
||||
lib.optionals useBlas [ "BLAS" ]
|
||||
++ lib.optionals useCuda [ "CUDA" ]
|
||||
++ lib.optionals useMetalKit [ "MetalKit" ]
|
||||
++ lib.optionals useMpi [ "MPI" ]
|
||||
++ lib.optionals useRocm [ "ROCm" ]
|
||||
++ lib.optionals useVulkan [ "Vulkan" ];
|
||||
|
||||
pnameSuffix =
|
||||
strings.optionalString (suffices != [ ])
|
||||
"-${strings.concatMapStringsSep "-" strings.toLower suffices}";
|
||||
descriptionSuffix = strings.optionalString (
|
||||
suffices != [ ]
|
||||
) ", accelerated with ${strings.concatStringsSep ", " suffices}";
|
||||
|
||||
xcrunHost = runCommand "xcrunHost" { } ''
|
||||
mkdir -p $out/bin
|
||||
ln -s /usr/bin/xcrun $out/bin
|
||||
'';
|
||||
|
||||
# apple_sdk is supposed to choose sane defaults, no need to handle isAarch64
|
||||
# separately
|
||||
darwinBuildInputs =
|
||||
with darwin.apple_sdk.frameworks;
|
||||
[
|
||||
Accelerate
|
||||
CoreVideo
|
||||
CoreGraphics
|
||||
]
|
||||
++ optionals useMetalKit [ MetalKit ];
|
||||
|
||||
cudaBuildInputs = with cudaPackages; [
|
||||
cuda_cudart
|
||||
cuda_cccl # <nv/target>
|
||||
libcublas
|
||||
];
|
||||
|
||||
rocmBuildInputs = with rocmPackages; [
|
||||
clr
|
||||
hipblas
|
||||
rocblas
|
||||
];
|
||||
|
||||
vulkanBuildInputs = [
|
||||
vulkan-headers
|
||||
vulkan-loader
|
||||
shaderc
|
||||
];
|
||||
in
|
||||
|
||||
effectiveStdenv.mkDerivation (finalAttrs: {
|
||||
pname = "llama-cpp${pnameSuffix}";
|
||||
version = llamaVersion;
|
||||
|
||||
# Note: none of the files discarded here are visible in the sandbox or
|
||||
# affect the output hash. This also means they can be modified without
|
||||
# triggering a rebuild.
|
||||
src = lib.cleanSourceWith {
|
||||
filter =
|
||||
name: type:
|
||||
let
|
||||
noneOf = builtins.all (x: !x);
|
||||
baseName = baseNameOf name;
|
||||
in
|
||||
noneOf [
|
||||
(lib.hasSuffix ".nix" name) # Ignore *.nix files when computing outPaths
|
||||
(lib.hasSuffix ".md" name) # Ignore *.md changes whe computing outPaths
|
||||
(lib.hasPrefix "." baseName) # Skip hidden files and directories
|
||||
(baseName == "flake.lock")
|
||||
];
|
||||
src = lib.cleanSource ../../.;
|
||||
};
|
||||
|
||||
postPatch = ''
|
||||
substituteInPlace ./ggml/src/ggml-metal/ggml-metal.m \
|
||||
--replace '[bundle pathForResource:@"ggml-metal" ofType:@"metal"];' "@\"$out/bin/ggml-metal.metal\";"
|
||||
substituteInPlace ./ggml/src/ggml-metal/ggml-metal.m \
|
||||
--replace '[bundle pathForResource:@"default" ofType:@"metallib"];' "@\"$out/bin/default.metallib\";"
|
||||
'';
|
||||
|
||||
# With PR#6015 https://github.com/ggerganov/llama.cpp/pull/6015,
|
||||
# `default.metallib` may be compiled with Metal compiler from XCode
|
||||
# and we need to escape sandbox on MacOS to access Metal compiler.
|
||||
# `xcrun` is used find the path of the Metal compiler, which is varible
|
||||
# and not on $PATH
|
||||
# see https://github.com/ggerganov/llama.cpp/pull/6118 for discussion
|
||||
__noChroot = effectiveStdenv.isDarwin && useMetalKit && precompileMetalShaders;
|
||||
|
||||
nativeBuildInputs =
|
||||
[
|
||||
cmake
|
||||
ninja
|
||||
pkg-config
|
||||
git
|
||||
]
|
||||
++ optionals useCuda [
|
||||
cudaPackages.cuda_nvcc
|
||||
|
||||
autoAddDriverRunpath
|
||||
]
|
||||
++ optionals (effectiveStdenv.hostPlatform.isGnu && enableStatic) [ glibc.static ]
|
||||
++ optionals (effectiveStdenv.isDarwin && useMetalKit && precompileMetalShaders) [ xcrunHost ];
|
||||
|
||||
buildInputs =
|
||||
optionals effectiveStdenv.isDarwin darwinBuildInputs
|
||||
++ optionals useCuda cudaBuildInputs
|
||||
++ optionals useMpi [ mpi ]
|
||||
++ optionals useRocm rocmBuildInputs
|
||||
++ optionals useBlas [ blas ]
|
||||
++ optionals useVulkan vulkanBuildInputs
|
||||
++ optionals enableCurl [ curl ];
|
||||
|
||||
cmakeFlags =
|
||||
[
|
||||
(cmakeBool "LLAMA_BUILD_SERVER" true)
|
||||
(cmakeBool "BUILD_SHARED_LIBS" (!enableStatic))
|
||||
(cmakeBool "CMAKE_SKIP_BUILD_RPATH" true)
|
||||
(cmakeBool "LLAMA_CURL" enableCurl)
|
||||
(cmakeBool "GGML_NATIVE" false)
|
||||
(cmakeBool "GGML_BLAS" useBlas)
|
||||
(cmakeBool "GGML_CUDA" useCuda)
|
||||
(cmakeBool "GGML_HIP" useRocm)
|
||||
(cmakeBool "GGML_METAL" useMetalKit)
|
||||
(cmakeBool "GGML_VULKAN" useVulkan)
|
||||
(cmakeBool "GGML_STATIC" enableStatic)
|
||||
]
|
||||
++ optionals useCuda [
|
||||
(
|
||||
with cudaPackages.flags;
|
||||
cmakeFeature "CMAKE_CUDA_ARCHITECTURES" (
|
||||
builtins.concatStringsSep ";" (map dropDot cudaCapabilities)
|
||||
)
|
||||
)
|
||||
]
|
||||
++ optionals useRocm [
|
||||
(cmakeFeature "CMAKE_HIP_COMPILER" "${rocmPackages.llvm.clang}/bin/clang")
|
||||
(cmakeFeature "CMAKE_HIP_ARCHITECTURES" rocmGpuTargets)
|
||||
]
|
||||
++ optionals useMetalKit [
|
||||
(lib.cmakeFeature "CMAKE_C_FLAGS" "-D__ARM_FEATURE_DOTPROD=1")
|
||||
(cmakeBool "GGML_METAL_EMBED_LIBRARY" (!precompileMetalShaders))
|
||||
];
|
||||
|
||||
# Environment variables needed for ROCm
|
||||
env = optionals useRocm {
|
||||
ROCM_PATH = "${rocmPackages.clr}";
|
||||
HIP_DEVICE_LIB_PATH = "${rocmPackages.rocm-device-libs}/amdgcn/bitcode";
|
||||
};
|
||||
|
||||
# TODO(SomeoneSerge): It's better to add proper install targets at the CMake level,
|
||||
# if they haven't been added yet.
|
||||
postInstall = ''
|
||||
mkdir -p $out/include
|
||||
cp $src/include/llama.h $out/include/
|
||||
'';
|
||||
|
||||
meta = {
|
||||
# Configurations we don't want even the CI to evaluate. Results in the
|
||||
# "unsupported platform" messages. This is mostly a no-op, because
|
||||
# cudaPackages would've refused to evaluate anyway.
|
||||
badPlatforms = optionals useCuda lib.platforms.darwin;
|
||||
|
||||
# Configurations that are known to result in build failures. Can be
|
||||
# overridden by importing Nixpkgs with `allowBroken = true`.
|
||||
broken = (useMetalKit && !effectiveStdenv.isDarwin);
|
||||
|
||||
description = "Inference of LLaMA model in pure C/C++${descriptionSuffix}";
|
||||
homepage = "https://github.com/ggerganov/llama.cpp/";
|
||||
license = lib.licenses.mit;
|
||||
|
||||
# Accommodates `nix run` and `lib.getExe`
|
||||
mainProgram = "llama-cli";
|
||||
|
||||
# These people might respond, on the best effort basis, if you ping them
|
||||
# in case of Nix-specific regressions or for reviewing Nix-specific PRs.
|
||||
# Consider adding yourself to this list if you want to ensure this flake
|
||||
# stays maintained and you're willing to invest your time. Do not add
|
||||
# other people without their consent. Consider removing people after
|
||||
# they've been unreachable for long periods of time.
|
||||
|
||||
# Note that lib.maintainers is defined in Nixpkgs, but you may just add
|
||||
# an attrset following the same format as in
|
||||
# https://github.com/NixOS/nixpkgs/blob/f36a80e54da29775c78d7eff0e628c2b4e34d1d7/maintainers/maintainer-list.nix
|
||||
maintainers = with lib.maintainers; [
|
||||
philiptaron
|
||||
SomeoneSerge
|
||||
];
|
||||
|
||||
# Extend `badPlatforms` instead
|
||||
platforms = lib.platforms.all;
|
||||
};
|
||||
})
|
|
@ -1,66 +0,0 @@
|
|||
{
|
||||
lib,
|
||||
stdenv,
|
||||
buildPythonPackage,
|
||||
poetry-core,
|
||||
mkShell,
|
||||
python3Packages,
|
||||
gguf-py,
|
||||
}@inputs:
|
||||
|
||||
let
|
||||
llama-python-deps = with python3Packages; [
|
||||
numpy
|
||||
sentencepiece
|
||||
transformers
|
||||
protobuf
|
||||
torchWithoutCuda
|
||||
gguf-py
|
||||
tqdm
|
||||
|
||||
# for scripts/compare-llama-bench.py
|
||||
gitpython
|
||||
tabulate
|
||||
|
||||
# for examples/pydantic-models-to-grammar-examples.py
|
||||
docstring-parser
|
||||
pydantic
|
||||
|
||||
];
|
||||
|
||||
llama-python-test-deps = with python3Packages; [
|
||||
# Server bench
|
||||
matplotlib
|
||||
|
||||
# server tests
|
||||
openai
|
||||
pytest
|
||||
prometheus-client
|
||||
];
|
||||
in
|
||||
|
||||
buildPythonPackage ({
|
||||
pname = "llama-scripts";
|
||||
version = "0.0.0";
|
||||
pyproject = true;
|
||||
|
||||
# NOTE: The files filtered out here are not visible in the build sandbox, neither
|
||||
# do they affect the output hash. They can be modified without triggering a rebuild.
|
||||
src = lib.cleanSourceWith {
|
||||
filter =
|
||||
name: type:
|
||||
let
|
||||
any = builtins.any (x: x);
|
||||
baseName = builtins.baseNameOf name;
|
||||
in
|
||||
any [
|
||||
(lib.hasSuffix ".py" name)
|
||||
(baseName == "README.md")
|
||||
(baseName == "pyproject.toml")
|
||||
];
|
||||
src = lib.cleanSource ../../.;
|
||||
};
|
||||
nativeBuildInputs = [ poetry-core ];
|
||||
nativeCheckInputs = llama-python-test-deps;
|
||||
dependencies = llama-python-deps;
|
||||
})
|
|
@ -1,41 +0,0 @@
|
|||
{
|
||||
lib,
|
||||
newScope,
|
||||
python3,
|
||||
llamaVersion ? "0.0.0",
|
||||
}:
|
||||
|
||||
let
|
||||
pythonPackages = python3.pkgs;
|
||||
buildPythonPackage = pythonPackages.buildPythonPackage;
|
||||
numpy = pythonPackages.numpy;
|
||||
tqdm = pythonPackages.tqdm;
|
||||
sentencepiece = pythonPackages.sentencepiece;
|
||||
pyyaml = pythonPackages.pyyaml;
|
||||
poetry-core = pythonPackages.poetry-core;
|
||||
pytestCheckHook = pythonPackages.pytestCheckHook;
|
||||
in
|
||||
|
||||
# We're using `makeScope` instead of just writing out an attrset
|
||||
# because it allows users to apply overlays later using `overrideScope'`.
|
||||
# Cf. https://noogle.dev/f/lib/makeScope
|
||||
|
||||
lib.makeScope newScope (self: {
|
||||
inherit llamaVersion;
|
||||
gguf-py = self.callPackage ./package-gguf-py.nix {
|
||||
inherit
|
||||
buildPythonPackage
|
||||
numpy
|
||||
tqdm
|
||||
sentencepiece
|
||||
poetry-core
|
||||
pyyaml
|
||||
pytestCheckHook
|
||||
;
|
||||
};
|
||||
python-scripts = self.callPackage ./python-scripts.nix { inherit buildPythonPackage poetry-core; };
|
||||
llama-cpp = self.callPackage ./package.nix { };
|
||||
docker = self.callPackage ./docker.nix { };
|
||||
docker-min = self.callPackage ./docker.nix { interactive = false; };
|
||||
sif = self.callPackage ./sif.nix { };
|
||||
})
|
|
@ -1,27 +0,0 @@
|
|||
{
|
||||
lib,
|
||||
singularity-tools,
|
||||
llama-cpp,
|
||||
bashInteractive,
|
||||
interactive ? false,
|
||||
}:
|
||||
|
||||
let
|
||||
optionalInt = cond: x: if cond then x else 0;
|
||||
in
|
||||
singularity-tools.buildImage rec {
|
||||
inherit (llama-cpp) name;
|
||||
contents = [ llama-cpp ] ++ lib.optionals interactive [ bashInteractive ];
|
||||
|
||||
# These are excessive (but safe) for most variants. Building singularity
|
||||
# images requires superuser privileges, so we build them inside a VM in a
|
||||
# writable image of pre-determined size.
|
||||
#
|
||||
# ROCm is currently affected by https://github.com/NixOS/nixpkgs/issues/276846
|
||||
#
|
||||
# Expected image sizes:
|
||||
# - cpu/blas: 150M,
|
||||
# - cuda, all gencodes: 560M,
|
||||
diskSize = 4096 + optionalInt llama-cpp.useRocm 16384;
|
||||
memSize = diskSize;
|
||||
}
|
|
@ -1,113 +0,0 @@
|
|||
ARG UBUNTU_VERSION=24.04
|
||||
|
||||
# This needs to generally match the container host's environment.
|
||||
ARG ROCM_VERSION=6.3
|
||||
ARG AMDGPU_VERSION=6.3
|
||||
|
||||
# Target the CUDA build image
|
||||
ARG BASE_ROCM_DEV_CONTAINER=rocm/dev-ubuntu-${UBUNTU_VERSION}:${ROCM_VERSION}-complete
|
||||
|
||||
### Build image
|
||||
FROM ${BASE_ROCM_DEV_CONTAINER} AS build
|
||||
|
||||
# Unless otherwise specified, we make a fat build.
|
||||
# List from https://github.com/ggerganov/llama.cpp/pull/1087#issuecomment-1682807878
|
||||
# This is mostly tied to rocBLAS supported archs.
|
||||
# gfx803, gfx900, gfx1032, gfx1101, gfx1102,not officialy supported
|
||||
# gfx906 is deprecated
|
||||
#check https://rocm.docs.amd.com/projects/install-on-linux/en/docs-6.2.4/reference/system-requirements.html
|
||||
|
||||
#ARG ROCM_DOCKER_ARCH='gfx803,gfx900,gfx906,gfx908,gfx90a,gfx942,gfx1010,gfx1030,gfx1032,gfx1100,gfx1101,gfx1102'
|
||||
ARG ROCM_DOCKER_ARCH=gfx1100
|
||||
|
||||
# Set nvcc architectured
|
||||
ENV AMDGPU_TARGETS=${ROCM_DOCKER_ARCH}
|
||||
# Enable ROCm
|
||||
# ENV CC=/opt/rocm/llvm/bin/clang
|
||||
# ENV CXX=/opt/rocm/llvm/bin/clang++
|
||||
|
||||
RUN apt-get update \
|
||||
&& apt-get install -y \
|
||||
build-essential \
|
||||
cmake \
|
||||
git \
|
||||
libcurl4-openssl-dev \
|
||||
curl \
|
||||
libgomp1
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
COPY . .
|
||||
|
||||
RUN HIPCXX="$(hipconfig -l)/clang" HIP_PATH="$(hipconfig -R)" \
|
||||
cmake -S . -B build -DGGML_HIP=ON -DAMDGPU_TARGETS=$ROCM_DOCKER_ARCH -DCMAKE_BUILD_TYPE=Release -DLLAMA_CURL=ON \
|
||||
&& cmake --build build --config Release -j$(nproc)
|
||||
|
||||
RUN mkdir -p /app/lib \
|
||||
&& find build -name "*.so" -exec cp {} /app/lib \;
|
||||
|
||||
RUN mkdir -p /app/full \
|
||||
&& cp build/bin/* /app/full \
|
||||
&& cp *.py /app/full \
|
||||
&& cp -r gguf-py /app/full \
|
||||
&& cp -r requirements /app/full \
|
||||
&& cp requirements.txt /app/full \
|
||||
&& cp .devops/tools.sh /app/full/tools.sh
|
||||
|
||||
## Base image
|
||||
FROM ${BASE_ROCM_DEV_CONTAINER} AS base
|
||||
|
||||
RUN apt-get update \
|
||||
&& apt-get install -y libgomp1 curl\
|
||||
&& apt autoremove -y \
|
||||
&& apt clean -y \
|
||||
&& rm -rf /tmp/* /var/tmp/* \
|
||||
&& find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
|
||||
&& find /var/cache -type f -delete
|
||||
|
||||
COPY --from=build /app/lib/ /app
|
||||
|
||||
### Full
|
||||
FROM base AS full
|
||||
|
||||
COPY --from=build /app/full /app
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
RUN apt-get update \
|
||||
&& apt-get install -y \
|
||||
git \
|
||||
python3-pip \
|
||||
python3 \
|
||||
python3-wheel\
|
||||
&& pip install --break-system-packages --upgrade setuptools \
|
||||
&& pip install --break-system-packages -r requirements.txt \
|
||||
&& apt autoremove -y \
|
||||
&& apt clean -y \
|
||||
&& rm -rf /tmp/* /var/tmp/* \
|
||||
&& find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
|
||||
&& find /var/cache -type f -delete
|
||||
|
||||
ENTRYPOINT ["/app/tools.sh"]
|
||||
|
||||
### Light, CLI only
|
||||
FROM base AS light
|
||||
|
||||
COPY --from=build /app/full/llama-cli /app
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
ENTRYPOINT [ "/app/llama-cli" ]
|
||||
|
||||
### Server, Server only
|
||||
FROM base AS server
|
||||
|
||||
ENV LLAMA_ARG_HOST=0.0.0.0
|
||||
|
||||
COPY --from=build /app/full/llama-server /app
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ]
|
||||
|
||||
ENTRYPOINT [ "/app/llama-server" ]
|
|
@ -8,40 +8,36 @@ arg1="$1"
|
|||
shift
|
||||
|
||||
if [[ "$arg1" == '--convert' || "$arg1" == '-c' ]]; then
|
||||
exec python3 ./convert_hf_to_gguf.py "$@"
|
||||
python3 ./convert.py "$@"
|
||||
elif [[ "$arg1" == '--quantize' || "$arg1" == '-q' ]]; then
|
||||
exec ./llama-quantize "$@"
|
||||
./quantize "$@"
|
||||
elif [[ "$arg1" == '--run' || "$arg1" == '-r' ]]; then
|
||||
exec ./llama-cli "$@"
|
||||
elif [[ "$arg1" == '--bench' || "$arg1" == '-b' ]]; then
|
||||
exec ./llama-bench "$@"
|
||||
elif [[ "$arg1" == '--perplexity' || "$arg1" == '-p' ]]; then
|
||||
exec ./llama-perplexity "$@"
|
||||
./main "$@"
|
||||
elif [[ "$arg1" == '--finetune' || "$arg1" == '-f' ]]; then
|
||||
./finetune "$@"
|
||||
elif [[ "$arg1" == '--all-in-one' || "$arg1" == '-a' ]]; then
|
||||
echo "Converting PTH to GGML..."
|
||||
for i in $(ls $1/$2/ggml-model-f16.bin*); do
|
||||
for i in `ls $1/$2/ggml-model-f16.bin*`; do
|
||||
if [ -f "${i/f16/q4_0}" ]; then
|
||||
echo "Skip model quantization, it already exists: ${i/f16/q4_0}"
|
||||
else
|
||||
echo "Converting PTH to GGML: $i into ${i/f16/q4_0}..."
|
||||
exec ./llama-quantize "$i" "${i/f16/q4_0}" q4_0
|
||||
./quantize "$i" "${i/f16/q4_0}" q4_0
|
||||
fi
|
||||
done
|
||||
elif [[ "$arg1" == '--server' || "$arg1" == '-s' ]]; then
|
||||
exec ./llama-server "$@"
|
||||
./server "$@"
|
||||
else
|
||||
echo "Unknown command: $arg1"
|
||||
echo "Available commands: "
|
||||
echo " --run (-r): Run a model previously converted into ggml"
|
||||
echo " ex: -m /models/7B/ggml-model-q4_0.bin -p \"Building a website can be done in 10 simple steps:\" -n 512"
|
||||
echo " --bench (-b): Benchmark the performance of the inference for various parameters."
|
||||
echo " ex: -m model.gguf"
|
||||
echo " --perplexity (-p): Measure the perplexity of a model over a given text."
|
||||
echo " ex: -m model.gguf -f file.txt"
|
||||
echo " --convert (-c): Convert a llama model into ggml"
|
||||
echo " ex: --outtype f16 \"/models/7B/\" "
|
||||
echo " --quantize (-q): Optimize with quantization process ggml"
|
||||
echo " ex: \"/models/7B/ggml-model-f16.bin\" \"/models/7B/ggml-model-q4_0.bin\" 2"
|
||||
echo " --finetune (-f): Run finetune command to create a lora finetune of the model"
|
||||
echo " See documentation for finetune for command-line parameters"
|
||||
echo " --all-in-one (-a): Execute --convert & --quantize"
|
||||
echo " ex: \"/models/\" 7B"
|
||||
echo " --server (-s): Run a model on the server"
|
||||
|
|
|
@ -1,89 +0,0 @@
|
|||
ARG UBUNTU_VERSION=24.04
|
||||
|
||||
FROM ubuntu:$UBUNTU_VERSION AS build
|
||||
|
||||
# Install build tools
|
||||
RUN apt update && apt install -y git build-essential cmake wget
|
||||
|
||||
# Install Vulkan SDK and cURL
|
||||
RUN wget -qO - https://packages.lunarg.com/lunarg-signing-key-pub.asc | apt-key add - && \
|
||||
wget -qO /etc/apt/sources.list.d/lunarg-vulkan-noble.list https://packages.lunarg.com/vulkan/lunarg-vulkan-noble.list && \
|
||||
apt update -y && \
|
||||
apt-get install -y vulkan-sdk libcurl4-openssl-dev curl
|
||||
|
||||
# Build it
|
||||
WORKDIR /app
|
||||
|
||||
COPY . .
|
||||
|
||||
RUN cmake -B build -DGGML_NATIVE=OFF -DGGML_VULKAN=1 -DLLAMA_CURL=1 && \
|
||||
cmake --build build --config Release -j$(nproc)
|
||||
|
||||
RUN mkdir -p /app/lib && \
|
||||
find build -name "*.so" -exec cp {} /app/lib \;
|
||||
|
||||
RUN mkdir -p /app/full \
|
||||
&& cp build/bin/* /app/full \
|
||||
&& cp *.py /app/full \
|
||||
&& cp -r gguf-py /app/full \
|
||||
&& cp -r requirements /app/full \
|
||||
&& cp requirements.txt /app/full \
|
||||
&& cp .devops/tools.sh /app/full/tools.sh
|
||||
|
||||
## Base image
|
||||
FROM ubuntu:$UBUNTU_VERSION AS base
|
||||
|
||||
RUN apt-get update \
|
||||
&& apt-get install -y libgomp1 curl libvulkan-dev \
|
||||
&& apt autoremove -y \
|
||||
&& apt clean -y \
|
||||
&& rm -rf /tmp/* /var/tmp/* \
|
||||
&& find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
|
||||
&& find /var/cache -type f -delete
|
||||
|
||||
COPY --from=build /app/lib/ /app
|
||||
|
||||
### Full
|
||||
FROM base AS full
|
||||
|
||||
COPY --from=build /app/full /app
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
RUN apt-get update \
|
||||
&& apt-get install -y \
|
||||
git \
|
||||
python3 \
|
||||
python3-pip \
|
||||
python3-wheel \
|
||||
&& pip install --break-system-packages --upgrade setuptools \
|
||||
&& pip install --break-system-packages -r requirements.txt \
|
||||
&& apt autoremove -y \
|
||||
&& apt clean -y \
|
||||
&& rm -rf /tmp/* /var/tmp/* \
|
||||
&& find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
|
||||
&& find /var/cache -type f -delete
|
||||
|
||||
ENTRYPOINT ["/app/tools.sh"]
|
||||
|
||||
### Light, CLI only
|
||||
FROM base AS light
|
||||
|
||||
COPY --from=build /app/full/llama-cli /app
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
ENTRYPOINT [ "/app/llama-cli" ]
|
||||
|
||||
### Server, Server only
|
||||
FROM base AS server
|
||||
|
||||
ENV LLAMA_ARG_HOST=0.0.0.0
|
||||
|
||||
COPY --from=build /app/full/llama-server /app
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ]
|
||||
|
||||
ENTRYPOINT [ "/app/llama-server" ]
|
|
@ -1,7 +1,7 @@
|
|||
*.o
|
||||
*.a
|
||||
.cache/
|
||||
# Do not ignore .git directory, otherwise the reported build number will always be 0
|
||||
.git/
|
||||
.github/
|
||||
.gitignore
|
||||
.vs/
|
||||
|
@ -12,8 +12,8 @@ build*/
|
|||
|
||||
models/*
|
||||
|
||||
/llama-cli
|
||||
/llama-quantize
|
||||
/main
|
||||
/quantize
|
||||
|
||||
arm_neon.h
|
||||
compile_commands.json
|
||||
|
|
1
.ecrc
1
.ecrc
|
@ -1,5 +1,4 @@
|
|||
{
|
||||
"Exclude": ["^\\.gitmodules$", "stb_image\\.h"],
|
||||
"Disable": {
|
||||
"IndentSize": true
|
||||
}
|
||||
|
|
|
@ -24,27 +24,5 @@ insert_final_newline = unset
|
|||
[examples/server/public/*]
|
||||
indent_size = 2
|
||||
|
||||
[examples/server/public/deps_*]
|
||||
trim_trailing_whitespace = unset
|
||||
indent_style = unset
|
||||
indent_size = unset
|
||||
|
||||
[examples/server/deps_*]
|
||||
trim_trailing_whitespace = unset
|
||||
indent_style = unset
|
||||
indent_size = unset
|
||||
|
||||
[examples/llama.swiftui/llama.swiftui.xcodeproj/*]
|
||||
indent_style = tab
|
||||
|
||||
[examples/cvector-generator/*.txt]
|
||||
trim_trailing_whitespace = unset
|
||||
insert_final_newline = unset
|
||||
|
||||
[models/templates/*.jinja]
|
||||
indent_style = unset
|
||||
indent_size = unset
|
||||
end_of_line = unset
|
||||
charset = unset
|
||||
trim_trailing_whitespace = unset
|
||||
insert_final_newline = unset
|
||||
|
|
15
.flake8
15
.flake8
|
@ -1,17 +1,2 @@
|
|||
[flake8]
|
||||
max-line-length = 125
|
||||
ignore = E203,E211,E221,E225,E231,E241,E251,E261,E266,E501,E701,E704,W503
|
||||
exclude =
|
||||
# Do not traverse examples
|
||||
examples,
|
||||
# Do not include package initializers
|
||||
__init__.py,
|
||||
# No need to traverse our git directory
|
||||
.git,
|
||||
# There's no value in checking cache directories
|
||||
__pycache__,
|
||||
# No need to include the build path
|
||||
build,
|
||||
# This contains builds that we don't want to check
|
||||
dist # This is generated with `python build .` for package releases
|
||||
# max-complexity = 10
|
||||
|
|
87
.github/ISSUE_TEMPLATE/010-bug-compilation.yml
vendored
87
.github/ISSUE_TEMPLATE/010-bug-compilation.yml
vendored
|
@ -1,87 +0,0 @@
|
|||
name: Bug (compilation)
|
||||
description: Something goes wrong when trying to compile llama.cpp.
|
||||
title: "Compile bug: "
|
||||
labels: ["bug-unconfirmed", "compilation"]
|
||||
body:
|
||||
- type: markdown
|
||||
attributes:
|
||||
value: >
|
||||
Thanks for taking the time to fill out this bug report!
|
||||
This issue template is intended for bug reports where the compilation of llama.cpp fails.
|
||||
Before opening an issue, please confirm that the compilation still fails with `-DGGML_CCACHE=OFF`.
|
||||
If the compilation succeeds with ccache disabled you should be able to permanently fix the issue
|
||||
by clearing `~/.cache/ccache` (on Linux).
|
||||
- type: textarea
|
||||
id: commit
|
||||
attributes:
|
||||
label: Git commit
|
||||
description: Which commit are you trying to compile?
|
||||
placeholder: |
|
||||
$git rev-parse HEAD
|
||||
84a07a17b1b08cf2b9747c633a2372782848a27f
|
||||
validations:
|
||||
required: true
|
||||
- type: dropdown
|
||||
id: operating-system
|
||||
attributes:
|
||||
label: Operating systems
|
||||
description: Which operating systems do you know to be affected?
|
||||
multiple: true
|
||||
options:
|
||||
- Linux
|
||||
- Mac
|
||||
- Windows
|
||||
- BSD
|
||||
- Other? (Please let us know in description)
|
||||
validations:
|
||||
required: true
|
||||
- type: dropdown
|
||||
id: backends
|
||||
attributes:
|
||||
label: GGML backends
|
||||
description: Which GGML backends do you know to be affected?
|
||||
options: [AMX, BLAS, CPU, CUDA, HIP, Kompute, Metal, Musa, RPC, SYCL, Vulkan]
|
||||
multiple: true
|
||||
validations:
|
||||
required: true
|
||||
- type: textarea
|
||||
id: info
|
||||
attributes:
|
||||
label: Problem description & steps to reproduce
|
||||
description: >
|
||||
Please give us a summary of the problem and tell us how to reproduce it.
|
||||
If you can narrow down the bug to specific compile flags, that information would be very much appreciated by us.
|
||||
placeholder: >
|
||||
I'm trying to compile llama.cpp with CUDA support on a fresh install of Ubuntu and get error XY.
|
||||
Here are the exact commands that I used: ...
|
||||
validations:
|
||||
required: true
|
||||
- type: textarea
|
||||
id: first_bad_commit
|
||||
attributes:
|
||||
label: First Bad Commit
|
||||
description: >
|
||||
If the bug was not present on an earlier version: when did it start appearing?
|
||||
If possible, please do a git bisect and identify the exact commit that introduced the bug.
|
||||
validations:
|
||||
required: false
|
||||
- type: textarea
|
||||
id: command
|
||||
attributes:
|
||||
label: Compile command
|
||||
description: >
|
||||
Please provide the exact command you used to compile llama.cpp. For example: `cmake -B ...`.
|
||||
This will be automatically formatted into code, so no need for backticks.
|
||||
render: shell
|
||||
validations:
|
||||
required: true
|
||||
- type: textarea
|
||||
id: logs
|
||||
attributes:
|
||||
label: Relevant log output
|
||||
description: >
|
||||
Please copy and paste any relevant log output, including any generated text.
|
||||
This will be automatically formatted into code, so no need for backticks.
|
||||
render: shell
|
||||
validations:
|
||||
required: true
|
101
.github/ISSUE_TEMPLATE/011-bug-results.yml
vendored
101
.github/ISSUE_TEMPLATE/011-bug-results.yml
vendored
|
@ -1,101 +0,0 @@
|
|||
name: Bug (model use)
|
||||
description: Something goes wrong when using a model (in general, not specific to a single llama.cpp module).
|
||||
title: "Eval bug: "
|
||||
labels: ["bug-unconfirmed", "model evaluation"]
|
||||
body:
|
||||
- type: markdown
|
||||
attributes:
|
||||
value: >
|
||||
Thanks for taking the time to fill out this bug report!
|
||||
This issue template is intended for bug reports where the model evaluation results
|
||||
(i.e. the generated text) are incorrect or llama.cpp crashes during model evaluation.
|
||||
If you encountered the issue while using an external UI (e.g. ollama),
|
||||
please reproduce your issue using one of the examples/binaries in this repository.
|
||||
The `llama-cli` binary can be used for simple and reproducible model inference.
|
||||
- type: textarea
|
||||
id: version
|
||||
attributes:
|
||||
label: Name and Version
|
||||
description: Which version of our software are you running? (use `--version` to get a version string)
|
||||
placeholder: |
|
||||
$./llama-cli --version
|
||||
version: 2999 (42b4109e)
|
||||
built with cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 for x86_64-linux-gnu
|
||||
validations:
|
||||
required: true
|
||||
- type: dropdown
|
||||
id: operating-system
|
||||
attributes:
|
||||
label: Operating systems
|
||||
description: Which operating systems do you know to be affected?
|
||||
multiple: true
|
||||
options:
|
||||
- Linux
|
||||
- Mac
|
||||
- Windows
|
||||
- BSD
|
||||
- Other? (Please let us know in description)
|
||||
validations:
|
||||
required: true
|
||||
- type: dropdown
|
||||
id: backends
|
||||
attributes:
|
||||
label: GGML backends
|
||||
description: Which GGML backends do you know to be affected?
|
||||
options: [AMX, BLAS, CPU, CUDA, HIP, Kompute, Metal, Musa, RPC, SYCL, Vulkan]
|
||||
multiple: true
|
||||
validations:
|
||||
required: true
|
||||
- type: textarea
|
||||
id: hardware
|
||||
attributes:
|
||||
label: Hardware
|
||||
description: Which CPUs/GPUs are you using?
|
||||
placeholder: >
|
||||
e.g. Ryzen 5950X + 2x RTX 4090
|
||||
validations:
|
||||
required: true
|
||||
- type: textarea
|
||||
id: model
|
||||
attributes:
|
||||
label: Models
|
||||
description: >
|
||||
Which model(s) at which quantization were you using when encountering the bug?
|
||||
If you downloaded a GGUF file off of Huggingface, please provide a link.
|
||||
placeholder: >
|
||||
e.g. Meta LLaMA 3.1 Instruct 8b q4_K_M
|
||||
validations:
|
||||
required: false
|
||||
- type: textarea
|
||||
id: info
|
||||
attributes:
|
||||
label: Problem description & steps to reproduce
|
||||
description: >
|
||||
Please give us a summary of the problem and tell us how to reproduce it.
|
||||
If you can narrow down the bug to specific hardware, compile flags, or command line arguments,
|
||||
that information would be very much appreciated by us.
|
||||
placeholder: >
|
||||
e.g. when I run llama-cli with -ngl 99 I get garbled outputs.
|
||||
When I use -ngl 0 it works correctly.
|
||||
Here are the exact commands that I used: ...
|
||||
validations:
|
||||
required: true
|
||||
- type: textarea
|
||||
id: first_bad_commit
|
||||
attributes:
|
||||
label: First Bad Commit
|
||||
description: >
|
||||
If the bug was not present on an earlier version: when did it start appearing?
|
||||
If possible, please do a git bisect and identify the exact commit that introduced the bug.
|
||||
validations:
|
||||
required: false
|
||||
- type: textarea
|
||||
id: logs
|
||||
attributes:
|
||||
label: Relevant log output
|
||||
description: >
|
||||
Please copy and paste any relevant log output, including the command that you entered and any generated text.
|
||||
This will be automatically formatted into code, so no need for backticks.
|
||||
render: shell
|
||||
validations:
|
||||
required: true
|
91
.github/ISSUE_TEMPLATE/019-bug-misc.yml
vendored
91
.github/ISSUE_TEMPLATE/019-bug-misc.yml
vendored
|
@ -1,91 +0,0 @@
|
|||
name: Bug (misc.)
|
||||
description: Something is not working the way it should (and it's not covered by any of the above cases).
|
||||
title: "Misc. bug: "
|
||||
labels: ["bug-unconfirmed"]
|
||||
body:
|
||||
- type: markdown
|
||||
attributes:
|
||||
value: >
|
||||
Thanks for taking the time to fill out this bug report!
|
||||
This issue template is intended for miscellaneous bugs that don't fit into any other category.
|
||||
If you encountered the issue while using an external UI (e.g. ollama),
|
||||
please reproduce your issue using one of the examples/binaries in this repository.
|
||||
- type: textarea
|
||||
id: version
|
||||
attributes:
|
||||
label: Name and Version
|
||||
description: Which version of our software is affected? (You can use `--version` to get a version string.)
|
||||
placeholder: |
|
||||
$./llama-cli --version
|
||||
version: 2999 (42b4109e)
|
||||
built with cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 for x86_64-linux-gnu
|
||||
validations:
|
||||
required: true
|
||||
- type: dropdown
|
||||
id: operating-system
|
||||
attributes:
|
||||
label: Operating systems
|
||||
description: Which operating systems do you know to be affected?
|
||||
multiple: true
|
||||
options:
|
||||
- Linux
|
||||
- Mac
|
||||
- Windows
|
||||
- BSD
|
||||
- Other? (Please let us know in description)
|
||||
validations:
|
||||
required: false
|
||||
- type: dropdown
|
||||
id: module
|
||||
attributes:
|
||||
label: Which llama.cpp modules do you know to be affected?
|
||||
multiple: true
|
||||
options:
|
||||
- Documentation/Github
|
||||
- libllama (core library)
|
||||
- llama-cli
|
||||
- llama-server
|
||||
- llama-bench
|
||||
- llama-quantize
|
||||
- Python/Bash scripts
|
||||
- Test code
|
||||
- Other (Please specify in the next section)
|
||||
validations:
|
||||
required: false
|
||||
- type: textarea
|
||||
id: command
|
||||
attributes:
|
||||
label: Command line
|
||||
description: >
|
||||
Please provide the exact commands you entered, if applicable. For example: `llama-server -m ... -c ...`, `llama-cli -m ...`, etc.
|
||||
This will be automatically formatted into code, so no need for backticks.
|
||||
render: shell
|
||||
validations:
|
||||
required: false
|
||||
- type: textarea
|
||||
id: info
|
||||
attributes:
|
||||
label: Problem description & steps to reproduce
|
||||
description: >
|
||||
Please give us a summary of the problem and tell us how to reproduce it (if applicable).
|
||||
validations:
|
||||
required: true
|
||||
- type: textarea
|
||||
id: first_bad_commit
|
||||
attributes:
|
||||
label: First Bad Commit
|
||||
description: >
|
||||
If the bug was not present on an earlier version and it's not trivial to track down: when did it start appearing?
|
||||
If possible, please do a git bisect and identify the exact commit that introduced the bug.
|
||||
validations:
|
||||
required: false
|
||||
- type: textarea
|
||||
id: logs
|
||||
attributes:
|
||||
label: Relevant log output
|
||||
description: >
|
||||
If applicable, please copy and paste any relevant log output, including any generated text.
|
||||
This will be automatically formatted into code, so no need for backticks.
|
||||
render: shell
|
||||
validations:
|
||||
required: false
|
51
.github/ISSUE_TEMPLATE/020-enhancement.yml
vendored
51
.github/ISSUE_TEMPLATE/020-enhancement.yml
vendored
|
@ -1,51 +0,0 @@
|
|||
name: Enhancement
|
||||
description: Used to request enhancements for llama.cpp.
|
||||
title: "Feature Request: "
|
||||
labels: ["enhancement"]
|
||||
body:
|
||||
- type: markdown
|
||||
attributes:
|
||||
value: |
|
||||
[Please post your idea first in Discussion if there is not yet a consensus for this enhancement request. This will help to keep this issue tracker focused on enhancements that the community has agreed needs to be implemented.](https://github.com/ggerganov/llama.cpp/discussions/categories/ideas)
|
||||
|
||||
- type: checkboxes
|
||||
id: prerequisites
|
||||
attributes:
|
||||
label: Prerequisites
|
||||
description: Please confirm the following before submitting your enhancement request.
|
||||
options:
|
||||
- label: I am running the latest code. Mention the version if possible as well.
|
||||
required: true
|
||||
- label: I carefully followed the [README.md](https://github.com/ggerganov/llama.cpp/blob/master/README.md).
|
||||
required: true
|
||||
- label: I searched using keywords relevant to my issue to make sure that I am creating a new issue that is not already open (or closed).
|
||||
required: true
|
||||
- label: I reviewed the [Discussions](https://github.com/ggerganov/llama.cpp/discussions), and have a new and useful enhancement to share.
|
||||
required: true
|
||||
|
||||
- type: textarea
|
||||
id: feature-description
|
||||
attributes:
|
||||
label: Feature Description
|
||||
description: Please provide a detailed written description of what you were trying to do, and what you expected `llama.cpp` to do as an enhancement.
|
||||
placeholder: Detailed description of the enhancement
|
||||
validations:
|
||||
required: true
|
||||
|
||||
- type: textarea
|
||||
id: motivation
|
||||
attributes:
|
||||
label: Motivation
|
||||
description: Please provide a detailed written description of reasons why this feature is necessary and how it is useful to `llama.cpp` users.
|
||||
placeholder: Explanation of why this feature is needed and its benefits
|
||||
validations:
|
||||
required: true
|
||||
|
||||
- type: textarea
|
||||
id: possible-implementation
|
||||
attributes:
|
||||
label: Possible Implementation
|
||||
description: If you have an idea as to how it can be implemented, please write a detailed description. Feel free to give links to external sources or share visuals that might be helpful to understand the details better.
|
||||
placeholder: Detailed description of potential implementation
|
||||
validations:
|
||||
required: false
|
52
.github/ISSUE_TEMPLATE/030-research.yml
vendored
52
.github/ISSUE_TEMPLATE/030-research.yml
vendored
|
@ -1,52 +0,0 @@
|
|||
name: Research
|
||||
description: Track new technical research area.
|
||||
title: "Research: "
|
||||
labels: ["research 🔬"]
|
||||
body:
|
||||
- type: markdown
|
||||
attributes:
|
||||
value: |
|
||||
Don't forget to check for any [duplicate research issue tickets](https://github.com/ggerganov/llama.cpp/issues?q=is%3Aopen+is%3Aissue+label%3A%22research+%F0%9F%94%AC%22)
|
||||
|
||||
- type: checkboxes
|
||||
id: research-stage
|
||||
attributes:
|
||||
label: Research Stage
|
||||
description: Track general state of this research ticket
|
||||
options:
|
||||
- label: Background Research (Let's try to avoid reinventing the wheel)
|
||||
- label: Hypothesis Formed (How do you think this will work and it's effect?)
|
||||
- label: Strategy / Implementation Forming
|
||||
- label: Analysis of results
|
||||
- label: Debrief / Documentation (So people in the future can learn from us)
|
||||
|
||||
- type: textarea
|
||||
id: background
|
||||
attributes:
|
||||
label: Previous existing literature and research
|
||||
description: Whats the current state of the art and whats the motivation for this research?
|
||||
|
||||
- type: textarea
|
||||
id: hypothesis
|
||||
attributes:
|
||||
label: Hypothesis
|
||||
description: How do you think this will work and it's effect?
|
||||
|
||||
- type: textarea
|
||||
id: implementation
|
||||
attributes:
|
||||
label: Implementation
|
||||
description: Got an approach? e.g. a PR ready to go?
|
||||
|
||||
- type: textarea
|
||||
id: analysis
|
||||
attributes:
|
||||
label: Analysis
|
||||
description: How does the proposed implementation behave?
|
||||
|
||||
- type: textarea
|
||||
id: logs
|
||||
attributes:
|
||||
label: Relevant log output
|
||||
description: Please copy and paste any relevant log output. This will be automatically formatted into code, so no need for backticks.
|
||||
render: shell
|
28
.github/ISSUE_TEMPLATE/040-refactor.yml
vendored
28
.github/ISSUE_TEMPLATE/040-refactor.yml
vendored
|
@ -1,28 +0,0 @@
|
|||
name: Refactor (Maintainers)
|
||||
description: Used to track refactoring opportunities.
|
||||
title: "Refactor: "
|
||||
labels: ["refactor"]
|
||||
body:
|
||||
- type: markdown
|
||||
attributes:
|
||||
value: |
|
||||
Don't forget to [check for existing refactor issue tickets](https://github.com/ggerganov/llama.cpp/issues?q=is%3Aopen+is%3Aissue+label%3Arefactoring) in case it's already covered.
|
||||
Also you may want to check [Pull request refactor label as well](https://github.com/ggerganov/llama.cpp/pulls?q=is%3Aopen+is%3Apr+label%3Arefactoring) for duplicates too.
|
||||
|
||||
- type: textarea
|
||||
id: background-description
|
||||
attributes:
|
||||
label: Background Description
|
||||
description: Please provide a detailed written description of the pain points you are trying to solve.
|
||||
placeholder: Detailed description behind your motivation to request refactor
|
||||
validations:
|
||||
required: true
|
||||
|
||||
- type: textarea
|
||||
id: possible-approaches
|
||||
attributes:
|
||||
label: Possible Refactor Approaches
|
||||
description: If you have some idea of possible approaches to solve this problem. You may want to make it a todo list.
|
||||
placeholder: Your idea of possible refactoring opportunity/approaches
|
||||
validations:
|
||||
required: false
|
184
.github/ISSUE_TEMPLATE/bug.md
vendored
Normal file
184
.github/ISSUE_TEMPLATE/bug.md
vendored
Normal file
|
@ -0,0 +1,184 @@
|
|||
---
|
||||
name: Bug template
|
||||
about: Used to report bugs in llama.cpp
|
||||
labels: ["bug-unconfirmed"]
|
||||
assignees: ''
|
||||
|
||||
---
|
||||
|
||||
# Prerequisites
|
||||
|
||||
Please answer the following questions for yourself before submitting an issue.
|
||||
|
||||
- [ ] I am running the latest code. Development is very rapid so there are no tagged versions as of now.
|
||||
- [ ] I carefully followed the [README.md](https://github.com/ggerganov/llama.cpp/blob/master/README.md).
|
||||
- [ ] I [searched using keywords relevant to my issue](https://docs.github.com/en/issues/tracking-your-work-with-issues/filtering-and-searching-issues-and-pull-requests) to make sure that I am creating a new issue that is not already open (or closed).
|
||||
- [ ] I reviewed the [Discussions](https://github.com/ggerganov/llama.cpp/discussions), and have a new bug or useful enhancement to share.
|
||||
|
||||
# Expected Behavior
|
||||
|
||||
Please provide a detailed written description of what you were trying to do, and what you expected `llama.cpp` to do.
|
||||
|
||||
# Current Behavior
|
||||
|
||||
Please provide a detailed written description of what `llama.cpp` did, instead.
|
||||
|
||||
# Environment and Context
|
||||
|
||||
Please provide detailed information about your computer setup. This is important in case the issue is not reproducible except for under certain specific conditions.
|
||||
|
||||
* Physical (or virtual) hardware you are using, e.g. for Linux:
|
||||
|
||||
`$ lscpu`
|
||||
|
||||
* Operating System, e.g. for Linux:
|
||||
|
||||
`$ uname -a`
|
||||
|
||||
* SDK version, e.g. for Linux:
|
||||
|
||||
```
|
||||
$ python3 --version
|
||||
$ make --version
|
||||
$ g++ --version
|
||||
```
|
||||
|
||||
# Failure Information (for bugs)
|
||||
|
||||
Please help provide information about the failure / bug.
|
||||
|
||||
# Steps to Reproduce
|
||||
|
||||
Please provide detailed steps for reproducing the issue. We are not sitting in front of your screen, so the more detail the better.
|
||||
|
||||
1. step 1
|
||||
2. step 2
|
||||
3. step 3
|
||||
4. etc.
|
||||
|
||||
# Failure Logs
|
||||
|
||||
Please include any relevant log snippets or files. If it works under one configuration but not under another, please provide logs for both configurations and their corresponding outputs so it is easy to see where behavior changes.
|
||||
|
||||
Also, please try to **avoid using screenshots** if at all possible. Instead, copy/paste the console output and use [Github's markdown](https://docs.github.com/en/get-started/writing-on-github/getting-started-with-writing-and-formatting-on-github/basic-writing-and-formatting-syntax) to cleanly format your logs for easy readability.
|
||||
|
||||
Example environment info:
|
||||
```
|
||||
llama.cpp$ git log | head -1
|
||||
commit 2af23d30434a677c6416812eea52ccc0af65119c
|
||||
|
||||
llama.cpp$ lscpu | egrep "AMD|Flags"
|
||||
Vendor ID: AuthenticAMD
|
||||
Model name: AMD Ryzen Threadripper 1950X 16-Core Processor
|
||||
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc cpuid extd_apicid amd_dcm aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb hw_pstate ssbd ibpb vmmcall fsgsbase bmi1 avx2 smep bmi2 rdseed adx smap clflushopt sha_ni xsaveopt xsavec xgetbv1 xsaves clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif overflow_recov succor smca sme sev
|
||||
Virtualization: AMD-V
|
||||
|
||||
llama.cpp$ python3 --version
|
||||
Python 3.10.9
|
||||
|
||||
llama.cpp$ pip list | egrep "torch|numpy|sentencepiece"
|
||||
numpy 1.24.2
|
||||
numpydoc 1.5.0
|
||||
sentencepiece 0.1.97
|
||||
torch 1.13.1
|
||||
torchvision 0.14.1
|
||||
|
||||
llama.cpp$ make --version | head -1
|
||||
GNU Make 4.3
|
||||
|
||||
$ md5sum ./models/65B/ggml-model-q4_0.bin
|
||||
dbdd682cce80e2d6e93cefc7449df487 ./models/65B/ggml-model-q4_0.bin
|
||||
```
|
||||
|
||||
Example run with the Linux command [perf](https://www.brendangregg.com/perf.html)
|
||||
```
|
||||
llama.cpp$ perf stat ./main -m ./models/65B/ggml-model-q4_0.bin -t 16 -n 1024 -p "Please close your issue when it has been answered."
|
||||
main: seed = 1679149377
|
||||
llama_model_load: loading model from './models/65B/ggml-model-q4_0.bin' - please wait ...
|
||||
llama_model_load: n_vocab = 32000
|
||||
llama_model_load: n_ctx = 512
|
||||
llama_model_load: n_embd = 8192
|
||||
llama_model_load: n_mult = 256
|
||||
llama_model_load: n_head = 64
|
||||
llama_model_load: n_layer = 80
|
||||
llama_model_load: n_rot = 128
|
||||
llama_model_load: f16 = 2
|
||||
llama_model_load: n_ff = 22016
|
||||
llama_model_load: n_parts = 8
|
||||
llama_model_load: ggml ctx size = 41477.73 MB
|
||||
llama_model_load: memory_size = 2560.00 MB, n_mem = 40960
|
||||
llama_model_load: loading model part 1/8 from './models/65B/ggml-model-q4_0.bin'
|
||||
llama_model_load: .......................................................................................... done
|
||||
llama_model_load: model size = 4869.09 MB / num tensors = 723
|
||||
llama_model_load: loading model part 2/8 from './models/65B/ggml-model-q4_0.bin.1'
|
||||
llama_model_load: .......................................................................................... done
|
||||
llama_model_load: model size = 4869.09 MB / num tensors = 723
|
||||
llama_model_load: loading model part 3/8 from './models/65B/ggml-model-q4_0.bin.2'
|
||||
llama_model_load: .......................................................................................... done
|
||||
llama_model_load: model size = 4869.09 MB / num tensors = 723
|
||||
llama_model_load: loading model part 4/8 from './models/65B/ggml-model-q4_0.bin.3'
|
||||
llama_model_load: .......................................................................................... done
|
||||
llama_model_load: model size = 4869.09 MB / num tensors = 723
|
||||
llama_model_load: loading model part 5/8 from './models/65B/ggml-model-q4_0.bin.4'
|
||||
llama_model_load: .......................................................................................... done
|
||||
llama_model_load: model size = 4869.09 MB / num tensors = 723
|
||||
llama_model_load: loading model part 6/8 from './models/65B/ggml-model-q4_0.bin.5'
|
||||
llama_model_load: .......................................................................................... done
|
||||
llama_model_load: model size = 4869.09 MB / num tensors = 723
|
||||
llama_model_load: loading model part 7/8 from './models/65B/ggml-model-q4_0.bin.6'
|
||||
llama_model_load: .......................................................................................... done
|
||||
llama_model_load: model size = 4869.09 MB / num tensors = 723
|
||||
llama_model_load: loading model part 8/8 from './models/65B/ggml-model-q4_0.bin.7'
|
||||
llama_model_load: .......................................................................................... done
|
||||
llama_model_load: model size = 4869.09 MB / num tensors = 723
|
||||
|
||||
system_info: n_threads = 16 / 32 | AVX = 1 | AVX2 = 1 | AVX512 = 0 | FMA = 1 | NEON = 0 | ARM_FMA = 0 | F16C = 1 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 0 | SSE3 = 1 | VSX = 0 |
|
||||
|
||||
main: prompt: 'Please close your issue when it has been answered.'
|
||||
main: number of tokens in prompt = 11
|
||||
1 -> ''
|
||||
12148 -> 'Please'
|
||||
3802 -> ' close'
|
||||
596 -> ' your'
|
||||
2228 -> ' issue'
|
||||
746 -> ' when'
|
||||
372 -> ' it'
|
||||
756 -> ' has'
|
||||
1063 -> ' been'
|
||||
7699 -> ' answered'
|
||||
29889 -> '.'
|
||||
|
||||
sampling parameters: temp = 0.800000, top_k = 40, top_p = 0.950000, repeat_last_n = 64, repeat_penalty = 1.300000
|
||||
|
||||
|
||||
Please close your issue when it has been answered.
|
||||
@duncan-donut: I'm trying to figure out what kind of "support" you need for this script and why, exactly? Is there a question about how the code works that hasn't already been addressed in one or more comments below this ticket, or are we talking something else entirely like some sorta bugfixing job because your server setup is different from mine??
|
||||
I can understand if your site needs to be running smoothly and you need help with a fix of sorts but there should really be nothing wrong here that the code itself could not handle. And given that I'm getting reports about how it works perfectly well on some other servers, what exactly are we talking? A detailed report will do wonders in helping us get this resolved for ya quickly so please take your time and describe the issue(s) you see as clearly & concisely as possible!!
|
||||
@duncan-donut: I'm not sure if you have access to cPanel but you could try these instructions. It is worth a shot! Let me know how it goes (or what error message, exactly!) when/if ya give that code a go? [end of text]
|
||||
|
||||
|
||||
main: mem per token = 71159620 bytes
|
||||
main: load time = 19309.95 ms
|
||||
main: sample time = 168.62 ms
|
||||
main: predict time = 223895.61 ms / 888.47 ms per token
|
||||
main: total time = 246406.42 ms
|
||||
|
||||
Performance counter stats for './main -m ./models/65B/ggml-model-q4_0.bin -t 16 -n 1024 -p Please close your issue when it has been answered.':
|
||||
|
||||
3636882.89 msec task-clock # 14.677 CPUs utilized
|
||||
13509 context-switches # 3.714 /sec
|
||||
2436 cpu-migrations # 0.670 /sec
|
||||
10476679 page-faults # 2.881 K/sec
|
||||
13133115082869 cycles # 3.611 GHz (16.77%)
|
||||
29314462753 stalled-cycles-frontend # 0.22% frontend cycles idle (16.76%)
|
||||
10294402631459 stalled-cycles-backend # 78.39% backend cycles idle (16.74%)
|
||||
23479217109614 instructions # 1.79 insn per cycle
|
||||
# 0.44 stalled cycles per insn (16.76%)
|
||||
2353072268027 branches # 647.002 M/sec (16.77%)
|
||||
1998682780 branch-misses # 0.08% of all branches (16.76%)
|
||||
|
||||
247.802177522 seconds time elapsed
|
||||
|
||||
3618.573072000 seconds user
|
||||
18.491698000 seconds sys
|
||||
```
|
11
.github/ISSUE_TEMPLATE/config.yml
vendored
11
.github/ISSUE_TEMPLATE/config.yml
vendored
|
@ -1,11 +0,0 @@
|
|||
blank_issues_enabled: true
|
||||
contact_links:
|
||||
- name: Got an idea?
|
||||
url: https://github.com/ggerganov/llama.cpp/discussions/categories/ideas
|
||||
about: Pop it there. It may then become an enhancement ticket.
|
||||
- name: Got a question?
|
||||
url: https://github.com/ggerganov/llama.cpp/discussions/categories/q-a
|
||||
about: Ask a question there!
|
||||
- name: Want to contribute?
|
||||
url: https://github.com/ggerganov/llama.cpp/wiki/contribute
|
||||
about: Head to the contribution guide page of the wiki for areas you can help with
|
28
.github/ISSUE_TEMPLATE/enhancement.md
vendored
Normal file
28
.github/ISSUE_TEMPLATE/enhancement.md
vendored
Normal file
|
@ -0,0 +1,28 @@
|
|||
---
|
||||
name: Enhancement template
|
||||
about: Used to request enhancements for llama.cpp
|
||||
labels: ["enhancement"]
|
||||
assignees: ''
|
||||
|
||||
---
|
||||
|
||||
# Prerequisites
|
||||
|
||||
Please answer the following questions for yourself before submitting an issue.
|
||||
|
||||
- [ ] I am running the latest code. Development is very rapid so there are no tagged versions as of now.
|
||||
- [ ] I carefully followed the [README.md](https://github.com/ggerganov/llama.cpp/blob/master/README.md).
|
||||
- [ ] I [searched using keywords relevant to my issue](https://docs.github.com/en/issues/tracking-your-work-with-issues/filtering-and-searching-issues-and-pull-requests) to make sure that I am creating a new issue that is not already open (or closed).
|
||||
- [ ] I reviewed the [Discussions](https://github.com/ggerganov/llama.cpp/discussions), and have a new bug or useful enhancement to share.
|
||||
|
||||
# Feature Description
|
||||
|
||||
Please provide a detailed written description of what you were trying to do, and what you expected `llama.cpp` to do as an enhancement.
|
||||
|
||||
# Motivation
|
||||
|
||||
Please provide a detailed written description of reasons why this feature is necessary and how it is useful to `llama.cpp` users.
|
||||
|
||||
# Possible Implementation
|
||||
|
||||
If you have an idea as to how it can be implemented, please write a detailed description. Feel free to give links to external sources or share visuals that might be helpful to understand the details better.
|
86
.github/labeler.yml
vendored
86
.github/labeler.yml
vendored
|
@ -1,86 +0,0 @@
|
|||
# https://github.com/actions/labeler
|
||||
Kompute:
|
||||
- changed-files:
|
||||
- any-glob-to-any-file:
|
||||
- ggml/include/ggml-kompute.h
|
||||
- ggml/src/ggml-kompute/**
|
||||
- README-kompute.md
|
||||
Apple Metal:
|
||||
- changed-files:
|
||||
- any-glob-to-any-file:
|
||||
- ggml/include/ggml-metal.h
|
||||
- ggml/src/ggml-metal/**
|
||||
- README-metal.md
|
||||
SYCL:
|
||||
- changed-files:
|
||||
- any-glob-to-any-file:
|
||||
- ggml/include/ggml-sycl.h
|
||||
- ggml/src/ggml-sycl/**
|
||||
- docs/backend/SYCL.md
|
||||
- examples/sycl/**
|
||||
Nvidia GPU:
|
||||
- changed-files:
|
||||
- any-glob-to-any-file:
|
||||
- ggml/include/ggml-cuda.h
|
||||
- ggml/src/ggml-cuda/**
|
||||
Vulkan:
|
||||
- changed-files:
|
||||
- any-glob-to-any-file:
|
||||
- ggml/include/ggml-vulkan.h
|
||||
- ggml/src/ggml-vulkan/**
|
||||
documentation:
|
||||
- changed-files:
|
||||
- any-glob-to-any-file:
|
||||
- docs/**
|
||||
- media/**
|
||||
testing:
|
||||
- changed-files:
|
||||
- any-glob-to-any-file:
|
||||
- tests/**
|
||||
build:
|
||||
- changed-files:
|
||||
- any-glob-to-any-file:
|
||||
- cmake/**
|
||||
- CMakeLists.txt
|
||||
- CMakePresets.json
|
||||
examples:
|
||||
- changed-files:
|
||||
- any-glob-to-any-file: examples/**
|
||||
devops:
|
||||
- changed-files:
|
||||
- any-glob-to-any-file:
|
||||
- .devops/**
|
||||
- .github/**
|
||||
- ci/**
|
||||
python:
|
||||
- changed-files:
|
||||
- any-glob-to-any-file:
|
||||
- "**/*.py"
|
||||
- requirements/**
|
||||
- gguf-py/**
|
||||
- .flake8
|
||||
script:
|
||||
- changed-files:
|
||||
- any-glob-to-any-file:
|
||||
- scripts/**
|
||||
android:
|
||||
- changed-files:
|
||||
- any-glob-to-any-file:
|
||||
- examples/llama.android/**
|
||||
server:
|
||||
- changed-files:
|
||||
- any-glob-to-any-file:
|
||||
- examples/server/**
|
||||
ggml:
|
||||
- changed-files:
|
||||
- any-glob-to-any-file:
|
||||
- ggml/**
|
||||
nix:
|
||||
- changed-files:
|
||||
- any-glob-to-any-file:
|
||||
- "**/*.nix"
|
||||
- .github/workflows/nix-*.yml
|
||||
- .devops/nix/nixpkgs-instances.nix
|
||||
embedding:
|
||||
- changed-files:
|
||||
- any-glob-to-any-file: examples/embedding/
|
1
.github/pull_request_template.md
vendored
1
.github/pull_request_template.md
vendored
|
@ -1 +0,0 @@
|
|||
*Make sure to read the [contributing guidelines](https://github.com/ggerganov/llama.cpp/blob/master/CONTRIBUTING.md) before submitting a PR*
|
315
.github/workflows/bench.yml.disabled
vendored
315
.github/workflows/bench.yml.disabled
vendored
|
@ -1,315 +0,0 @@
|
|||
# TODO: there have been some issues with the workflow, so disabling for now
|
||||
# https://github.com/ggerganov/llama.cpp/issues/7893
|
||||
#
|
||||
# Benchmark
|
||||
name: Benchmark
|
||||
|
||||
on:
|
||||
workflow_dispatch:
|
||||
inputs:
|
||||
gpu-series:
|
||||
description: 'Azure GPU series to run with'
|
||||
required: true
|
||||
type: choice
|
||||
options:
|
||||
- Standard_NC4as_T4_v3
|
||||
- Standard_NC24ads_A100_v4
|
||||
- Standard_NC80adis_H100_v5
|
||||
sha:
|
||||
description: 'Commit SHA1 to build'
|
||||
required: false
|
||||
type: string
|
||||
duration:
|
||||
description: 'Duration of the bench'
|
||||
type: string
|
||||
default: 10m
|
||||
|
||||
push:
|
||||
branches:
|
||||
- master
|
||||
paths: ['llama.cpp', 'ggml.c', 'ggml-backend.cpp', 'ggml-quants.c', '**/*.cu', 'examples/server/*.h*', 'examples/server/*.cpp']
|
||||
pull_request_target:
|
||||
types: [opened, synchronize, reopened]
|
||||
paths: ['llama.cpp', 'ggml.c', 'ggml-backend.cpp', 'ggml-quants.c', '**/*.cu', 'examples/server/*.h*', 'examples/server/*.cpp']
|
||||
schedule:
|
||||
- cron: '04 2 * * *'
|
||||
|
||||
concurrency:
|
||||
group: ${{ github.workflow }}-${{ github.ref }}-${{ github.head_ref || github.run_id }}-${{ github.event.inputs.sha }}
|
||||
cancel-in-progress: true
|
||||
|
||||
jobs:
|
||||
bench-server-baseline:
|
||||
runs-on: Standard_NC4as_T4_v3
|
||||
env:
|
||||
RUNNER_LABEL: Standard_NC4as_T4_v3 # FIXME Do not find a way to not duplicate it
|
||||
N_USERS: 8
|
||||
DURATION: 10m
|
||||
|
||||
strategy:
|
||||
matrix:
|
||||
model: [phi-2]
|
||||
ftype: [q4_0, q8_0, f16]
|
||||
include:
|
||||
- model: phi-2
|
||||
ftype: q4_0
|
||||
pr_comment_enabled: "true"
|
||||
|
||||
if: |
|
||||
inputs.gpu-series == 'Standard_NC4as_T4_v3'
|
||||
|| (
|
||||
github.event_name == 'schedule'
|
||||
&& github.ref_name == 'master'
|
||||
&& github.repository_owner == 'ggerganov'
|
||||
)
|
||||
|| github.event_name == 'pull_request_target'
|
||||
|| (
|
||||
github.event_name == 'push'
|
||||
&& github.event.ref == 'refs/heads/master'
|
||||
&& github.repository_owner == 'ggerganov'
|
||||
)
|
||||
steps:
|
||||
- name: Clone
|
||||
id: checkout
|
||||
uses: actions/checkout@v4
|
||||
with:
|
||||
fetch-depth: 0
|
||||
ref: ${{ github.event.inputs.sha || github.event.pull_request.head.sha || github.sha || github.head_ref || github.ref_name }}
|
||||
|
||||
- name: Install python env
|
||||
id: pipenv
|
||||
run: |
|
||||
cd examples/server/bench
|
||||
python3 -m venv venv
|
||||
source venv/bin/activate
|
||||
pip install -r requirements.txt
|
||||
|
||||
- name: Prometheus
|
||||
id: install_prometheus
|
||||
run: |
|
||||
wget --quiet https://github.com/prometheus/prometheus/releases/download/v2.51.0/prometheus-2.51.0.linux-amd64.tar.gz
|
||||
tar xzf prometheus*.tar.gz --strip-components=1
|
||||
./prometheus --config.file=examples/server/bench/prometheus.yml &
|
||||
while ! nc -z localhost 9090; do
|
||||
sleep 0.1
|
||||
done
|
||||
|
||||
- name: Set up Go
|
||||
uses: actions/setup-go@v5
|
||||
with:
|
||||
go-version: '1.21'
|
||||
|
||||
- name: Install k6 and xk6-sse
|
||||
id: k6_installation
|
||||
run: |
|
||||
cd examples/server/bench
|
||||
go install go.k6.io/xk6/cmd/xk6@latest
|
||||
xk6 build master \
|
||||
--with github.com/phymbert/xk6-sse
|
||||
|
||||
- name: Build
|
||||
id: cmake_build
|
||||
run: |
|
||||
set -eux
|
||||
cmake -B build \
|
||||
-DGGML_NATIVE=OFF \
|
||||
-DLLAMA_BUILD_SERVER=ON \
|
||||
-DLLAMA_CURL=ON \
|
||||
-DLLAMA_CUBLAS=ON \
|
||||
-DCUDAToolkit_ROOT=/usr/local/cuda \
|
||||
-DCMAKE_CUDA_COMPILER=/usr/local/cuda/bin/nvcc \
|
||||
-DCMAKE_CUDA_ARCHITECTURES=75 \
|
||||
-DLLAMA_FATAL_WARNINGS=OFF \
|
||||
-DLLAMA_ALL_WARNINGS=OFF \
|
||||
-DCMAKE_BUILD_TYPE=Release;
|
||||
cmake --build build --config Release -j $(nproc) --target llama-server
|
||||
|
||||
- name: Download the dataset
|
||||
id: download_dataset
|
||||
run: |
|
||||
cd examples/server/bench
|
||||
wget --quiet https://huggingface.co/datasets/anon8231489123/ShareGPT_Vicuna_unfiltered/resolve/main/ShareGPT_V3_unfiltered_cleaned_split.json
|
||||
|
||||
- name: Server bench
|
||||
id: server_bench
|
||||
env:
|
||||
HEAD_REF: ${{ github.head_ref || github.ref_name }}
|
||||
run: |
|
||||
set -eux
|
||||
|
||||
cd examples/server/bench
|
||||
source venv/bin/activate
|
||||
python bench.py \
|
||||
--runner-label ${{ env.RUNNER_LABEL }} \
|
||||
--name ${{ github.job }} \
|
||||
--branch $HEAD_REF \
|
||||
--commit ${{ github.event.inputs.sha || github.event.pull_request.head.sha || github.sha }} \
|
||||
--scenario script.js \
|
||||
--duration ${{ github.event.inputs.duration || env.DURATION }} \
|
||||
--hf-repo ggml-org/models \
|
||||
--hf-file ${{ matrix.model }}/ggml-model-${{ matrix.ftype }}.gguf \
|
||||
--model-path-prefix /models \
|
||||
--parallel ${{ env.N_USERS }} \
|
||||
-ngl 33 \
|
||||
--batch-size 2048 \
|
||||
--ubatch-size 256 \
|
||||
--ctx-size 16384 \
|
||||
--n-prompts 1000 \
|
||||
--max-prompt-tokens 1024 \
|
||||
--max-tokens 2048
|
||||
|
||||
cat results.github.env >> $GITHUB_ENV
|
||||
|
||||
# Remove dataset as we do not want it in the artefact
|
||||
rm ShareGPT_V3_unfiltered_cleaned_split.json
|
||||
|
||||
- uses: actions/upload-artifact@v4
|
||||
with:
|
||||
name: bench-server-${{ github.job }}-${{ env.RUNNER_LABEL }}-${{ matrix.model }}-${{ matrix.ftype }}
|
||||
compression-level: 9
|
||||
path: |
|
||||
examples/server/bench/*.jpg
|
||||
examples/server/bench/*.json
|
||||
examples/server/bench/*.log
|
||||
|
||||
- name: Commit status
|
||||
uses: Sibz/github-status-action@v1
|
||||
with:
|
||||
authToken: ${{secrets.GITHUB_TOKEN}}
|
||||
sha: ${{ inputs.sha || github.event.pull_request.head.sha || github.sha }}
|
||||
context: bench-server-${{ github.job }}-${{ env.RUNNER_LABEL }}-${{ matrix.model }}-${{ matrix.ftype }}
|
||||
description: |
|
||||
${{ env.BENCH_RESULTS }}
|
||||
state: 'success'
|
||||
|
||||
- name: Upload benchmark images
|
||||
uses: devicons/public-upload-to-imgur@v2.2.2
|
||||
continue-on-error: true # Important as it looks unstable: 503
|
||||
id: imgur_step
|
||||
with:
|
||||
client_id: ${{secrets.IMGUR_CLIENT_ID}}
|
||||
path: |
|
||||
examples/server/bench/prompt_tokens_seconds.jpg
|
||||
examples/server/bench/predicted_tokens_seconds.jpg
|
||||
examples/server/bench/kv_cache_usage_ratio.jpg
|
||||
examples/server/bench/requests_processing.jpg
|
||||
|
||||
- name: Extract mermaid
|
||||
id: set_mermaid
|
||||
run: |
|
||||
set -eux
|
||||
|
||||
cd examples/server/bench
|
||||
PROMPT_TOKENS_SECONDS=$(cat prompt_tokens_seconds.mermaid)
|
||||
echo "PROMPT_TOKENS_SECONDS<<EOF" >> $GITHUB_ENV
|
||||
echo "$PROMPT_TOKENS_SECONDS" >> $GITHUB_ENV
|
||||
echo "EOF" >> $GITHUB_ENV
|
||||
|
||||
PREDICTED_TOKENS_SECONDS=$(cat predicted_tokens_seconds.mermaid)
|
||||
echo "PREDICTED_TOKENS_SECONDS<<EOF" >> $GITHUB_ENV
|
||||
echo "$PREDICTED_TOKENS_SECONDS" >> $GITHUB_ENV
|
||||
echo "EOF" >> $GITHUB_ENV
|
||||
|
||||
KV_CACHE_USAGE_RATIO=$(cat kv_cache_usage_ratio.mermaid)
|
||||
echo "KV_CACHE_USAGE_RATIO<<EOF" >> $GITHUB_ENV
|
||||
echo "$KV_CACHE_USAGE_RATIO" >> $GITHUB_ENV
|
||||
echo "EOF" >> $GITHUB_ENV
|
||||
|
||||
REQUESTS_PROCESSING=$(cat requests_processing.mermaid)
|
||||
echo "REQUESTS_PROCESSING<<EOF" >> $GITHUB_ENV
|
||||
echo "$REQUESTS_PROCESSING" >> $GITHUB_ENV
|
||||
echo "EOF" >> $GITHUB_ENV
|
||||
|
||||
- name: Extract image url
|
||||
id: extract_image_url
|
||||
continue-on-error: true
|
||||
run: |
|
||||
set -eux
|
||||
|
||||
echo "IMAGE_O=${{ fromJSON(steps.imgur_step.outputs.imgur_urls)[0] }}" >> $GITHUB_ENV
|
||||
echo "IMAGE_1=${{ fromJSON(steps.imgur_step.outputs.imgur_urls)[1] }}" >> $GITHUB_ENV
|
||||
echo "IMAGE_2=${{ fromJSON(steps.imgur_step.outputs.imgur_urls)[2] }}" >> $GITHUB_ENV
|
||||
echo "IMAGE_3=${{ fromJSON(steps.imgur_step.outputs.imgur_urls)[3] }}" >> $GITHUB_ENV
|
||||
|
||||
- name: Comment PR
|
||||
uses: mshick/add-pr-comment@v2
|
||||
id: comment_pr
|
||||
if: ${{ github.event.pull_request != '' && matrix.pr_comment_enabled == 'true' }}
|
||||
with:
|
||||
message-id: bench-server-${{ github.job }}-${{ env.RUNNER_LABEL }}-${{ matrix.model }}-${{ matrix.ftype }}
|
||||
message: |
|
||||
<p align="center">
|
||||
|
||||
📈 **llama.cpp server** for _${{ github.job }}_ on _${{ env.RUNNER_LABEL }}_ for `${{ matrix.model }}`-`${{ matrix.ftype }}`: **${{ env.BENCH_ITERATIONS}} iterations** 🚀
|
||||
|
||||
</p>
|
||||
|
||||
<details>
|
||||
|
||||
<summary>Expand details for performance related PR only</summary>
|
||||
|
||||
- Concurrent users: ${{ env.N_USERS }}, duration: ${{ github.event.inputs.duration || env.DURATION }}
|
||||
- HTTP request : avg=${{ env.HTTP_REQ_DURATION_AVG }}ms p(95)=${{ env.HTTP_REQ_DURATION_P_95_ }}ms fails=${{ env.HTTP_REQ_FAILED_PASSES }}, finish reason: stop=${{ env.LLAMACPP_COMPLETIONS_STOP_RATE_PASSES }} truncated=${{ env.LLAMACPP_COMPLETIONS_TRUNCATED_RATE_PASSES }}
|
||||
- Prompt processing (pp): avg=${{ env.LLAMACPP_PROMPT_PROCESSING_SECOND_AVG }}tk/s p(95)=${{ env.LLAMACPP_PROMPT_PROCESSING_SECOND_P_95_ }}tk/s
|
||||
- Token generation (tg): avg=${{ env.LLAMACPP_TOKENS_SECOND_AVG }}tk/s p(95)=${{ env.LLAMACPP_TOKENS_SECOND_P_95_ }}tk/s
|
||||
- ${{ env.BENCH_GRAPH_XLABEL }}
|
||||
|
||||
|
||||
<p align="center">
|
||||
|
||||
<img width="100%" height="100%" src="${{ env.IMAGE_O }}" alt="prompt_tokens_seconds" />
|
||||
|
||||
<details>
|
||||
|
||||
<summary>More</summary>
|
||||
|
||||
```mermaid
|
||||
${{ env.PROMPT_TOKENS_SECONDS }}
|
||||
```
|
||||
|
||||
</details>
|
||||
|
||||
<img width="100%" height="100%" src="${{ env.IMAGE_1 }}" alt="predicted_tokens_seconds"/>
|
||||
|
||||
<details>
|
||||
<summary>More</summary>
|
||||
|
||||
```mermaid
|
||||
${{ env.PREDICTED_TOKENS_SECONDS }}
|
||||
```
|
||||
|
||||
</details>
|
||||
|
||||
</p>
|
||||
|
||||
<details>
|
||||
|
||||
<summary>Details</summary>
|
||||
|
||||
<p align="center">
|
||||
|
||||
<img width="100%" height="100%" src="${{ env.IMAGE_2 }}" alt="kv_cache_usage_ratio" />
|
||||
|
||||
<details>
|
||||
<summary>More</summary>
|
||||
|
||||
```mermaid
|
||||
${{ env.KV_CACHE_USAGE_RATIO }}
|
||||
```
|
||||
|
||||
</details>
|
||||
|
||||
<img width="100%" height="100%" src="${{ env.IMAGE_3 }}" alt="requests_processing"/>
|
||||
|
||||
<details>
|
||||
<summary>More</summary>
|
||||
|
||||
```mermaid
|
||||
${{ env.REQUESTS_PROCESSING }}
|
||||
```
|
||||
|
||||
</details>
|
||||
|
||||
</p>
|
||||
</details>
|
||||
</details>
|
1283
.github/workflows/build.yml
vendored
1283
.github/workflows/build.yml
vendored
File diff suppressed because it is too large
Load diff
28
.github/workflows/close-issue.yml
vendored
28
.github/workflows/close-issue.yml
vendored
|
@ -1,28 +0,0 @@
|
|||
name: Close inactive issues
|
||||
on:
|
||||
schedule:
|
||||
- cron: "42 0 * * *"
|
||||
|
||||
# Fine-grant permission
|
||||
# https://docs.github.com/en/actions/security-for-github-actions/security-guides/automatic-token-authentication#modifying-the-permissions-for-the-github_token
|
||||
permissions:
|
||||
issues: write
|
||||
|
||||
jobs:
|
||||
close-issues:
|
||||
runs-on: ubuntu-latest
|
||||
permissions:
|
||||
issues: write
|
||||
pull-requests: write
|
||||
steps:
|
||||
- uses: actions/stale@v5
|
||||
with:
|
||||
exempt-issue-labels: "refactor,help wanted,good first issue,research,bug,roadmap"
|
||||
days-before-issue-stale: 30
|
||||
days-before-issue-close: 14
|
||||
stale-issue-label: "stale"
|
||||
close-issue-message: "This issue was closed because it has been inactive for 14 days since being marked as stale."
|
||||
days-before-pr-stale: -1
|
||||
days-before-pr-close: -1
|
||||
operations-per-run: 10000
|
||||
repo-token: ${{ secrets.GITHUB_TOKEN }}
|
36
.github/workflows/code-coverage.yml
vendored
Normal file
36
.github/workflows/code-coverage.yml
vendored
Normal file
|
@ -0,0 +1,36 @@
|
|||
name: Code Coverage
|
||||
on: [push, pull_request]
|
||||
|
||||
env:
|
||||
GGML_NLOOP: 3
|
||||
GGML_N_THREADS: 1
|
||||
|
||||
jobs:
|
||||
run:
|
||||
runs-on: ubuntu-20.04
|
||||
steps:
|
||||
- name: Checkout
|
||||
uses: actions/checkout@v3
|
||||
|
||||
- name: Dependencies
|
||||
run: |
|
||||
sudo apt-get update
|
||||
sudo apt-get install build-essential gcc-8 lcov
|
||||
|
||||
- name: Build
|
||||
run: CC=gcc-8 make -j LLAMA_CODE_COVERAGE=1 tests
|
||||
|
||||
- name: Run tests
|
||||
run: CC=gcc-8 make test
|
||||
|
||||
- name: Generate coverage report
|
||||
run: |
|
||||
make coverage
|
||||
make lcov-report
|
||||
|
||||
- name: Upload coverage to Codecov
|
||||
uses: codecov/codecov-action@v3
|
||||
env:
|
||||
CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}
|
||||
with:
|
||||
files: lcov-report/coverage.info
|
153
.github/workflows/docker.yml
vendored
153
.github/workflows/docker.yml
vendored
|
@ -10,50 +10,40 @@
|
|||
name: Publish Docker image
|
||||
|
||||
on:
|
||||
workflow_dispatch: # allows manual triggering
|
||||
schedule:
|
||||
# Rebuild daily rather than on every push because it is expensive
|
||||
- cron: '12 4 * * *'
|
||||
|
||||
concurrency:
|
||||
group: ${{ github.workflow }}-${{ github.head_ref && github.ref || github.run_id }}
|
||||
cancel-in-progress: true
|
||||
|
||||
# Fine-grant permission
|
||||
# https://docs.github.com/en/actions/security-for-github-actions/security-guides/automatic-token-authentication#modifying-the-permissions-for-the-github_token
|
||||
permissions:
|
||||
packages: write
|
||||
pull_request:
|
||||
push:
|
||||
branches:
|
||||
- master
|
||||
|
||||
jobs:
|
||||
push_to_registry:
|
||||
name: Push Docker image to Docker Hub
|
||||
if: github.event.pull_request.draft == false
|
||||
|
||||
runs-on: ubuntu-22.04
|
||||
runs-on: ubuntu-latest
|
||||
env:
|
||||
COMMIT_SHA: ${{ github.sha }}
|
||||
strategy:
|
||||
fail-fast: false
|
||||
matrix:
|
||||
config:
|
||||
# Multi-stage build
|
||||
- { tag: "cpu", dockerfile: ".devops/cpu.Dockerfile", platforms: "linux/amd64,linux/arm64", full: true, light: true, server: true, freediskspace: false}
|
||||
- { tag: "cuda", dockerfile: ".devops/cuda.Dockerfile", platforms: "linux/amd64", full: true, light: true, server: true, freediskspace: false}
|
||||
- { tag: "musa", dockerfile: ".devops/musa.Dockerfile", platforms: "linux/amd64", full: true, light: true, server: true, freediskspace: false}
|
||||
- { tag: "intel", dockerfile: ".devops/intel.Dockerfile", platforms: "linux/amd64", full: true, light: true, server: true, freediskspace: false}
|
||||
- { tag: "vulkan", dockerfile: ".devops/vulkan.Dockerfile", platforms: "linux/amd64", full: true, light: true, server: true, freediskspace: false}
|
||||
# Note: the rocm images are failing due to a compiler error and are disabled until this is fixed to allow the workflow to complete
|
||||
#- {tag: "rocm", dockerfile: ".devops/rocm.Dockerfile", platforms: "linux/amd64,linux/arm64", full: true, light: true, server: true, freediskspace: true }
|
||||
- { tag: "light", dockerfile: ".devops/main.Dockerfile", platforms: "linux/amd64,linux/arm64" }
|
||||
- { tag: "full", dockerfile: ".devops/full.Dockerfile", platforms: "linux/amd64,linux/arm64" }
|
||||
# NOTE(canardletter): The CUDA builds on arm64 are very slow, so I
|
||||
# have disabled them for now until the reason why
|
||||
# is understood.
|
||||
- { tag: "light-cuda", dockerfile: ".devops/main-cuda.Dockerfile", platforms: "linux/amd64" }
|
||||
- { tag: "full-cuda", dockerfile: ".devops/full-cuda.Dockerfile", platforms: "linux/amd64" }
|
||||
- { tag: "light-rocm", dockerfile: ".devops/main-rocm.Dockerfile", platforms: "linux/amd64,linux/arm64" }
|
||||
- { tag: "full-rocm", dockerfile: ".devops/full-rocm.Dockerfile", platforms: "linux/amd64,linux/arm64" }
|
||||
steps:
|
||||
- name: Check out the repo
|
||||
uses: actions/checkout@v4
|
||||
with:
|
||||
fetch-depth: 0 # preserve git history, so we can determine the build number
|
||||
uses: actions/checkout@v3
|
||||
|
||||
- name: Set up QEMU
|
||||
uses: docker/setup-qemu-action@v3
|
||||
uses: docker/setup-qemu-action@v2
|
||||
|
||||
- name: Set up Docker Buildx
|
||||
uses: docker/setup-buildx-action@v3
|
||||
uses: docker/setup-buildx-action@v2
|
||||
|
||||
- name: Log in to Docker Hub
|
||||
uses: docker/login-action@v2
|
||||
|
@ -62,112 +52,21 @@ jobs:
|
|||
username: ${{ github.repository_owner }}
|
||||
password: ${{ secrets.GITHUB_TOKEN }}
|
||||
|
||||
- name: Determine tag name
|
||||
id: tag
|
||||
shell: bash
|
||||
run: |
|
||||
BUILD_NUMBER="$(git rev-list --count HEAD)"
|
||||
SHORT_HASH="$(git rev-parse --short=7 HEAD)"
|
||||
REPO_OWNER="${GITHUB_REPOSITORY_OWNER@L}" # to lower case
|
||||
REPO_NAME="${{ github.event.repository.name }}"
|
||||
|
||||
# determine tag name postfix (build number, commit hash)
|
||||
if [[ "${{ env.GITHUB_BRANCH_NAME }}" == "master" ]]; then
|
||||
TAG_POSTFIX="-b${BUILD_NUMBER}"
|
||||
else
|
||||
SAFE_NAME=$(echo "${{ env.GITHUB_BRANCH_NAME }}" | tr '/' '-')
|
||||
TAG_POSTFIX="-${SAFE_NAME}-${SHORT_HASH}"
|
||||
fi
|
||||
# list all tags possible
|
||||
if [[ "${{ matrix.config.tag }}" == "cpu" ]]; then
|
||||
TYPE=""
|
||||
else
|
||||
TYPE="-${{ matrix.config.tag }}"
|
||||
fi
|
||||
PREFIX="ghcr.io/${REPO_OWNER}/${REPO_NAME}:"
|
||||
FULLTAGS="${PREFIX}full${TYPE},${PREFIX}full${TYPE}${TAG_POSTFIX}"
|
||||
LIGHTTAGS="${PREFIX}light${TYPE},${PREFIX}light${TYPE}${TAG_POSTFIX}"
|
||||
SERVERTAGS="${PREFIX}server${TYPE},${PREFIX}server${TYPE}${TAG_POSTFIX}"
|
||||
echo "full_output_tags=$FULLTAGS" >> $GITHUB_OUTPUT
|
||||
echo "light_output_tags=$LIGHTTAGS" >> $GITHUB_OUTPUT
|
||||
echo "server_output_tags=$SERVERTAGS" >> $GITHUB_OUTPUT
|
||||
echo "full_output_tags=$FULLTAGS" # print out for debugging
|
||||
echo "light_output_tags=$LIGHTTAGS" # print out for debugging
|
||||
echo "server_output_tags=$SERVERTAGS" # print out for debugging
|
||||
env:
|
||||
GITHUB_BRANCH_NAME: ${{ github.head_ref || github.ref_name }}
|
||||
GITHUB_REPOSITORY_OWNER: '${{ github.repository_owner }}'
|
||||
|
||||
- name: Free Disk Space (Ubuntu)
|
||||
if: ${{ matrix.config.free_disk_space == true }}
|
||||
uses: ggml-org/free-disk-space@v1.3.1
|
||||
with:
|
||||
# this might remove tools that are actually needed,
|
||||
# if set to "true" but frees about 6 GB
|
||||
tool-cache: false
|
||||
|
||||
# all of these default to true, but feel free to set to
|
||||
# "false" if necessary for your workflow
|
||||
android: true
|
||||
dotnet: true
|
||||
haskell: true
|
||||
large-packages: true
|
||||
docker-images: true
|
||||
swap-storage: true
|
||||
|
||||
- name: Build and push Full Docker image (tagged + versioned)
|
||||
if: ${{ (github.event_name == 'push' || github.event_name == 'schedule' || github.event_name == 'workflow_dispatch') && matrix.config.full == true }}
|
||||
uses: docker/build-push-action@v6
|
||||
- name: Build and push Docker image (versioned)
|
||||
if: github.event_name == 'push'
|
||||
uses: docker/build-push-action@v4
|
||||
with:
|
||||
context: .
|
||||
push: true
|
||||
platforms: ${{ matrix.config.platforms }}
|
||||
# tag list is generated from step above
|
||||
tags: ${{ steps.tag.outputs.full_output_tags }}
|
||||
tags: "ghcr.io/ggerganov/llama.cpp:${{ matrix.config.tag }}-${{ env.COMMIT_SHA }}"
|
||||
file: ${{ matrix.config.dockerfile }}
|
||||
target: full
|
||||
provenance: false
|
||||
# using github experimental cache
|
||||
cache-from: type=gha
|
||||
cache-to: type=gha,mode=max
|
||||
# return to this if the experimental github cache is having issues
|
||||
#cache-to: type=local,dest=/tmp/.buildx-cache
|
||||
#cache-from: type=local,src=/tmp/.buildx-cache
|
||||
|
||||
- name: Build and push Light Docker image (tagged + versioned)
|
||||
if: ${{ (github.event_name == 'push' || github.event_name == 'schedule' || github.event_name == 'workflow_dispatch') && matrix.config.light == true }}
|
||||
uses: docker/build-push-action@v6
|
||||
- name: Build and push Docker image (tagged)
|
||||
uses: docker/build-push-action@v4
|
||||
with:
|
||||
context: .
|
||||
push: true
|
||||
push: ${{ github.event_name == 'push' }}
|
||||
platforms: ${{ matrix.config.platforms }}
|
||||
# tag list is generated from step above
|
||||
tags: ${{ steps.tag.outputs.light_output_tags }}
|
||||
tags: "ghcr.io/ggerganov/llama.cpp:${{ matrix.config.tag }}"
|
||||
file: ${{ matrix.config.dockerfile }}
|
||||
target: light
|
||||
provenance: false
|
||||
# using github experimental cache
|
||||
cache-from: type=gha
|
||||
cache-to: type=gha,mode=max
|
||||
# return to this if the experimental github cache is having issues
|
||||
#cache-to: type=local,dest=/tmp/.buildx-cache
|
||||
#cache-from: type=local,src=/tmp/.buildx-cache
|
||||
|
||||
- name: Build and push Server Docker image (tagged + versioned)
|
||||
if: ${{ (github.event_name == 'push' || github.event_name == 'schedule' || github.event_name == 'workflow_dispatch') && matrix.config.server == true }}
|
||||
uses: docker/build-push-action@v6
|
||||
with:
|
||||
context: .
|
||||
push: true
|
||||
platforms: ${{ matrix.config.platforms }}
|
||||
# tag list is generated from step above
|
||||
tags: ${{ steps.tag.outputs.server_output_tags }}
|
||||
file: ${{ matrix.config.dockerfile }}
|
||||
target: server
|
||||
provenance: false
|
||||
# using github experimental cache
|
||||
cache-from: type=gha
|
||||
cache-to: type=gha,mode=max
|
||||
# return to this if the experimental github cache is having issues
|
||||
#cache-to: type=local,dest=/tmp/.buildx-cache
|
||||
#cache-from: type=local,src=/tmp/.buildx-cache
|
||||
|
|
16
.github/workflows/editorconfig.yml
vendored
16
.github/workflows/editorconfig.yml
vendored
|
@ -1,12 +1,6 @@
|
|||
name: EditorConfig Checker
|
||||
|
||||
on:
|
||||
workflow_dispatch: # allows manual triggering
|
||||
inputs:
|
||||
create_release:
|
||||
description: 'Create new release'
|
||||
required: true
|
||||
type: boolean
|
||||
push:
|
||||
branches:
|
||||
- master
|
||||
|
@ -14,16 +8,10 @@ on:
|
|||
branches:
|
||||
- master
|
||||
|
||||
concurrency:
|
||||
group: ${{ github.workflow }}-${{ github.head_ref && github.ref || github.run_id }}
|
||||
cancel-in-progress: true
|
||||
|
||||
jobs:
|
||||
editorconfig:
|
||||
runs-on: ubuntu-latest
|
||||
steps:
|
||||
- uses: actions/checkout@v4
|
||||
- uses: editorconfig-checker/action-editorconfig-checker@v2
|
||||
with:
|
||||
version: v3.0.3
|
||||
- uses: actions/checkout@v3
|
||||
- uses: editorconfig-checker/action-editorconfig-checker@main
|
||||
- run: editorconfig-checker
|
||||
|
|
4
.github/workflows/gguf-publish.yml
vendored
4
.github/workflows/gguf-publish.yml
vendored
|
@ -24,9 +24,9 @@ jobs:
|
|||
runs-on: ubuntu-latest
|
||||
|
||||
steps:
|
||||
- uses: actions/checkout@v4
|
||||
- uses: actions/checkout@v3
|
||||
- name: Set up Python
|
||||
uses: actions/setup-python@v5
|
||||
uses: actions/setup-python@v2
|
||||
with:
|
||||
python-version: '3.9.x'
|
||||
- name: Install dependencies
|
||||
|
|
17
.github/workflows/labeler.yml
vendored
17
.github/workflows/labeler.yml
vendored
|
@ -1,17 +0,0 @@
|
|||
name: "Pull Request Labeler"
|
||||
on:
|
||||
- pull_request_target
|
||||
|
||||
jobs:
|
||||
labeler:
|
||||
permissions:
|
||||
contents: read
|
||||
pull-requests: write
|
||||
runs-on: ubuntu-latest
|
||||
steps:
|
||||
- uses: actions/checkout@v4
|
||||
with:
|
||||
repository: "ggerganov/llama.cpp"
|
||||
- uses: actions/labeler@v5
|
||||
with:
|
||||
configuration-path: '.github/labeler.yml'
|
33
.github/workflows/python-check-requirements.yml
vendored
33
.github/workflows/python-check-requirements.yml
vendored
|
@ -1,33 +0,0 @@
|
|||
name: Python check requirements.txt
|
||||
|
||||
on:
|
||||
push:
|
||||
paths:
|
||||
- '.github/workflows/python-check-requirements.yml'
|
||||
- 'scripts/check-requirements.sh'
|
||||
- 'convert*.py'
|
||||
- '**/requirements*.txt'
|
||||
pull_request:
|
||||
paths:
|
||||
- '.github/workflows/python-check-requirements.yml'
|
||||
- 'scripts/check-requirements.sh'
|
||||
- 'convert*.py'
|
||||
- '**/requirements*.txt'
|
||||
|
||||
concurrency:
|
||||
group: ${{ github.workflow }}-${{ github.head_ref && github.ref || github.run_id }}
|
||||
cancel-in-progress: true
|
||||
|
||||
jobs:
|
||||
python-check-requirements:
|
||||
runs-on: ubuntu-latest
|
||||
name: check-requirements
|
||||
steps:
|
||||
- name: Check out source repository
|
||||
uses: actions/checkout@v4
|
||||
- name: Set up Python environment
|
||||
uses: actions/setup-python@v5
|
||||
with:
|
||||
python-version: "3.11"
|
||||
- name: Run check-requirements.sh script
|
||||
run: bash scripts/check-requirements.sh
|
20
.github/workflows/python-lint.yml
vendored
20
.github/workflows/python-lint.yml
vendored
|
@ -1,17 +1,6 @@
|
|||
name: flake8 Lint
|
||||
|
||||
on:
|
||||
push:
|
||||
branches:
|
||||
- master
|
||||
paths: ['.github/workflows/python-lint.yml', '**/*.py']
|
||||
pull_request:
|
||||
types: [opened, synchronize, reopened]
|
||||
paths: ['.github/workflows/python-lint.yml', '**/*.py']
|
||||
|
||||
concurrency:
|
||||
group: ${{ github.workflow }}-${{ github.head_ref && github.ref || github.run_id }}
|
||||
cancel-in-progress: true
|
||||
on: [push, pull_request]
|
||||
|
||||
jobs:
|
||||
flake8-lint:
|
||||
|
@ -19,12 +8,13 @@ jobs:
|
|||
name: Lint
|
||||
steps:
|
||||
- name: Check out source repository
|
||||
uses: actions/checkout@v4
|
||||
uses: actions/checkout@v3
|
||||
- name: Set up Python environment
|
||||
uses: actions/setup-python@v5
|
||||
uses: actions/setup-python@v4
|
||||
with:
|
||||
python-version: "3.11"
|
||||
- name: flake8 Lint
|
||||
uses: py-actions/flake8@v2
|
||||
with:
|
||||
plugins: "flake8-no-print"
|
||||
ignore: "E203,E211,E221,E225,E231,E241,E251,E261,E266,E501,E701,E704"
|
||||
exclude: "examples/*,examples/*/**,*/**/__init__.py"
|
||||
|
|
40
.github/workflows/python-type-check.yml
vendored
40
.github/workflows/python-type-check.yml
vendored
|
@ -1,40 +0,0 @@
|
|||
name: Python Type-Check
|
||||
|
||||
on:
|
||||
push:
|
||||
paths:
|
||||
- '.github/workflows/python-type-check.yml'
|
||||
- 'pyrightconfig.json'
|
||||
- '**.py'
|
||||
- '**/requirements*.txt'
|
||||
pull_request:
|
||||
paths:
|
||||
- '.github/workflows/python-type-check.yml'
|
||||
- 'pyrightconfig.json'
|
||||
- '**.py'
|
||||
- '**/requirements*.txt'
|
||||
|
||||
concurrency:
|
||||
group: ${{ github.workflow }}-${{ github.head_ref && github.ref || github.run_id }}
|
||||
cancel-in-progress: true
|
||||
|
||||
jobs:
|
||||
python-type-check:
|
||||
runs-on: ubuntu-latest
|
||||
name: pyright type-check
|
||||
steps:
|
||||
- name: Check out source repository
|
||||
uses: actions/checkout@v4
|
||||
- name: Set up Python environment
|
||||
uses: actions/setup-python@v5
|
||||
with:
|
||||
python-version: "3.11"
|
||||
- name: Install Python dependencies
|
||||
# TODO: use a venv
|
||||
run: pip install -r requirements/requirements-all.txt
|
||||
- name: Type-check with Pyright
|
||||
uses: jakebailey/pyright-action@v2
|
||||
with:
|
||||
version: 1.1.382
|
||||
level: warning
|
||||
warnings: true
|
239
.github/workflows/server.yml
vendored
239
.github/workflows/server.yml
vendored
|
@ -1,239 +0,0 @@
|
|||
# Server build and tests
|
||||
name: Server
|
||||
|
||||
on:
|
||||
workflow_dispatch: # allows manual triggering
|
||||
inputs:
|
||||
sha:
|
||||
description: 'Commit SHA1 to build'
|
||||
required: false
|
||||
type: string
|
||||
slow_tests:
|
||||
description: 'Run slow tests'
|
||||
required: true
|
||||
type: boolean
|
||||
push:
|
||||
branches:
|
||||
- master
|
||||
paths: ['.github/workflows/server.yml', '**/CMakeLists.txt', '**/Makefile', '**/*.h', '**/*.hpp', '**/*.c', '**/*.cpp', '**/*.cu', '**/*.swift', '**/*.m', 'examples/server/**.*']
|
||||
pull_request:
|
||||
types: [opened, synchronize, reopened]
|
||||
paths: ['.github/workflows/server.yml', '**/CMakeLists.txt', '**/Makefile', '**/*.h', '**/*.hpp', '**/*.c', '**/*.cpp', '**/*.cu', '**/*.swift', '**/*.m', 'examples/server/**.*']
|
||||
|
||||
env:
|
||||
LLAMA_LOG_COLORS: 1
|
||||
LLAMA_LOG_PREFIX: 1
|
||||
LLAMA_LOG_TIMESTAMPS: 1
|
||||
LLAMA_LOG_VERBOSITY: 10
|
||||
|
||||
concurrency:
|
||||
group: ${{ github.workflow }}-${{ github.ref }}-${{ github.head_ref || github.run_id }}
|
||||
cancel-in-progress: true
|
||||
|
||||
jobs:
|
||||
server:
|
||||
runs-on: ubuntu-latest
|
||||
|
||||
strategy:
|
||||
matrix:
|
||||
sanitizer: [ADDRESS, UNDEFINED] # THREAD is broken
|
||||
build_type: [RelWithDebInfo]
|
||||
include:
|
||||
- build_type: Release
|
||||
sanitizer: ""
|
||||
fail-fast: false # While -DLLAMA_SANITIZE_THREAD=ON is broken
|
||||
|
||||
steps:
|
||||
- name: Dependencies
|
||||
id: depends
|
||||
run: |
|
||||
sudo apt-get update
|
||||
sudo apt-get -y install \
|
||||
build-essential \
|
||||
xxd \
|
||||
git \
|
||||
cmake \
|
||||
curl \
|
||||
wget \
|
||||
language-pack-en \
|
||||
libcurl4-openssl-dev
|
||||
|
||||
- name: Clone
|
||||
id: checkout
|
||||
uses: actions/checkout@v4
|
||||
with:
|
||||
fetch-depth: 0
|
||||
ref: ${{ github.event.inputs.sha || github.event.pull_request.head.sha || github.sha || github.head_ref || github.ref_name }}
|
||||
|
||||
- name: Python setup
|
||||
id: setup_python
|
||||
uses: actions/setup-python@v5
|
||||
with:
|
||||
python-version: '3.11'
|
||||
|
||||
- name: Tests dependencies
|
||||
id: test_dependencies
|
||||
run: |
|
||||
pip install -r examples/server/tests/requirements.txt
|
||||
|
||||
# Setup nodejs (to be used for verifying bundled index.html)
|
||||
- uses: actions/setup-node@v4
|
||||
with:
|
||||
node-version: '22.11.0'
|
||||
|
||||
- name: WebUI - Install dependencies
|
||||
id: webui_lint
|
||||
run: |
|
||||
cd examples/server/webui
|
||||
npm ci
|
||||
|
||||
- name: WebUI - Check code format
|
||||
id: webui_format
|
||||
run: |
|
||||
git config --global --add safe.directory $(realpath .)
|
||||
cd examples/server/webui
|
||||
git status
|
||||
|
||||
npm run format
|
||||
git status
|
||||
modified_files="$(git status -s)"
|
||||
echo "Modified files: ${modified_files}"
|
||||
if [ -n "${modified_files}" ]; then
|
||||
echo "Files do not follow coding style. To fix: npm run format"
|
||||
echo "${modified_files}"
|
||||
exit 1
|
||||
fi
|
||||
|
||||
- name: Verify bundled index.html
|
||||
id: verify_server_index_html
|
||||
run: |
|
||||
git config --global --add safe.directory $(realpath .)
|
||||
cd examples/server/webui
|
||||
git status
|
||||
|
||||
npm run build
|
||||
git status
|
||||
modified_files="$(git status -s)"
|
||||
echo "Modified files: ${modified_files}"
|
||||
if [ -n "${modified_files}" ]; then
|
||||
echo "Repository is dirty or server/webui is not built as expected"
|
||||
echo "Hint: You may need to follow Web UI build guide in server/README.md"
|
||||
echo "${modified_files}"
|
||||
exit 1
|
||||
fi
|
||||
|
||||
- name: Build (no OpenMP)
|
||||
id: cmake_build_no_openmp
|
||||
if: ${{ matrix.sanitizer == 'THREAD' }}
|
||||
run: |
|
||||
cmake -B build \
|
||||
-DGGML_NATIVE=OFF \
|
||||
-DLLAMA_BUILD_SERVER=ON \
|
||||
-DLLAMA_CURL=ON \
|
||||
-DCMAKE_BUILD_TYPE=${{ matrix.build_type }} \
|
||||
-DLLAMA_SANITIZE_${{ matrix.sanitizer }}=ON \
|
||||
-DGGML_OPENMP=OFF ;
|
||||
cmake --build build --config ${{ matrix.build_type }} -j $(nproc) --target llama-server
|
||||
|
||||
- name: Build (sanitizers)
|
||||
id: cmake_build_sanitizers
|
||||
if: ${{ matrix.sanitizer != '' && matrix.sanitizer != 'THREAD' }}
|
||||
run: |
|
||||
cmake -B build \
|
||||
-DGGML_NATIVE=OFF \
|
||||
-DLLAMA_BUILD_SERVER=ON \
|
||||
-DLLAMA_CURL=ON \
|
||||
-DCMAKE_BUILD_TYPE=${{ matrix.build_type }} \
|
||||
-DLLAMA_SANITIZE_${{ matrix.sanitizer }}=ON ;
|
||||
cmake --build build --config ${{ matrix.build_type }} -j $(nproc) --target llama-server
|
||||
|
||||
- name: Build (sanitizers)
|
||||
id: cmake_build
|
||||
if: ${{ matrix.sanitizer == '' }}
|
||||
run: |
|
||||
cmake -B build \
|
||||
-DGGML_NATIVE=OFF \
|
||||
-DLLAMA_BUILD_SERVER=ON \
|
||||
-DLLAMA_CURL=ON \
|
||||
-DCMAKE_BUILD_TYPE=${{ matrix.build_type }} ;
|
||||
cmake --build build --config ${{ matrix.build_type }} -j $(nproc) --target llama-server
|
||||
|
||||
- name: Tests
|
||||
id: server_integration_tests
|
||||
if: ${{ matrix.sanitizer == '' }}
|
||||
run: |
|
||||
cd examples/server/tests
|
||||
./tests.sh
|
||||
|
||||
- name: Tests (sanitizers)
|
||||
id: server_integration_tests_sanitizers
|
||||
if: ${{ matrix.sanitizer != '' }}
|
||||
run: |
|
||||
cd examples/server/tests
|
||||
LLAMA_SANITIZE=1 ./tests.sh
|
||||
|
||||
- name: Slow tests
|
||||
id: server_integration_tests_slow
|
||||
if: ${{ (github.event.schedule || github.event.inputs.slow_tests == 'true') && matrix.build_type == 'Release' }}
|
||||
run: |
|
||||
cd examples/server/tests
|
||||
SLOW_TESTS=1 ./tests.sh
|
||||
|
||||
|
||||
server-windows:
|
||||
runs-on: windows-2019
|
||||
|
||||
steps:
|
||||
- name: Clone
|
||||
id: checkout
|
||||
uses: actions/checkout@v4
|
||||
with:
|
||||
fetch-depth: 0
|
||||
ref: ${{ github.event.inputs.sha || github.event.pull_request.head.sha || github.sha || github.head_ref || github.ref_name }}
|
||||
|
||||
- name: libCURL
|
||||
id: get_libcurl
|
||||
env:
|
||||
CURL_VERSION: 8.6.0_6
|
||||
run: |
|
||||
curl.exe -o $env:RUNNER_TEMP/curl.zip -L "https://curl.se/windows/dl-${env:CURL_VERSION}/curl-${env:CURL_VERSION}-win64-mingw.zip"
|
||||
mkdir $env:RUNNER_TEMP/libcurl
|
||||
tar.exe -xvf $env:RUNNER_TEMP/curl.zip --strip-components=1 -C $env:RUNNER_TEMP/libcurl
|
||||
|
||||
- name: Build
|
||||
id: cmake_build
|
||||
run: |
|
||||
cmake -B build -DLLAMA_CURL=ON -DCURL_LIBRARY="$env:RUNNER_TEMP/libcurl/lib/libcurl.dll.a" -DCURL_INCLUDE_DIR="$env:RUNNER_TEMP/libcurl/include"
|
||||
cmake --build build --config Release -j ${env:NUMBER_OF_PROCESSORS} --target llama-server
|
||||
|
||||
- name: Python setup
|
||||
id: setup_python
|
||||
uses: actions/setup-python@v5
|
||||
with:
|
||||
python-version: '3.11'
|
||||
|
||||
- name: Tests dependencies
|
||||
id: test_dependencies
|
||||
run: |
|
||||
pip install -r examples/server/tests/requirements.txt
|
||||
|
||||
- name: Copy Libcurl
|
||||
id: prepare_libcurl
|
||||
run: |
|
||||
cp $env:RUNNER_TEMP/libcurl/bin/libcurl-x64.dll ./build/bin/Release/libcurl-x64.dll
|
||||
|
||||
- name: Tests
|
||||
id: server_integration_tests
|
||||
if: ${{ !matrix.disabled_on_pr || !github.event.pull_request }}
|
||||
run: |
|
||||
cd examples/server/tests
|
||||
$env:PYTHONIOENCODING = ":replace"
|
||||
pytest -v -x -m "not slow"
|
||||
|
||||
- name: Slow tests
|
||||
id: server_integration_tests_slow
|
||||
if: ${{ (github.event.schedule || github.event.inputs.slow_tests == 'true') && matrix.build_type == 'Release' }}
|
||||
run: |
|
||||
cd examples/server/tests
|
||||
$env:SLOW_TESTS = "1"
|
||||
pytest -v -x
|
20
.github/workflows/tidy-post.yml
vendored
Normal file
20
.github/workflows/tidy-post.yml
vendored
Normal file
|
@ -0,0 +1,20 @@
|
|||
name: clang-tidy review post comments
|
||||
|
||||
on:
|
||||
workflow_dispatch:
|
||||
workflows: ["clang-tidy-review"]
|
||||
types:
|
||||
- completed
|
||||
|
||||
jobs:
|
||||
build:
|
||||
runs-on: ubuntu-latest
|
||||
|
||||
steps:
|
||||
- uses: ZedThree/clang-tidy-review/post@v0.13.0
|
||||
# lgtm_comment_body, max_comments, and annotations need to be set on the posting workflow in a split setup
|
||||
with:
|
||||
# adjust options as necessary
|
||||
lgtm_comment_body: ''
|
||||
annotations: false
|
||||
max_comments: 25
|
23
.github/workflows/tidy-review.yml
vendored
Normal file
23
.github/workflows/tidy-review.yml
vendored
Normal file
|
@ -0,0 +1,23 @@
|
|||
name: clang-tidy-review
|
||||
|
||||
on:
|
||||
pull_request:
|
||||
branches:
|
||||
- master
|
||||
|
||||
jobs:
|
||||
clang-tidy-review:
|
||||
runs-on: ubuntu-latest
|
||||
|
||||
steps:
|
||||
- uses: actions/checkout@v3
|
||||
|
||||
- uses: ZedThree/clang-tidy-review@v0.13.0
|
||||
id: review
|
||||
with:
|
||||
lgtm_comment_body: ''
|
||||
build_dir: build
|
||||
cmake_command: cmake . -B build -DCMAKE_EXPORT_COMPILE_COMMANDS=on
|
||||
split_workflow: true
|
||||
|
||||
- uses: ZedThree/clang-tidy-review/upload@v0.13.0
|
25
.github/workflows/zig-build.yml
vendored
Normal file
25
.github/workflows/zig-build.yml
vendored
Normal file
|
@ -0,0 +1,25 @@
|
|||
name: Zig CI
|
||||
|
||||
on:
|
||||
pull_request:
|
||||
push:
|
||||
branches:
|
||||
- master
|
||||
|
||||
jobs:
|
||||
build:
|
||||
strategy:
|
||||
fail-fast: false
|
||||
matrix:
|
||||
runs-on: [ubuntu-latest, macos-latest, windows-latest]
|
||||
runs-on: ${{ matrix.runs-on }}
|
||||
steps:
|
||||
- uses: actions/checkout@v3
|
||||
with:
|
||||
submodules: recursive
|
||||
fetch-depth: 0
|
||||
- uses: goto-bus-stop/setup-zig@v2
|
||||
with:
|
||||
version: 0.11.0
|
||||
- name: Build Summary
|
||||
run: zig build --summary all -freference-trace
|
175
.gitignore
vendored
175
.gitignore
vendored
|
@ -1,145 +1,104 @@
|
|||
# Extensions
|
||||
|
||||
*.a
|
||||
*.bat
|
||||
*.bin
|
||||
*.d
|
||||
*.dll
|
||||
*.dot
|
||||
*.etag
|
||||
*.exe
|
||||
*.gcda
|
||||
*.gcno
|
||||
*.gcov
|
||||
*.gguf
|
||||
*.gguf.json
|
||||
*.lastModified
|
||||
*.log
|
||||
*.metallib
|
||||
*.o
|
||||
*.a
|
||||
*.so
|
||||
*.swp
|
||||
*.tmp
|
||||
|
||||
# IDE / OS
|
||||
|
||||
*.gguf
|
||||
*.bin
|
||||
*.exe
|
||||
*.dll
|
||||
*.log
|
||||
*.gcov
|
||||
*.gcno
|
||||
*.gcda
|
||||
*.dot
|
||||
*.bat
|
||||
*.metallib
|
||||
.DS_Store
|
||||
.build/
|
||||
.cache/
|
||||
.ccls-cache/
|
||||
.direnv/
|
||||
.DS_Store
|
||||
.envrc
|
||||
.idea/
|
||||
.swiftpm
|
||||
.venv
|
||||
.clang-tidy
|
||||
.vs/
|
||||
.vscode/
|
||||
nppBackup
|
||||
|
||||
|
||||
# Coverage
|
||||
|
||||
gcovr-report/
|
||||
lcov-report/
|
||||
gcovr-report/
|
||||
|
||||
# Build Artifacts
|
||||
|
||||
tags
|
||||
.build/
|
||||
build*
|
||||
!build-info.cmake
|
||||
!build-info.cpp.in
|
||||
!build-info.sh
|
||||
!build.zig
|
||||
!docs/build.md
|
||||
/libllama.so
|
||||
/llama-*
|
||||
/vulkan-shaders-gen
|
||||
android-ndk-*
|
||||
arm_neon.h
|
||||
cmake-build-*
|
||||
CMakeSettings.json
|
||||
compile_commands.json
|
||||
ggml-metal-embed.metal
|
||||
llama-batched-swift
|
||||
/rpc-server
|
||||
build*/
|
||||
out/
|
||||
tmp/
|
||||
autogen-*.md
|
||||
|
||||
# Deprecated
|
||||
|
||||
/main
|
||||
/server
|
||||
|
||||
# CI
|
||||
|
||||
!.github/workflows/*.yml
|
||||
|
||||
# Models
|
||||
|
||||
models/*
|
||||
models-mnt
|
||||
!models/.editorconfig
|
||||
!models/ggml-vocab-*.gguf*
|
||||
|
||||
# Zig
|
||||
/Pipfile
|
||||
/baby-llama
|
||||
/beam-search
|
||||
/benchmark-matmult
|
||||
/convert-llama2c-to-ggml
|
||||
/embd-input-test
|
||||
/embedding
|
||||
/gguf
|
||||
/gguf-llama-simple
|
||||
/infill
|
||||
/libllama.so
|
||||
/llama-bench
|
||||
/llava-cli
|
||||
/lookahead
|
||||
/main
|
||||
/metal
|
||||
/perplexity
|
||||
/q8dot
|
||||
/quantize
|
||||
/quantize-stats
|
||||
/result
|
||||
/save-load-state
|
||||
/server
|
||||
/simple
|
||||
/batched
|
||||
/batched-bench
|
||||
/export-lora
|
||||
/finetune
|
||||
/speculative
|
||||
/parallel
|
||||
/train-text-from-scratch
|
||||
/tokenize
|
||||
/vdot
|
||||
/common/build-info.cpp
|
||||
arm_neon.h
|
||||
compile_commands.json
|
||||
CMakeSettings.json
|
||||
|
||||
__pycache__
|
||||
dist
|
||||
|
||||
zig-out/
|
||||
zig-cache/
|
||||
|
||||
# Logs
|
||||
|
||||
ppl-*.txt
|
||||
qnt-*.txt
|
||||
perf-*.txt
|
||||
|
||||
# Examples
|
||||
|
||||
examples/jeopardy/results.txt
|
||||
examples/server/*.css.hpp
|
||||
examples/server/*.html.hpp
|
||||
examples/server/*.js.hpp
|
||||
examples/server/*.mjs.hpp
|
||||
!build_64.sh
|
||||
!examples/*.bat
|
||||
!examples/*/*.kts
|
||||
!examples/*/*/*.kts
|
||||
!examples/sycl/*.bat
|
||||
!examples/sycl/*.sh
|
||||
|
||||
# Server Web UI temporary files
|
||||
node_modules
|
||||
examples/server/webui/dist
|
||||
|
||||
# Python
|
||||
|
||||
/.venv
|
||||
__pycache__/
|
||||
*/poetry.lock
|
||||
poetry.lock
|
||||
poetry.toml
|
||||
|
||||
# Nix
|
||||
/result
|
||||
|
||||
# Test binaries
|
||||
/tests/test-backend-ops
|
||||
/tests/test-double-float
|
||||
/tests/test-grad0
|
||||
/tests/test-grammar-parser
|
||||
/tests/test-llama-grammar
|
||||
/tests/test-double-float
|
||||
/tests/test-grad0
|
||||
/tests/test-opt
|
||||
/tests/test-quantize-fns
|
||||
/tests/test-quantize-perf
|
||||
/tests/test-rope
|
||||
/tests/test-sampling
|
||||
/tests/test-tokenizer-0
|
||||
/tests/test-tokenizer-0-llama
|
||||
/tests/test-tokenizer-0-falcon
|
||||
/tests/test-tokenizer-1-llama
|
||||
/tests/test-tokenizer-1-bpe
|
||||
/tests/test-tokenizer-1-spm
|
||||
|
||||
# Scripts
|
||||
!/scripts/install-oneapi.bat
|
||||
|
||||
# Test models for lora adapters
|
||||
/lora-tests
|
||||
|
||||
# Local scripts
|
||||
/run-vim.sh
|
||||
/run-chat.sh
|
||||
/tests/test-rope
|
||||
/tests/test-backend-ops
|
||||
|
|
3
.gitmodules
vendored
3
.gitmodules
vendored
|
@ -1,3 +0,0 @@
|
|||
[submodule "kompute"]
|
||||
path = ggml/src/ggml-kompute/kompute
|
||||
url = https://github.com/nomic-ai/kompute.git
|
|
@ -3,14 +3,13 @@
|
|||
exclude: prompts/.*.txt
|
||||
repos:
|
||||
- repo: https://github.com/pre-commit/pre-commit-hooks
|
||||
rev: v4.6.0
|
||||
rev: v3.2.0
|
||||
hooks:
|
||||
- id: trailing-whitespace
|
||||
- id: end-of-file-fixer
|
||||
- id: check-yaml
|
||||
- id: check-added-large-files
|
||||
- repo: https://github.com/PyCQA/flake8
|
||||
rev: 7.0.0
|
||||
rev: 6.0.0
|
||||
hooks:
|
||||
- id: flake8
|
||||
additional_dependencies: [flake8-no-print]
|
||||
|
|
874
CMakeLists.txt
874
CMakeLists.txt
|
@ -1,9 +1,5 @@
|
|||
cmake_minimum_required(VERSION 3.14) # for add_link_options and implicit target directories.
|
||||
cmake_minimum_required(VERSION 3.13) # for add_link_options
|
||||
project("llama.cpp" C CXX)
|
||||
include(CheckIncludeFileCXX)
|
||||
|
||||
#set(CMAKE_WARN_DEPRECATED YES)
|
||||
set(CMAKE_WARN_UNUSED_CLI YES)
|
||||
|
||||
set(CMAKE_EXPORT_COMPILE_COMMANDS ON)
|
||||
|
||||
|
@ -12,17 +8,11 @@ if (NOT XCODE AND NOT MSVC AND NOT CMAKE_BUILD_TYPE)
|
|||
set_property(CACHE CMAKE_BUILD_TYPE PROPERTY STRINGS "Debug" "Release" "MinSizeRel" "RelWithDebInfo")
|
||||
endif()
|
||||
|
||||
# Add path to modules
|
||||
list(APPEND CMAKE_MODULE_PATH "${CMAKE_CURRENT_SOURCE_DIR}/cmake/")
|
||||
|
||||
set(CMAKE_RUNTIME_OUTPUT_DIRECTORY ${CMAKE_BINARY_DIR}/bin)
|
||||
set(CMAKE_LIBRARY_OUTPUT_DIRECTORY ${CMAKE_BINARY_DIR}/bin)
|
||||
|
||||
if (CMAKE_SOURCE_DIR STREQUAL CMAKE_CURRENT_SOURCE_DIR)
|
||||
set(LLAMA_STANDALONE ON)
|
||||
|
||||
include(git-vars)
|
||||
|
||||
# configure project version
|
||||
# TODO
|
||||
else()
|
||||
|
@ -41,139 +31,738 @@ else()
|
|||
endif()
|
||||
endif()
|
||||
|
||||
option(BUILD_SHARED_LIBS "build shared libraries" ${BUILD_SHARED_LIBS_DEFAULT})
|
||||
|
||||
if (WIN32)
|
||||
add_compile_definitions(_CRT_SECURE_NO_WARNINGS)
|
||||
endif()
|
||||
|
||||
if (MSVC)
|
||||
add_compile_options("$<$<COMPILE_LANGUAGE:C>:/utf-8>")
|
||||
add_compile_options("$<$<COMPILE_LANGUAGE:CXX>:/utf-8>")
|
||||
add_compile_options("$<$<COMPILE_LANGUAGE:C>:/bigobj>")
|
||||
add_compile_options("$<$<COMPILE_LANGUAGE:CXX>:/bigobj>")
|
||||
endif()
|
||||
|
||||
#
|
||||
# option list
|
||||
# Option list
|
||||
#
|
||||
|
||||
if (APPLE)
|
||||
set(LLAMA_METAL_DEFAULT ON)
|
||||
else()
|
||||
set(LLAMA_METAL_DEFAULT OFF)
|
||||
endif()
|
||||
|
||||
# general
|
||||
option(BUILD_SHARED_LIBS "build shared libraries" OFF)
|
||||
option(LLAMA_STATIC "llama: static link libraries" OFF)
|
||||
option(LLAMA_NATIVE "llama: enable -march=native flag" ON)
|
||||
option(LLAMA_LTO "llama: enable link time optimization" OFF)
|
||||
|
||||
# debug
|
||||
option(LLAMA_ALL_WARNINGS "llama: enable all compiler warnings" ON)
|
||||
option(LLAMA_ALL_WARNINGS_3RD_PARTY "llama: enable all compiler warnings in 3rd party libs" OFF)
|
||||
|
||||
# build
|
||||
option(LLAMA_FATAL_WARNINGS "llama: enable -Werror flag" OFF)
|
||||
option(LLAMA_ALL_WARNINGS "llama: enable all compiler warnings" ON)
|
||||
option(LLAMA_ALL_WARNINGS_3RD_PARTY "llama: enable all compiler warnings in 3rd party libs" OFF)
|
||||
option(LLAMA_GPROF "llama: enable gprof" OFF)
|
||||
|
||||
# sanitizers
|
||||
option(LLAMA_SANITIZE_THREAD "llama: enable thread sanitizer" OFF)
|
||||
option(LLAMA_SANITIZE_ADDRESS "llama: enable address sanitizer" OFF)
|
||||
option(LLAMA_SANITIZE_UNDEFINED "llama: enable undefined sanitizer" OFF)
|
||||
option(LLAMA_SANITIZE_THREAD "llama: enable thread sanitizer" OFF)
|
||||
option(LLAMA_SANITIZE_ADDRESS "llama: enable address sanitizer" OFF)
|
||||
option(LLAMA_SANITIZE_UNDEFINED "llama: enable undefined sanitizer" OFF)
|
||||
|
||||
# utils
|
||||
option(LLAMA_BUILD_COMMON "llama: build common utils library" ${LLAMA_STANDALONE})
|
||||
# instruction set specific
|
||||
if (LLAMA_NATIVE)
|
||||
set(INS_ENB OFF)
|
||||
else()
|
||||
set(INS_ENB ON)
|
||||
endif()
|
||||
|
||||
# extra artifacts
|
||||
option(LLAMA_BUILD_TESTS "llama: build tests" ${LLAMA_STANDALONE})
|
||||
option(LLAMA_BUILD_EXAMPLES "llama: build examples" ${LLAMA_STANDALONE})
|
||||
option(LLAMA_BUILD_SERVER "llama: build server example" ${LLAMA_STANDALONE})
|
||||
option(LLAMA_AVX "llama: enable AVX" ${INS_ENB})
|
||||
option(LLAMA_AVX2 "llama: enable AVX2" ${INS_ENB})
|
||||
option(LLAMA_AVX512 "llama: enable AVX512" OFF)
|
||||
option(LLAMA_AVX512_VBMI "llama: enable AVX512-VBMI" OFF)
|
||||
option(LLAMA_AVX512_VNNI "llama: enable AVX512-VNNI" OFF)
|
||||
option(LLAMA_FMA "llama: enable FMA" ${INS_ENB})
|
||||
# in MSVC F16C is implied with AVX2/AVX512
|
||||
if (NOT MSVC)
|
||||
option(LLAMA_F16C "llama: enable F16C" ${INS_ENB})
|
||||
endif()
|
||||
|
||||
# 3rd party libs
|
||||
option(LLAMA_CURL "llama: use libcurl to download model from an URL" OFF)
|
||||
option(LLAMA_LLGUIDANCE "llama-common: include LLGuidance library for structured output in common utils" OFF)
|
||||
option(LLAMA_ACCELERATE "llama: enable Accelerate framework" ON)
|
||||
option(LLAMA_BLAS "llama: use BLAS" OFF)
|
||||
set(LLAMA_BLAS_VENDOR "Generic" CACHE STRING "llama: BLAS library vendor")
|
||||
option(LLAMA_CUBLAS "llama: use CUDA" OFF)
|
||||
#option(LLAMA_CUDA_CUBLAS "llama: use cuBLAS for prompt processing" OFF)
|
||||
option(LLAMA_CUDA_FORCE_DMMV "llama: use dmmv instead of mmvq CUDA kernels" OFF)
|
||||
option(LLAMA_CUDA_FORCE_MMQ "llama: use mmq kernels instead of cuBLAS" OFF)
|
||||
set(LLAMA_CUDA_DMMV_X "32" CACHE STRING "llama: x stride for dmmv CUDA kernels")
|
||||
set(LLAMA_CUDA_MMV_Y "1" CACHE STRING "llama: y block size for mmv CUDA kernels")
|
||||
option(LLAMA_CUDA_F16 "llama: use 16 bit floats for some calculations" OFF)
|
||||
set(LLAMA_CUDA_KQUANTS_ITER "2" CACHE STRING "llama: iters./thread per block for Q2_K/Q6_K")
|
||||
set(LLAMA_CUDA_PEER_MAX_BATCH_SIZE "128" CACHE STRING
|
||||
"llama: max. batch size for using peer access")
|
||||
option(LLAMA_HIPBLAS "llama: use hipBLAS" OFF)
|
||||
option(LLAMA_CLBLAST "llama: use CLBlast" OFF)
|
||||
option(LLAMA_METAL "llama: use Metal" ${LLAMA_METAL_DEFAULT})
|
||||
option(LLAMA_METAL_NDEBUG "llama: disable Metal debugging" OFF)
|
||||
option(LLAMA_MPI "llama: use MPI" OFF)
|
||||
option(LLAMA_QKK_64 "llama: use super-block size of 64 for k-quants" OFF)
|
||||
|
||||
option(LLAMA_BUILD_TESTS "llama: build tests" ${LLAMA_STANDALONE})
|
||||
option(LLAMA_BUILD_EXAMPLES "llama: build examples" ${LLAMA_STANDALONE})
|
||||
option(LLAMA_BUILD_SERVER "llama: build server example" ON)
|
||||
|
||||
# Required for relocatable CMake package
|
||||
include(${CMAKE_CURRENT_SOURCE_DIR}/cmake/build-info.cmake)
|
||||
include(${CMAKE_CURRENT_SOURCE_DIR}/cmake/common.cmake)
|
||||
include(${CMAKE_CURRENT_SOURCE_DIR}/scripts/build-info.cmake)
|
||||
|
||||
# override ggml options
|
||||
set(GGML_ALL_WARNINGS ${LLAMA_ALL_WARNINGS})
|
||||
set(GGML_FATAL_WARNINGS ${LLAMA_FATAL_WARNINGS})
|
||||
#
|
||||
# Compile flags
|
||||
#
|
||||
|
||||
# change the default for these ggml options
|
||||
if (NOT DEFINED GGML_LLAMAFILE)
|
||||
set(GGML_LLAMAFILE_DEFAULT ON)
|
||||
set(CMAKE_CXX_STANDARD 11)
|
||||
set(CMAKE_CXX_STANDARD_REQUIRED true)
|
||||
set(CMAKE_C_STANDARD 11)
|
||||
set(CMAKE_C_STANDARD_REQUIRED true)
|
||||
set(THREADS_PREFER_PTHREAD_FLAG ON)
|
||||
find_package(Threads REQUIRED)
|
||||
include(CheckCXXCompilerFlag)
|
||||
|
||||
# enable libstdc++ assertions for debug builds
|
||||
if (CMAKE_SYSTEM_NAME MATCHES "Linux")
|
||||
add_compile_definitions($<$<CONFIG:Debug>:_GLIBCXX_ASSERTIONS>)
|
||||
endif()
|
||||
|
||||
if (NOT DEFINED GGML_CUDA_GRAPHS)
|
||||
set(GGML_CUDA_GRAPHS_DEFAULT ON)
|
||||
endif()
|
||||
|
||||
# transition helpers
|
||||
function (llama_option_depr TYPE OLD NEW)
|
||||
if (${OLD})
|
||||
message(${TYPE} "${OLD} is deprecated and will be removed in the future.\nUse ${NEW} instead\n")
|
||||
set(${NEW} ON PARENT_SCOPE)
|
||||
endif()
|
||||
endfunction()
|
||||
|
||||
llama_option_depr(FATAL_ERROR LLAMA_CUBLAS GGML_CUDA)
|
||||
llama_option_depr(WARNING LLAMA_CUDA GGML_CUDA)
|
||||
llama_option_depr(WARNING LLAMA_KOMPUTE GGML_KOMPUTE)
|
||||
llama_option_depr(WARNING LLAMA_METAL GGML_METAL)
|
||||
llama_option_depr(WARNING LLAMA_METAL_EMBED_LIBRARY GGML_METAL_EMBED_LIBRARY)
|
||||
llama_option_depr(WARNING LLAMA_NATIVE GGML_NATIVE)
|
||||
llama_option_depr(WARNING LLAMA_RPC GGML_RPC)
|
||||
llama_option_depr(WARNING LLAMA_SYCL GGML_SYCL)
|
||||
llama_option_depr(WARNING LLAMA_SYCL_F16 GGML_SYCL_F16)
|
||||
llama_option_depr(WARNING LLAMA_CANN GGML_CANN)
|
||||
|
||||
if (NOT MSVC)
|
||||
if (LLAMA_SANITIZE_THREAD)
|
||||
message(STATUS "Using -fsanitize=thread")
|
||||
|
||||
add_compile_options(-fsanitize=thread)
|
||||
link_libraries (-fsanitize=thread)
|
||||
link_libraries(-fsanitize=thread)
|
||||
endif()
|
||||
|
||||
if (LLAMA_SANITIZE_ADDRESS)
|
||||
message(STATUS "Using -fsanitize=address")
|
||||
|
||||
add_compile_options(-fsanitize=address -fno-omit-frame-pointer)
|
||||
link_libraries (-fsanitize=address)
|
||||
link_libraries(-fsanitize=address)
|
||||
endif()
|
||||
|
||||
if (LLAMA_SANITIZE_UNDEFINED)
|
||||
message(STATUS "Using -fsanitize=undefined")
|
||||
|
||||
add_compile_options(-fsanitize=undefined)
|
||||
link_libraries (-fsanitize=undefined)
|
||||
link_libraries(-fsanitize=undefined)
|
||||
endif()
|
||||
endif()
|
||||
|
||||
#
|
||||
# 3rd-party
|
||||
#
|
||||
if (APPLE AND LLAMA_ACCELERATE)
|
||||
find_library(ACCELERATE_FRAMEWORK Accelerate)
|
||||
if (ACCELERATE_FRAMEWORK)
|
||||
message(STATUS "Accelerate framework found")
|
||||
|
||||
if (NOT TARGET ggml)
|
||||
add_subdirectory(ggml)
|
||||
# ... otherwise assume ggml is added by a parent CMakeLists.txt
|
||||
add_compile_definitions(GGML_USE_ACCELERATE)
|
||||
add_compile_definitions(ACCELERATE_NEW_LAPACK)
|
||||
add_compile_definitions(ACCELERATE_LAPACK_ILP64)
|
||||
set(LLAMA_EXTRA_LIBS ${LLAMA_EXTRA_LIBS} ${ACCELERATE_FRAMEWORK})
|
||||
else()
|
||||
message(WARNING "Accelerate framework not found")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
if (LLAMA_METAL)
|
||||
find_library(FOUNDATION_LIBRARY Foundation REQUIRED)
|
||||
find_library(METAL_FRAMEWORK Metal REQUIRED)
|
||||
find_library(METALKIT_FRAMEWORK MetalKit REQUIRED)
|
||||
|
||||
message(STATUS "Metal framework found")
|
||||
set(GGML_HEADERS_METAL ggml-metal.h)
|
||||
set(GGML_SOURCES_METAL ggml-metal.m)
|
||||
|
||||
add_compile_definitions(GGML_USE_METAL)
|
||||
if (LLAMA_METAL_NDEBUG)
|
||||
add_compile_definitions(GGML_METAL_NDEBUG)
|
||||
endif()
|
||||
|
||||
# get full path to the file
|
||||
#add_compile_definitions(GGML_METAL_DIR_KERNELS="${CMAKE_CURRENT_SOURCE_DIR}/")
|
||||
|
||||
# copy ggml-metal.metal to bin directory
|
||||
configure_file(ggml-metal.metal ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-metal.metal COPYONLY)
|
||||
|
||||
set(LLAMA_EXTRA_LIBS ${LLAMA_EXTRA_LIBS}
|
||||
${FOUNDATION_LIBRARY}
|
||||
${METAL_FRAMEWORK}
|
||||
${METALKIT_FRAMEWORK}
|
||||
)
|
||||
endif()
|
||||
if (LLAMA_BLAS)
|
||||
if (LLAMA_STATIC)
|
||||
set(BLA_STATIC ON)
|
||||
endif()
|
||||
if ($(CMAKE_VERSION) VERSION_GREATER_EQUAL 3.22)
|
||||
set(BLA_SIZEOF_INTEGER 8)
|
||||
endif()
|
||||
|
||||
set(BLA_VENDOR ${LLAMA_BLAS_VENDOR})
|
||||
find_package(BLAS)
|
||||
|
||||
if (BLAS_FOUND)
|
||||
message(STATUS "BLAS found, Libraries: ${BLAS_LIBRARIES}")
|
||||
|
||||
if ("${BLAS_INCLUDE_DIRS}" STREQUAL "")
|
||||
# BLAS_INCLUDE_DIRS is missing in FindBLAS.cmake.
|
||||
# see https://gitlab.kitware.com/cmake/cmake/-/issues/20268
|
||||
find_package(PkgConfig REQUIRED)
|
||||
if (${LLAMA_BLAS_VENDOR} MATCHES "Generic")
|
||||
pkg_check_modules(DepBLAS REQUIRED blas)
|
||||
elseif (${LLAMA_BLAS_VENDOR} MATCHES "OpenBLAS")
|
||||
pkg_check_modules(DepBLAS REQUIRED openblas)
|
||||
elseif (${LLAMA_BLAS_VENDOR} MATCHES "FLAME")
|
||||
pkg_check_modules(DepBLAS REQUIRED blis)
|
||||
elseif (${LLAMA_BLAS_VENDOR} MATCHES "ATLAS")
|
||||
pkg_check_modules(DepBLAS REQUIRED blas-atlas)
|
||||
elseif (${LLAMA_BLAS_VENDOR} MATCHES "FlexiBLAS")
|
||||
pkg_check_modules(DepBLAS REQUIRED flexiblas_api)
|
||||
elseif (${LLAMA_BLAS_VENDOR} MATCHES "Intel")
|
||||
# all Intel* libraries share the same include path
|
||||
pkg_check_modules(DepBLAS REQUIRED mkl-sdl)
|
||||
elseif (${LLAMA_BLAS_VENDOR} MATCHES "NVHPC")
|
||||
# this doesn't provide pkg-config
|
||||
# suggest to assign BLAS_INCLUDE_DIRS on your own
|
||||
if ("${NVHPC_VERSION}" STREQUAL "")
|
||||
message(WARNING "Better to set NVHPC_VERSION")
|
||||
else()
|
||||
set(DepBLAS_FOUND ON)
|
||||
set(DepBLAS_INCLUDE_DIRS "/opt/nvidia/hpc_sdk/${CMAKE_SYSTEM_NAME}_${CMAKE_SYSTEM_PROCESSOR}/${NVHPC_VERSION}/math_libs/include")
|
||||
endif()
|
||||
endif()
|
||||
if (DepBLAS_FOUND)
|
||||
set(BLAS_INCLUDE_DIRS ${DepBLAS_INCLUDE_DIRS})
|
||||
else()
|
||||
message(WARNING "BLAS_INCLUDE_DIRS neither been provided nor been automatically"
|
||||
" detected by pkgconfig, trying to find cblas.h from possible paths...")
|
||||
find_path(BLAS_INCLUDE_DIRS
|
||||
NAMES cblas.h
|
||||
HINTS
|
||||
/usr/include
|
||||
/usr/local/include
|
||||
/usr/include/openblas
|
||||
/opt/homebrew/opt/openblas/include
|
||||
/usr/local/opt/openblas/include
|
||||
/usr/include/x86_64-linux-gnu/openblas/include
|
||||
)
|
||||
endif()
|
||||
endif()
|
||||
|
||||
message(STATUS "BLAS found, Includes: ${BLAS_INCLUDE_DIRS}")
|
||||
add_compile_options(${BLAS_LINKER_FLAGS})
|
||||
add_compile_definitions(GGML_USE_OPENBLAS)
|
||||
if (${BLAS_INCLUDE_DIRS} MATCHES "mkl" AND (${LLAMA_BLAS_VENDOR} MATCHES "Generic" OR ${LLAMA_BLAS_VENDOR} MATCHES "Intel"))
|
||||
add_compile_definitions(GGML_BLAS_USE_MKL)
|
||||
endif()
|
||||
set(LLAMA_EXTRA_LIBS ${LLAMA_EXTRA_LIBS} ${BLAS_LIBRARIES})
|
||||
set(LLAMA_EXTRA_INCLUDES ${LLAMA_EXTRA_INCLUDES} ${BLAS_INCLUDE_DIRS})
|
||||
|
||||
else()
|
||||
message(WARNING "BLAS not found, please refer to "
|
||||
"https://cmake.org/cmake/help/latest/module/FindBLAS.html#blas-lapack-vendors"
|
||||
" to set correct LLAMA_BLAS_VENDOR")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
if (LLAMA_QKK_64)
|
||||
add_compile_definitions(GGML_QKK_64)
|
||||
endif()
|
||||
|
||||
if (LLAMA_CUBLAS)
|
||||
cmake_minimum_required(VERSION 3.17)
|
||||
|
||||
find_package(CUDAToolkit)
|
||||
if (CUDAToolkit_FOUND)
|
||||
message(STATUS "cuBLAS found")
|
||||
|
||||
enable_language(CUDA)
|
||||
|
||||
set(GGML_HEADERS_CUDA ggml-cuda.h)
|
||||
set(GGML_SOURCES_CUDA ggml-cuda.cu)
|
||||
|
||||
add_compile_definitions(GGML_USE_CUBLAS)
|
||||
# if (LLAMA_CUDA_CUBLAS)
|
||||
# add_compile_definitions(GGML_CUDA_CUBLAS)
|
||||
# endif()
|
||||
if (LLAMA_CUDA_FORCE_DMMV)
|
||||
add_compile_definitions(GGML_CUDA_FORCE_DMMV)
|
||||
endif()
|
||||
if (LLAMA_CUDA_FORCE_MMQ)
|
||||
add_compile_definitions(GGML_CUDA_FORCE_MMQ)
|
||||
endif()
|
||||
add_compile_definitions(GGML_CUDA_DMMV_X=${LLAMA_CUDA_DMMV_X})
|
||||
add_compile_definitions(GGML_CUDA_MMV_Y=${LLAMA_CUDA_MMV_Y})
|
||||
if (DEFINED LLAMA_CUDA_DMMV_Y)
|
||||
add_compile_definitions(GGML_CUDA_MMV_Y=${LLAMA_CUDA_DMMV_Y}) # for backwards compatibility
|
||||
endif()
|
||||
if (LLAMA_CUDA_F16 OR LLAMA_CUDA_DMMV_F16)
|
||||
add_compile_definitions(GGML_CUDA_F16)
|
||||
endif()
|
||||
add_compile_definitions(K_QUANTS_PER_ITERATION=${LLAMA_CUDA_KQUANTS_ITER})
|
||||
add_compile_definitions(GGML_CUDA_PEER_MAX_BATCH_SIZE=${LLAMA_CUDA_PEER_MAX_BATCH_SIZE})
|
||||
|
||||
if (LLAMA_STATIC)
|
||||
if (WIN32)
|
||||
# As of 12.3.1 CUDA Tookit for Windows does not offer a static cublas library
|
||||
set(LLAMA_EXTRA_LIBS ${LLAMA_EXTRA_LIBS} CUDA::cudart_static CUDA::cublas CUDA::cublasLt)
|
||||
else ()
|
||||
set(LLAMA_EXTRA_LIBS ${LLAMA_EXTRA_LIBS} CUDA::cudart_static CUDA::cublas_static CUDA::cublasLt_static)
|
||||
endif()
|
||||
else()
|
||||
set(LLAMA_EXTRA_LIBS ${LLAMA_EXTRA_LIBS} CUDA::cudart CUDA::cublas CUDA::cublasLt)
|
||||
endif()
|
||||
|
||||
if (NOT DEFINED CMAKE_CUDA_ARCHITECTURES)
|
||||
# 52 == lowest CUDA 12 standard
|
||||
# 60 == f16 CUDA intrinsics
|
||||
# 61 == integer CUDA intrinsics
|
||||
# 70 == compute capability at which unrolling a loop in mul_mat_q kernels is faster
|
||||
if (LLAMA_CUDA_F16 OR LLAMA_CUDA_DMMV_F16)
|
||||
set(CMAKE_CUDA_ARCHITECTURES "60;61;70") # needed for f16 CUDA intrinsics
|
||||
else()
|
||||
set(CMAKE_CUDA_ARCHITECTURES "52;61;70") # lowest CUDA 12 standard + lowest for integer intrinsics
|
||||
#set(CMAKE_CUDA_ARCHITECTURES "") # use this to compile much faster, but only F16 models work
|
||||
endif()
|
||||
endif()
|
||||
message(STATUS "Using CUDA architectures: ${CMAKE_CUDA_ARCHITECTURES}")
|
||||
|
||||
else()
|
||||
message(WARNING "cuBLAS not found")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
if (LLAMA_MPI)
|
||||
cmake_minimum_required(VERSION 3.10)
|
||||
find_package(MPI)
|
||||
if (MPI_C_FOUND)
|
||||
message(STATUS "MPI found")
|
||||
set(GGML_HEADERS_MPI ggml-mpi.h)
|
||||
set(GGML_SOURCES_MPI ggml-mpi.c ggml-mpi.h)
|
||||
add_compile_definitions(GGML_USE_MPI)
|
||||
add_compile_definitions(${MPI_C_COMPILE_DEFINITIONS})
|
||||
if (NOT MSVC)
|
||||
add_compile_options(-Wno-cast-qual)
|
||||
endif()
|
||||
set(LLAMA_EXTRA_LIBS ${LLAMA_EXTRA_LIBS} ${MPI_C_LIBRARIES})
|
||||
set(LLAMA_EXTRA_INCLUDES ${LLAMA_EXTRA_INCLUDES} ${MPI_C_INCLUDE_DIRS})
|
||||
# Even if you're only using the C header, C++ programs may bring in MPI
|
||||
# C++ functions, so more linkage is needed
|
||||
if (MPI_CXX_FOUND)
|
||||
set(LLAMA_EXTRA_LIBS ${LLAMA_EXTRA_LIBS} ${MPI_CXX_LIBRARIES})
|
||||
endif()
|
||||
else()
|
||||
message(WARNING "MPI not found")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
if (LLAMA_CLBLAST)
|
||||
find_package(CLBlast)
|
||||
if (CLBlast_FOUND)
|
||||
message(STATUS "CLBlast found")
|
||||
|
||||
set(GGML_HEADERS_OPENCL ggml-opencl.h)
|
||||
set(GGML_SOURCES_OPENCL ggml-opencl.cpp)
|
||||
|
||||
add_compile_definitions(GGML_USE_CLBLAST)
|
||||
|
||||
set(LLAMA_EXTRA_LIBS ${LLAMA_EXTRA_LIBS} clblast)
|
||||
else()
|
||||
message(WARNING "CLBlast not found")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
if (LLAMA_HIPBLAS)
|
||||
list(APPEND CMAKE_PREFIX_PATH /opt/rocm)
|
||||
|
||||
if (NOT ${CMAKE_C_COMPILER_ID} MATCHES "Clang")
|
||||
message(WARNING "Only LLVM is supported for HIP, hint: CC=/opt/rocm/llvm/bin/clang")
|
||||
endif()
|
||||
if (NOT ${CMAKE_CXX_COMPILER_ID} MATCHES "Clang")
|
||||
message(WARNING "Only LLVM is supported for HIP, hint: CXX=/opt/rocm/llvm/bin/clang++")
|
||||
endif()
|
||||
|
||||
find_package(hip)
|
||||
find_package(hipblas)
|
||||
find_package(rocblas)
|
||||
|
||||
if (${hipblas_FOUND} AND ${hip_FOUND})
|
||||
message(STATUS "HIP and hipBLAS found")
|
||||
add_compile_definitions(GGML_USE_HIPBLAS GGML_USE_CUBLAS)
|
||||
add_library(ggml-rocm OBJECT ggml-cuda.cu ggml-cuda.h)
|
||||
if (BUILD_SHARED_LIBS)
|
||||
set_target_properties(ggml-rocm PROPERTIES POSITION_INDEPENDENT_CODE ON)
|
||||
endif()
|
||||
if (LLAMA_CUDA_FORCE_DMMV)
|
||||
target_compile_definitions(ggml-rocm PRIVATE GGML_CUDA_FORCE_DMMV)
|
||||
endif()
|
||||
if (LLAMA_CUDA_FORCE_MMQ)
|
||||
target_compile_definitions(ggml-rocm PRIVATE GGML_CUDA_FORCE_MMQ)
|
||||
endif()
|
||||
target_compile_definitions(ggml-rocm PRIVATE GGML_CUDA_DMMV_X=${LLAMA_CUDA_DMMV_X})
|
||||
target_compile_definitions(ggml-rocm PRIVATE GGML_CUDA_MMV_Y=${LLAMA_CUDA_MMV_Y})
|
||||
target_compile_definitions(ggml-rocm PRIVATE K_QUANTS_PER_ITERATION=${LLAMA_CUDA_KQUANTS_ITER})
|
||||
set_source_files_properties(ggml-cuda.cu PROPERTIES LANGUAGE CXX)
|
||||
target_link_libraries(ggml-rocm PRIVATE hip::device PUBLIC hip::host roc::rocblas roc::hipblas)
|
||||
|
||||
if (LLAMA_STATIC)
|
||||
message(FATAL_ERROR "Static linking not supported for HIP/ROCm")
|
||||
endif()
|
||||
set(LLAMA_EXTRA_LIBS ${LLAMA_EXTRA_LIBS} ggml-rocm)
|
||||
else()
|
||||
message(WARNING "hipBLAS or HIP not found. Try setting CMAKE_PREFIX_PATH=/opt/rocm")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
function(get_flags CCID CCVER)
|
||||
set(C_FLAGS "")
|
||||
set(CXX_FLAGS "")
|
||||
|
||||
if (CCID MATCHES "Clang")
|
||||
set(C_FLAGS -Wunreachable-code-break -Wunreachable-code-return)
|
||||
set(CXX_FLAGS -Wunreachable-code-break -Wunreachable-code-return -Wmissing-prototypes -Wextra-semi)
|
||||
|
||||
if (
|
||||
(CCID STREQUAL "Clang" AND CCVER VERSION_GREATER_EQUAL 3.8.0) OR
|
||||
(CCID STREQUAL "AppleClang" AND CCVER VERSION_GREATER_EQUAL 7.3.0)
|
||||
)
|
||||
set(C_FLAGS ${C_FLAGS} -Wdouble-promotion)
|
||||
endif()
|
||||
elseif (CCID STREQUAL "GNU")
|
||||
set(C_FLAGS -Wdouble-promotion)
|
||||
set(CXX_FLAGS -Wno-array-bounds)
|
||||
|
||||
if (CCVER VERSION_GREATER_EQUAL 7.1.0)
|
||||
set(CXX_FLAGS ${CXX_FLAGS} -Wno-format-truncation)
|
||||
endif()
|
||||
if (CCVER VERSION_GREATER_EQUAL 8.1.0)
|
||||
set(CXX_FLAGS ${CXX_FLAGS} -Wextra-semi)
|
||||
endif()
|
||||
endif()
|
||||
|
||||
set(GF_C_FLAGS ${C_FLAGS} PARENT_SCOPE)
|
||||
set(GF_CXX_FLAGS ${CXX_FLAGS} PARENT_SCOPE)
|
||||
endfunction()
|
||||
|
||||
if (LLAMA_ALL_WARNINGS)
|
||||
if (NOT MSVC)
|
||||
set(WARNING_FLAGS -Wall -Wextra -Wpedantic -Wcast-qual -Wno-unused-function)
|
||||
set(C_FLAGS -Wshadow -Wstrict-prototypes -Wpointer-arith -Wmissing-prototypes
|
||||
-Werror=implicit-int -Werror=implicit-function-declaration)
|
||||
set(CXX_FLAGS -Wmissing-declarations -Wmissing-noreturn)
|
||||
|
||||
set(C_FLAGS ${WARNING_FLAGS} ${C_FLAGS})
|
||||
set(CXX_FLAGS ${WARNING_FLAGS} ${CXX_FLAGS})
|
||||
|
||||
get_flags(${CMAKE_CXX_COMPILER_ID} ${CMAKE_CXX_COMPILER_VERSION})
|
||||
|
||||
add_compile_options("$<$<COMPILE_LANGUAGE:C>:${C_FLAGS};${GF_C_FLAGS}>"
|
||||
"$<$<COMPILE_LANGUAGE:CXX>:${CXX_FLAGS};${GF_CXX_FLAGS}>")
|
||||
else()
|
||||
# todo : msvc
|
||||
set(C_FLAGS "")
|
||||
set(CXX_FLAGS "")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
if (LLAMA_CUBLAS)
|
||||
set(CUDA_FLAGS ${CXX_FLAGS} -use_fast_math)
|
||||
if (NOT MSVC)
|
||||
set(CUDA_FLAGS ${CUDA_FLAGS} -Wno-pedantic)
|
||||
endif()
|
||||
|
||||
if (LLAMA_ALL_WARNINGS AND NOT MSVC)
|
||||
set(NVCC_CMD ${CMAKE_CUDA_COMPILER} .c)
|
||||
if (NOT CMAKE_CUDA_HOST_COMPILER STREQUAL "")
|
||||
set(NVCC_CMD ${NVCC_CMD} -ccbin ${CMAKE_CUDA_HOST_COMPILER})
|
||||
endif()
|
||||
|
||||
execute_process(
|
||||
COMMAND ${NVCC_CMD} -Xcompiler --version
|
||||
OUTPUT_VARIABLE CUDA_CCFULLVER
|
||||
ERROR_QUIET
|
||||
)
|
||||
|
||||
if (NOT CUDA_CCFULLVER MATCHES clang)
|
||||
set(CUDA_CCID "GNU")
|
||||
execute_process(
|
||||
COMMAND ${NVCC_CMD} -Xcompiler "-dumpfullversion -dumpversion"
|
||||
OUTPUT_VARIABLE CUDA_CCVER
|
||||
ERROR_QUIET
|
||||
)
|
||||
else()
|
||||
if (CUDA_CCFULLVER MATCHES Apple)
|
||||
set(CUDA_CCID "AppleClang")
|
||||
else()
|
||||
set(CUDA_CCID "Clang")
|
||||
endif()
|
||||
string(REGEX REPLACE "^.* version ([0-9.]*).*$" "\\1" CUDA_CCVER ${CUDA_CCFULLVER})
|
||||
endif()
|
||||
|
||||
message("-- CUDA host compiler is ${CUDA_CCID} ${CUDA_CCVER}")
|
||||
|
||||
get_flags(${CUDA_CCID} ${CUDA_CCVER})
|
||||
list(JOIN GF_CXX_FLAGS " " CUDA_CXX_FLAGS) # pass host compiler flags as a single argument
|
||||
if (NOT CUDA_CXX_FLAGS STREQUAL "")
|
||||
set(CUDA_FLAGS ${CUDA_FLAGS} -Xcompiler ${CUDA_CXX_FLAGS})
|
||||
endif()
|
||||
endif()
|
||||
|
||||
add_compile_options("$<$<COMPILE_LANGUAGE:CUDA>:${CUDA_FLAGS}>")
|
||||
endif()
|
||||
|
||||
if (WIN32)
|
||||
add_compile_definitions(_CRT_SECURE_NO_WARNINGS)
|
||||
|
||||
if (BUILD_SHARED_LIBS)
|
||||
set(CMAKE_WINDOWS_EXPORT_ALL_SYMBOLS ON)
|
||||
endif()
|
||||
endif()
|
||||
|
||||
if (LLAMA_LTO)
|
||||
include(CheckIPOSupported)
|
||||
check_ipo_supported(RESULT result OUTPUT output)
|
||||
if (result)
|
||||
set(CMAKE_INTERPROCEDURAL_OPTIMIZATION TRUE)
|
||||
else()
|
||||
message(WARNING "IPO is not supported: ${output}")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
# this version of Apple ld64 is buggy
|
||||
execute_process(
|
||||
COMMAND ${CMAKE_C_COMPILER} ${CMAKE_EXE_LINKER_FLAGS} -Wl,-v
|
||||
ERROR_VARIABLE output
|
||||
OUTPUT_QUIET
|
||||
)
|
||||
if (output MATCHES "dyld-1015\.7")
|
||||
add_compile_definitions(HAVE_BUGGY_APPLE_LINKER)
|
||||
endif()
|
||||
|
||||
# Architecture specific
|
||||
# TODO: probably these flags need to be tweaked on some architectures
|
||||
# feel free to update the Makefile for your architecture and send a pull request or issue
|
||||
message(STATUS "CMAKE_SYSTEM_PROCESSOR: ${CMAKE_SYSTEM_PROCESSOR}")
|
||||
if (MSVC)
|
||||
string(TOLOWER "${CMAKE_GENERATOR_PLATFORM}" CMAKE_GENERATOR_PLATFORM_LWR)
|
||||
message(STATUS "CMAKE_GENERATOR_PLATFORM: ${CMAKE_GENERATOR_PLATFORM}")
|
||||
else ()
|
||||
set(CMAKE_GENERATOR_PLATFORM_LWR "")
|
||||
endif ()
|
||||
|
||||
if (NOT MSVC)
|
||||
if (LLAMA_STATIC)
|
||||
add_link_options(-static)
|
||||
if (MINGW)
|
||||
add_link_options(-static-libgcc -static-libstdc++)
|
||||
endif()
|
||||
endif()
|
||||
if (LLAMA_GPROF)
|
||||
add_compile_options(-pg)
|
||||
endif()
|
||||
endif()
|
||||
|
||||
if ((${CMAKE_SYSTEM_PROCESSOR} MATCHES "arm") OR (${CMAKE_SYSTEM_PROCESSOR} MATCHES "aarch64") OR ("${CMAKE_GENERATOR_PLATFORM_LWR}" MATCHES "arm64"))
|
||||
message(STATUS "ARM detected")
|
||||
if (MSVC)
|
||||
add_compile_definitions(__ARM_NEON)
|
||||
add_compile_definitions(__ARM_FEATURE_FMA)
|
||||
add_compile_definitions(__ARM_FEATURE_DOTPROD)
|
||||
# add_compile_definitions(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) # MSVC doesn't support vdupq_n_f16, vld1q_f16, vst1q_f16
|
||||
add_compile_definitions(__aarch64__) # MSVC defines _M_ARM64 instead
|
||||
else()
|
||||
check_cxx_compiler_flag(-mfp16-format=ieee COMPILER_SUPPORTS_FP16_FORMAT_I3E)
|
||||
if (NOT "${COMPILER_SUPPORTS_FP16_FORMAT_I3E}" STREQUAL "")
|
||||
add_compile_options(-mfp16-format=ieee)
|
||||
endif()
|
||||
if (${CMAKE_SYSTEM_PROCESSOR} MATCHES "armv6")
|
||||
# Raspberry Pi 1, Zero
|
||||
add_compile_options(-mfpu=neon-fp-armv8 -mno-unaligned-access)
|
||||
endif()
|
||||
if (${CMAKE_SYSTEM_PROCESSOR} MATCHES "armv7")
|
||||
# Raspberry Pi 2
|
||||
add_compile_options(-mfpu=neon-fp-armv8 -mno-unaligned-access -funsafe-math-optimizations)
|
||||
endif()
|
||||
if (${CMAKE_SYSTEM_PROCESSOR} MATCHES "armv8")
|
||||
# Raspberry Pi 3, 4, Zero 2 (32-bit)
|
||||
add_compile_options(-mno-unaligned-access)
|
||||
endif()
|
||||
endif()
|
||||
elseif (${CMAKE_SYSTEM_PROCESSOR} MATCHES "^(x86_64|i686|AMD64)$" OR "${CMAKE_GENERATOR_PLATFORM_LWR}" MATCHES "^(x86_64|i686|amd64|x64)$" )
|
||||
message(STATUS "x86 detected")
|
||||
if (MSVC)
|
||||
# instruction set detection for MSVC only
|
||||
if (LLAMA_NATIVE)
|
||||
include(cmake/FindSIMD.cmake)
|
||||
endif ()
|
||||
if (LLAMA_AVX512)
|
||||
add_compile_options($<$<COMPILE_LANGUAGE:C>:/arch:AVX512>)
|
||||
add_compile_options($<$<COMPILE_LANGUAGE:CXX>:/arch:AVX512>)
|
||||
# MSVC has no compile-time flags enabling specific
|
||||
# AVX512 extensions, neither it defines the
|
||||
# macros corresponding to the extensions.
|
||||
# Do it manually.
|
||||
if (LLAMA_AVX512_VBMI)
|
||||
add_compile_definitions($<$<COMPILE_LANGUAGE:C>:__AVX512VBMI__>)
|
||||
add_compile_definitions($<$<COMPILE_LANGUAGE:CXX>:__AVX512VBMI__>)
|
||||
endif()
|
||||
if (LLAMA_AVX512_VNNI)
|
||||
add_compile_definitions($<$<COMPILE_LANGUAGE:C>:__AVX512VNNI__>)
|
||||
add_compile_definitions($<$<COMPILE_LANGUAGE:CXX>:__AVX512VNNI__>)
|
||||
endif()
|
||||
elseif (LLAMA_AVX2)
|
||||
add_compile_options($<$<COMPILE_LANGUAGE:C>:/arch:AVX2>)
|
||||
add_compile_options($<$<COMPILE_LANGUAGE:CXX>:/arch:AVX2>)
|
||||
elseif (LLAMA_AVX)
|
||||
add_compile_options($<$<COMPILE_LANGUAGE:C>:/arch:AVX>)
|
||||
add_compile_options($<$<COMPILE_LANGUAGE:CXX>:/arch:AVX>)
|
||||
endif()
|
||||
else()
|
||||
if (LLAMA_NATIVE)
|
||||
add_compile_options(-march=native)
|
||||
endif()
|
||||
if (LLAMA_F16C)
|
||||
add_compile_options(-mf16c)
|
||||
endif()
|
||||
if (LLAMA_FMA)
|
||||
add_compile_options(-mfma)
|
||||
endif()
|
||||
if (LLAMA_AVX)
|
||||
add_compile_options(-mavx)
|
||||
endif()
|
||||
if (LLAMA_AVX2)
|
||||
add_compile_options(-mavx2)
|
||||
endif()
|
||||
if (LLAMA_AVX512)
|
||||
add_compile_options(-mavx512f)
|
||||
add_compile_options(-mavx512bw)
|
||||
endif()
|
||||
if (LLAMA_AVX512_VBMI)
|
||||
add_compile_options(-mavx512vbmi)
|
||||
endif()
|
||||
if (LLAMA_AVX512_VNNI)
|
||||
add_compile_options(-mavx512vnni)
|
||||
endif()
|
||||
endif()
|
||||
elseif (${CMAKE_SYSTEM_PROCESSOR} MATCHES "ppc64")
|
||||
message(STATUS "PowerPC detected")
|
||||
if (${CMAKE_SYSTEM_PROCESSOR} MATCHES "ppc64le")
|
||||
add_compile_options(-mcpu=powerpc64le)
|
||||
else()
|
||||
add_compile_options(-mcpu=native -mtune=native)
|
||||
#TODO: Add targets for Power8/Power9 (Altivec/VSX) and Power10(MMA) and query for big endian systems (ppc64/le/be)
|
||||
endif()
|
||||
else()
|
||||
message(STATUS "Unknown architecture")
|
||||
endif()
|
||||
|
||||
if (MINGW)
|
||||
# Target Windows 8 for PrefetchVirtualMemory
|
||||
add_compile_definitions(_WIN32_WINNT=0x602)
|
||||
endif()
|
||||
|
||||
#
|
||||
# build the library
|
||||
# POSIX conformance
|
||||
#
|
||||
|
||||
add_subdirectory(src)
|
||||
# clock_gettime came in POSIX.1b (1993)
|
||||
# CLOCK_MONOTONIC came in POSIX.1-2001 / SUSv3 as optional
|
||||
# posix_memalign came in POSIX.1-2001 / SUSv3
|
||||
# M_PI is an XSI extension since POSIX.1-2001 / SUSv3, came in XPG1 (1985)
|
||||
add_compile_definitions(_XOPEN_SOURCE=600)
|
||||
|
||||
# Somehow in OpenBSD whenever POSIX conformance is specified
|
||||
# some string functions rely on locale_t availability,
|
||||
# which was introduced in POSIX.1-2008, forcing us to go higher
|
||||
if (CMAKE_SYSTEM_NAME MATCHES "OpenBSD")
|
||||
remove_definitions(-D_XOPEN_SOURCE=600)
|
||||
add_compile_definitions(_XOPEN_SOURCE=700)
|
||||
endif()
|
||||
|
||||
# Data types, macros and functions related to controlling CPU affinity and
|
||||
# some memory allocation are available on Linux through GNU extensions in libc
|
||||
if (CMAKE_SYSTEM_NAME MATCHES "Linux")
|
||||
add_compile_definitions(_GNU_SOURCE)
|
||||
endif()
|
||||
|
||||
# RLIMIT_MEMLOCK came in BSD, is not specified in POSIX.1,
|
||||
# and on macOS its availability depends on enabling Darwin extensions
|
||||
# similarly on DragonFly, enabling BSD extensions is necessary
|
||||
if (
|
||||
CMAKE_SYSTEM_NAME MATCHES "Darwin" OR
|
||||
CMAKE_SYSTEM_NAME MATCHES "iOS" OR
|
||||
CMAKE_SYSTEM_NAME MATCHES "tvOS" OR
|
||||
CMAKE_SYSTEM_NAME MATCHES "DragonFly"
|
||||
)
|
||||
add_compile_definitions(_DARWIN_C_SOURCE)
|
||||
endif()
|
||||
|
||||
# alloca is a non-standard interface that is not visible on BSDs when
|
||||
# POSIX conformance is specified, but not all of them provide a clean way
|
||||
# to enable it in such cases
|
||||
if (CMAKE_SYSTEM_NAME MATCHES "FreeBSD")
|
||||
add_compile_definitions(__BSD_VISIBLE)
|
||||
endif()
|
||||
if (CMAKE_SYSTEM_NAME MATCHES "NetBSD")
|
||||
add_compile_definitions(_NETBSD_SOURCE)
|
||||
endif()
|
||||
if (CMAKE_SYSTEM_NAME MATCHES "OpenBSD")
|
||||
add_compile_definitions(_BSD_SOURCE)
|
||||
endif()
|
||||
|
||||
#
|
||||
# utils, programs, examples and tests
|
||||
# libraries
|
||||
#
|
||||
|
||||
if (LLAMA_BUILD_COMMON)
|
||||
add_subdirectory(common)
|
||||
# ggml
|
||||
|
||||
if (GGML_USE_CPU_HBM)
|
||||
add_definitions(-DGGML_USE_CPU_HBM)
|
||||
find_library(memkind memkind REQUIRED)
|
||||
endif()
|
||||
|
||||
if (LLAMA_BUILD_COMMON AND LLAMA_BUILD_TESTS AND NOT CMAKE_JS_VERSION)
|
||||
include(CTest)
|
||||
add_subdirectory(tests)
|
||||
add_library(ggml OBJECT
|
||||
ggml.c
|
||||
ggml.h
|
||||
ggml-alloc.c
|
||||
ggml-alloc.h
|
||||
ggml-backend.c
|
||||
ggml-backend.h
|
||||
ggml-quants.c
|
||||
ggml-quants.h
|
||||
${GGML_SOURCES_CUDA} ${GGML_HEADERS_CUDA}
|
||||
${GGML_SOURCES_OPENCL} ${GGML_HEADERS_OPENCL}
|
||||
${GGML_SOURCES_METAL} ${GGML_HEADERS_METAL}
|
||||
${GGML_SOURCES_MPI} ${GGML_HEADERS_MPI}
|
||||
${GGML_SOURCES_EXTRA} ${GGML_HEADERS_EXTRA}
|
||||
)
|
||||
|
||||
target_include_directories(ggml PUBLIC . ${LLAMA_EXTRA_INCLUDES})
|
||||
target_compile_features(ggml PUBLIC c_std_11) # don't bump
|
||||
target_link_libraries(ggml PUBLIC Threads::Threads ${LLAMA_EXTRA_LIBS})
|
||||
if (GGML_USE_CPU_HBM)
|
||||
target_link_libraries(ggml PUBLIC memkind)
|
||||
endif()
|
||||
|
||||
if (LLAMA_BUILD_COMMON AND LLAMA_BUILD_EXAMPLES)
|
||||
add_subdirectory(examples)
|
||||
add_subdirectory(pocs)
|
||||
add_library(ggml_static STATIC $<TARGET_OBJECTS:ggml>)
|
||||
if (BUILD_SHARED_LIBS)
|
||||
set_target_properties(ggml PROPERTIES POSITION_INDEPENDENT_CODE ON)
|
||||
add_library(ggml_shared SHARED $<TARGET_OBJECTS:ggml>)
|
||||
target_link_libraries(ggml_shared PUBLIC Threads::Threads ${LLAMA_EXTRA_LIBS})
|
||||
install(TARGETS ggml_shared LIBRARY)
|
||||
endif()
|
||||
|
||||
# llama
|
||||
|
||||
add_library(llama
|
||||
llama.cpp
|
||||
llama.h
|
||||
)
|
||||
|
||||
target_include_directories(llama PUBLIC .)
|
||||
target_compile_features(llama PUBLIC cxx_std_11) # don't bump
|
||||
target_link_libraries(llama PRIVATE
|
||||
ggml
|
||||
${LLAMA_EXTRA_LIBS}
|
||||
)
|
||||
|
||||
if (BUILD_SHARED_LIBS)
|
||||
set_target_properties(llama PROPERTIES POSITION_INDEPENDENT_CODE ON)
|
||||
target_compile_definitions(llama PRIVATE LLAMA_SHARED LLAMA_BUILD)
|
||||
if (LLAMA_METAL)
|
||||
set_target_properties(llama PROPERTIES RESOURCE "${CMAKE_CURRENT_SOURCE_DIR}/ggml-metal.metal")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
|
||||
#
|
||||
# install
|
||||
#
|
||||
|
@ -181,43 +770,46 @@ endif()
|
|||
include(GNUInstallDirs)
|
||||
include(CMakePackageConfigHelpers)
|
||||
|
||||
set(LLAMA_BUILD_NUMBER ${BUILD_NUMBER})
|
||||
set(LLAMA_BUILD_COMMIT ${BUILD_COMMIT})
|
||||
set(LLAMA_INCLUDE_INSTALL_DIR ${CMAKE_INSTALL_INCLUDEDIR}
|
||||
CACHE PATH "Location of header files")
|
||||
set(LLAMA_LIB_INSTALL_DIR ${CMAKE_INSTALL_LIBDIR}
|
||||
CACHE PATH "Location of library files")
|
||||
set(LLAMA_BIN_INSTALL_DIR ${CMAKE_INSTALL_BINDIR}
|
||||
CACHE PATH "Location of binary files")
|
||||
set(LLAMA_BUILD_NUMBER ${BUILD_NUMBER})
|
||||
set(LLAMA_BUILD_COMMIT ${BUILD_COMMIT})
|
||||
set(LLAMA_INSTALL_VERSION 0.0.${BUILD_NUMBER})
|
||||
|
||||
set(LLAMA_INCLUDE_INSTALL_DIR ${CMAKE_INSTALL_INCLUDEDIR} CACHE PATH "Location of header files")
|
||||
set(LLAMA_LIB_INSTALL_DIR ${CMAKE_INSTALL_LIBDIR} CACHE PATH "Location of library files")
|
||||
set(LLAMA_BIN_INSTALL_DIR ${CMAKE_INSTALL_BINDIR} CACHE PATH "Location of binary files")
|
||||
|
||||
set(LLAMA_PUBLIC_HEADERS
|
||||
${CMAKE_CURRENT_SOURCE_DIR}/include/llama.h
|
||||
${CMAKE_CURRENT_SOURCE_DIR}/include/llama-cpp.h)
|
||||
|
||||
set_target_properties(llama
|
||||
PROPERTIES
|
||||
PUBLIC_HEADER "${LLAMA_PUBLIC_HEADERS}")
|
||||
|
||||
install(TARGETS llama LIBRARY PUBLIC_HEADER)
|
||||
get_directory_property(LLAMA_TRANSIENT_DEFINES COMPILE_DEFINITIONS)
|
||||
|
||||
configure_package_config_file(
|
||||
${CMAKE_CURRENT_SOURCE_DIR}/cmake/llama-config.cmake.in
|
||||
${CMAKE_CURRENT_BINARY_DIR}/llama-config.cmake
|
||||
INSTALL_DESTINATION ${CMAKE_INSTALL_LIBDIR}/cmake/llama
|
||||
${CMAKE_CURRENT_SOURCE_DIR}/scripts/LlamaConfig.cmake.in
|
||||
${CMAKE_CURRENT_BINARY_DIR}/LlamaConfig.cmake
|
||||
INSTALL_DESTINATION ${CMAKE_INSTALL_LIBDIR}/cmake/Llama
|
||||
PATH_VARS LLAMA_INCLUDE_INSTALL_DIR
|
||||
LLAMA_LIB_INSTALL_DIR
|
||||
LLAMA_BIN_INSTALL_DIR )
|
||||
|
||||
write_basic_package_version_file(
|
||||
${CMAKE_CURRENT_BINARY_DIR}/llama-version.cmake
|
||||
${CMAKE_CURRENT_BINARY_DIR}/LlamaConfigVersion.cmake
|
||||
VERSION ${LLAMA_INSTALL_VERSION}
|
||||
COMPATIBILITY SameMajorVersion)
|
||||
|
||||
install(FILES ${CMAKE_CURRENT_BINARY_DIR}/llama-config.cmake
|
||||
${CMAKE_CURRENT_BINARY_DIR}/llama-version.cmake
|
||||
DESTINATION ${CMAKE_INSTALL_LIBDIR}/cmake/llama)
|
||||
install(FILES ${CMAKE_CURRENT_BINARY_DIR}/LlamaConfig.cmake
|
||||
${CMAKE_CURRENT_BINARY_DIR}/LlamaConfigVersion.cmake
|
||||
DESTINATION ${CMAKE_INSTALL_LIBDIR}/cmake/Llama)
|
||||
|
||||
set(GGML_PUBLIC_HEADERS "ggml.h"
|
||||
"${GGML_HEADERS_CUDA}" "${GGML_HEADERS_OPENCL}"
|
||||
"${GGML_HEADERS_METAL}" "${GGML_HEADERS_MPI}" "${GGML_HEADERS_EXTRA}")
|
||||
|
||||
set_target_properties(ggml PROPERTIES PUBLIC_HEADER "${GGML_PUBLIC_HEADERS}")
|
||||
install(TARGETS ggml PUBLIC_HEADER)
|
||||
|
||||
set_target_properties(llama PROPERTIES PUBLIC_HEADER ${CMAKE_CURRENT_SOURCE_DIR}/llama.h)
|
||||
install(TARGETS llama LIBRARY PUBLIC_HEADER)
|
||||
|
||||
install(
|
||||
FILES convert_hf_to_gguf.py
|
||||
FILES convert.py
|
||||
PERMISSIONS
|
||||
OWNER_READ
|
||||
OWNER_WRITE
|
||||
|
@ -227,10 +819,40 @@ install(
|
|||
WORLD_READ
|
||||
WORLD_EXECUTE
|
||||
DESTINATION ${CMAKE_INSTALL_BINDIR})
|
||||
install(
|
||||
FILES convert-lora-to-ggml.py
|
||||
PERMISSIONS
|
||||
OWNER_READ
|
||||
OWNER_WRITE
|
||||
OWNER_EXECUTE
|
||||
GROUP_READ
|
||||
GROUP_EXECUTE
|
||||
WORLD_READ
|
||||
WORLD_EXECUTE
|
||||
DESTINATION ${CMAKE_INSTALL_BINDIR})
|
||||
if (LLAMA_METAL)
|
||||
install(
|
||||
FILES ggml-metal.metal
|
||||
PERMISSIONS
|
||||
OWNER_READ
|
||||
OWNER_WRITE
|
||||
GROUP_READ
|
||||
WORLD_READ
|
||||
DESTINATION ${CMAKE_INSTALL_BINDIR})
|
||||
endif()
|
||||
|
||||
configure_file(cmake/llama.pc.in
|
||||
"${CMAKE_CURRENT_BINARY_DIR}/llama.pc"
|
||||
@ONLY)
|
||||
#
|
||||
# programs, examples and tests
|
||||
#
|
||||
|
||||
install(FILES "${CMAKE_CURRENT_BINARY_DIR}/llama.pc"
|
||||
DESTINATION ${CMAKE_INSTALL_LIBDIR}/pkgconfig)
|
||||
add_subdirectory(common)
|
||||
|
||||
if (LLAMA_BUILD_TESTS AND NOT CMAKE_JS_VERSION)
|
||||
include(CTest)
|
||||
add_subdirectory(tests)
|
||||
endif ()
|
||||
|
||||
if (LLAMA_BUILD_EXAMPLES)
|
||||
add_subdirectory(examples)
|
||||
add_subdirectory(pocs)
|
||||
endif()
|
||||
|
|
|
@ -1,97 +0,0 @@
|
|||
{
|
||||
"version": 4,
|
||||
"configurePresets": [
|
||||
{
|
||||
"name": "base",
|
||||
"hidden": true,
|
||||
"generator": "Ninja",
|
||||
"binaryDir": "${sourceDir}/build-${presetName}",
|
||||
"cacheVariables": {
|
||||
"CMAKE_EXPORT_COMPILE_COMMANDS": "ON",
|
||||
"CMAKE_INSTALL_RPATH": "$ORIGIN;$ORIGIN/.."
|
||||
}
|
||||
},
|
||||
{
|
||||
"name": "sycl-base",
|
||||
"hidden": true,
|
||||
"generator": "Ninja",
|
||||
"binaryDir": "${sourceDir}/build-${presetName}",
|
||||
"cacheVariables": {
|
||||
"CMAKE_EXPORT_COMPILE_COMMANDS": "ON",
|
||||
"CMAKE_CXX_COMPILER": "icx",
|
||||
"CMAKE_C_COMPILER": "cl",
|
||||
"GGML_SYCL": "ON",
|
||||
"CMAKE_INSTALL_RPATH": "$ORIGIN;$ORIGIN/.."
|
||||
}
|
||||
},
|
||||
{ "name": "debug", "hidden": true, "cacheVariables": { "CMAKE_BUILD_TYPE": "Debug" } },
|
||||
{ "name": "release", "hidden": true, "cacheVariables": { "CMAKE_BUILD_TYPE": "Release" } },
|
||||
{ "name": "reldbg", "hidden": true, "cacheVariables": { "CMAKE_BUILD_TYPE": "RelWithDebInfo" } },
|
||||
{ "name": "static", "hidden": true, "cacheVariables": { "GGML_STATIC": "ON" } },
|
||||
{ "name": "sycl_f16", "hidden": true, "cacheVariables": { "GGML_SYCL_F16": "ON" } },
|
||||
{ "name": "vulkan", "hidden": true, "cacheVariables": { "GGML_VULKAN": "ON" } },
|
||||
|
||||
{
|
||||
"name": "x64-windows-llvm", "hidden": true,
|
||||
"cacheVariables": {
|
||||
"CMAKE_TOOLCHAIN_FILE": "${sourceDir}/cmake/x64-windows-llvm.cmake"
|
||||
}
|
||||
},
|
||||
|
||||
{
|
||||
"name": "arm64-windows-msvc", "hidden": true,
|
||||
"architecture": { "value": "arm64", "strategy": "external" },
|
||||
"toolset": { "value": "host=x64", "strategy": "external" },
|
||||
"cacheVariables": {
|
||||
"CMAKE_TOOLCHAIN_FILE": "${sourceDir}/cmake/arm64-windows-msvc.cmake"
|
||||
}
|
||||
},
|
||||
|
||||
{
|
||||
"name": "arm64-windows-llvm", "hidden": true,
|
||||
"architecture": { "value": "arm64", "strategy": "external" },
|
||||
"toolset": { "value": "host=x64", "strategy": "external" },
|
||||
"cacheVariables": {
|
||||
"CMAKE_TOOLCHAIN_FILE": "${sourceDir}/cmake/arm64-windows-llvm.cmake"
|
||||
}
|
||||
},
|
||||
|
||||
{
|
||||
"name": "arm64-apple-clang", "hidden": true,
|
||||
"architecture": { "value": "arm64", "strategy": "external" },
|
||||
"toolset": { "value": "host=x64", "strategy": "external" },
|
||||
"cacheVariables": {
|
||||
"CMAKE_TOOLCHAIN_FILE": "${sourceDir}/cmake/arm64-apple-clang.cmake"
|
||||
}
|
||||
},
|
||||
|
||||
{ "name": "arm64-windows-llvm-debug", "inherits": [ "base", "arm64-windows-llvm", "debug" ] },
|
||||
{ "name": "arm64-windows-llvm-release", "inherits": [ "base", "arm64-windows-llvm", "reldbg" ] },
|
||||
{ "name": "arm64-windows-llvm+static-release", "inherits": [ "base", "arm64-windows-llvm", "reldbg", "static" ] },
|
||||
|
||||
{ "name": "arm64-apple-clang-debug", "inherits": [ "base", "arm64-apple-clang", "debug" ] },
|
||||
{ "name": "arm64-apple-clang-release", "inherits": [ "base", "arm64-apple-clang", "reldbg" ] },
|
||||
{ "name": "arm64-apple-clang+static-release", "inherits": [ "base", "arm64-apple-clang", "reldbg", "static" ] },
|
||||
|
||||
{ "name": "arm64-windows-msvc-debug", "inherits": [ "base", "arm64-windows-msvc", "debug" ] },
|
||||
{ "name": "arm64-windows-msvc-release", "inherits": [ "base", "arm64-windows-msvc", "reldbg" ] },
|
||||
{ "name": "arm64-windows-msvc+static-release", "inherits": [ "base", "arm64-windows-msvc", "reldbg", "static" ] },
|
||||
|
||||
{ "name": "x64-windows-llvm-debug", "inherits": [ "base", "x64-windows-llvm", "debug" ] },
|
||||
{ "name": "x64-windows-llvm-release", "inherits": [ "base", "x64-windows-llvm", "release" ] },
|
||||
{ "name": "x64-windows-llvm-reldbg", "inherits": [ "base", "x64-windows-llvm", "reldbg" ] },
|
||||
{ "name": "x64-windows-llvm+static-release", "inherits": [ "base", "x64-windows-llvm", "reldbg", "static" ] },
|
||||
|
||||
{ "name": "x64-windows-msvc-debug", "inherits": [ "base", "debug" ] },
|
||||
{ "name": "x64-windows-msvc-release", "inherits": [ "base", "reldbg" ] },
|
||||
{ "name": "x64-windows-msvc+static-release", "inherits": [ "base", "reldbg", "static" ] },
|
||||
|
||||
{ "name": "x64-windows-sycl-debug", "inherits": [ "sycl-base", "debug" ] },
|
||||
{ "name": "x64-windows-sycl-debug-f16", "inherits": [ "sycl-base", "debug", "sycl_f16" ] },
|
||||
{ "name": "x64-windows-sycl-release", "inherits": [ "sycl-base", "release" ] },
|
||||
{ "name": "x64-windows-sycl-release-f16", "inherits": [ "sycl-base", "release", "sycl_f16" ] },
|
||||
|
||||
{ "name": "x64-windows-vulkan-debug", "inherits": [ "base", "vulkan", "debug" ] },
|
||||
{ "name": "x64-windows-vulkan-release", "inherits": [ "base", "vulkan", "release" ] }
|
||||
]
|
||||
}
|
11
CODEOWNERS
11
CODEOWNERS
|
@ -1,11 +0,0 @@
|
|||
# collaborators can optionally add themselves here to indicate their availability for reviewing related PRs
|
||||
|
||||
/ci/ @ggerganov
|
||||
/.devops/*.Dockerfile @ngxson
|
||||
/examples/server/ @ngxson
|
||||
/ggml/src/ggml-cuda/fattn* @JohannesGaessler
|
||||
/ggml/src/ggml-cuda/mmq.* @JohannesGaessler
|
||||
/ggml/src/ggml-cuda/mmv.* @JohannesGaessler
|
||||
/ggml/src/ggml-cuda/mmvq.* @JohannesGaessler
|
||||
/ggml/src/ggml-opt.cpp @JohannesGaessler
|
||||
/ggml/src/gguf.cpp @JohannesGaessler
|
125
CONTRIBUTING.md
125
CONTRIBUTING.md
|
@ -1,125 +0,0 @@
|
|||
# Pull requests (for contributors)
|
||||
|
||||
- Test your changes:
|
||||
- Execute [the full CI locally on your machine](ci/README.md) before publishing
|
||||
- Verify that the perplexity and the performance are not affected negatively by your changes (use `llama-perplexity` and `llama-bench`)
|
||||
- If you modified the `ggml` source, run the `test-backend-ops` tool to check whether different backend implementations of the `ggml` operators produce consistent results (this requires access to at least two different `ggml` backends)
|
||||
- If you modified a `ggml` operator or added a new one, add the corresponding test cases to `test-backend-ops`
|
||||
- Consider allowing write access to your branch for faster reviews, as reviewers can push commits directly
|
||||
- If your PR becomes stale, don't hesitate to ping the maintainers in the comments
|
||||
|
||||
# Pull requests (for collaborators)
|
||||
|
||||
- Squash-merge PRs
|
||||
- Use the following format for the squashed commit title: `<module> : <commit title> (#<issue_number>)`. For example: `utils : fix typo in utils.py (#1234)`
|
||||
- Optionally pick a `<module>` from here: https://github.com/ggerganov/llama.cpp/wiki/Modules
|
||||
- Consider adding yourself to [CODEOWNERS](CODEOWNERS)
|
||||
|
||||
# Coding guidelines
|
||||
|
||||
- Avoid adding third-party dependencies, extra files, extra headers, etc.
|
||||
- Always consider cross-compatibility with other operating systems and architectures
|
||||
- Avoid fancy-looking modern STL constructs, use basic `for` loops, avoid templates, keep it simple
|
||||
- Vertical alignment makes things more readable and easier to batch edit
|
||||
- Clean-up any trailing whitespaces, use 4 spaces for indentation, brackets on the same line, `void * ptr`, `int & a`
|
||||
- Use sized integer types such as `int32_t` in the public API, e.g. `size_t` may also be appropriate for allocation sizes or byte offsets
|
||||
- Declare structs with `struct foo {}` instead of `typedef struct foo {} foo`
|
||||
- In C++ code omit optional `struct` and `enum` keyword whenever they are not necessary
|
||||
```cpp
|
||||
// OK
|
||||
llama_context * ctx;
|
||||
const llama_rope_type rope_type;
|
||||
|
||||
// not OK
|
||||
struct llama_context * ctx;
|
||||
const enum llama_rope_type rope_type;
|
||||
```
|
||||
|
||||
_(NOTE: this guideline is yet to be applied to the `llama.cpp` codebase. New code should follow this guideline.)_
|
||||
|
||||
- Try to follow the existing patterns in the code (indentation, spaces, etc.). In case of doubt use `clang-format` to format the added code
|
||||
- For anything not covered in the current guidelines, refer to the [C++ Core Guidelines](https://isocpp.github.io/CppCoreGuidelines/CppCoreGuidelines)
|
||||
- Tensors store data in row-major order. We refer to dimension 0 as columns, 1 as rows, 2 as matrices
|
||||
- Matrix multiplication is unconventional: [`C = ggml_mul_mat(ctx, A, B)`](https://github.com/ggerganov/llama.cpp/blob/880e352277fc017df4d5794f0c21c44e1eae2b84/ggml.h#L1058-L1064) means $C^T = A B^T \Leftrightarrow C = B A^T.$
|
||||
|
||||

|
||||
|
||||
# Naming guidelines
|
||||
|
||||
- Use `snake_case` for function, variable and type names
|
||||
- Naming usually optimizes for longest common prefix (see https://github.com/ggerganov/ggml/pull/302#discussion_r1243240963)
|
||||
|
||||
```cpp
|
||||
// not OK
|
||||
int small_number;
|
||||
int big_number;
|
||||
|
||||
// OK
|
||||
int number_small;
|
||||
int number_big;
|
||||
```
|
||||
|
||||
- Enum values are always in upper case and prefixed with the enum name
|
||||
|
||||
```cpp
|
||||
enum llama_vocab_type {
|
||||
LLAMA_VOCAB_TYPE_NONE = 0,
|
||||
LLAMA_VOCAB_TYPE_SPM = 1,
|
||||
LLAMA_VOCAB_TYPE_BPE = 2,
|
||||
LLAMA_VOCAB_TYPE_WPM = 3,
|
||||
LLAMA_VOCAB_TYPE_UGM = 4,
|
||||
LLAMA_VOCAB_TYPE_RWKV = 5,
|
||||
};
|
||||
```
|
||||
|
||||
- The general naming pattern is `<class>_<method>`, with `<method>` being `<action>_<noun>`
|
||||
|
||||
```cpp
|
||||
llama_model_init(); // class: "llama_model", method: "init"
|
||||
llama_sampler_chain_remove(); // class: "llama_sampler_chain", method: "remove"
|
||||
llama_sampler_get_seed(); // class: "llama_sampler", method: "get_seed"
|
||||
llama_set_embeddings(); // class: "llama_context", method: "set_embeddings"
|
||||
llama_n_threads(); // class: "llama_context", method: "n_threads"
|
||||
llama_adapter_lora_free(); // class: "llama_adapter_lora", method: "free"
|
||||
```
|
||||
|
||||
- The `get` `<action>` can be omitted
|
||||
- The `<noun>` can be omitted if not necessary
|
||||
- The `_context` suffix of the `<class>` is optional. Use it to disambiguate symbols when needed
|
||||
- Use `init`/`free` for constructor/destructor `<action>`
|
||||
|
||||
- Use the `_t` suffix when a type is supposed to be opaque to the user - it's not relevant to them if it is a struct or anything else
|
||||
|
||||
```cpp
|
||||
typedef struct llama_context * llama_context_t;
|
||||
|
||||
enum llama_pooling_type llama_pooling_type(const llama_context_t ctx);
|
||||
```
|
||||
|
||||
_(NOTE: this guideline is yet to be applied to the `llama.cpp` codebase. New code should follow this guideline)_
|
||||
|
||||
- C/C++ filenames are all lowercase with dashes. Headers use the `.h` extension. Source files use the `.c` or `.cpp` extension
|
||||
- Python filenames are all lowercase with underscores
|
||||
|
||||
- _(TODO: abbreviations usage)_
|
||||
|
||||
# Preprocessor directives
|
||||
|
||||
- _(TODO: add guidelines with examples and apply them to the codebase)_
|
||||
|
||||
```cpp
|
||||
#ifdef FOO
|
||||
#endif // FOO
|
||||
```
|
||||
|
||||
# Documentation
|
||||
|
||||
- Documentation is a community effort
|
||||
- When you need to look into the source code to figure out how to use an API consider adding a short summary to the header file for future reference
|
||||
- When you notice incorrect or outdated documentation, please update it
|
||||
|
||||
# Resources
|
||||
|
||||
The Github issues, PRs and discussions contain a lot of information that can be useful to get familiar with the codebase. For convenience, some of the more important information is referenced from Github projects:
|
||||
|
||||
https://github.com/ggerganov/llama.cpp/projects
|
2
LICENSE
2
LICENSE
|
@ -1,6 +1,6 @@
|
|||
MIT License
|
||||
|
||||
Copyright (c) 2023-2024 The ggml authors
|
||||
Copyright (c) 2023 Georgi Gerganov
|
||||
|
||||
Permission is hereby granted, free of charge, to any person obtaining a copy
|
||||
of this software and associated documentation files (the "Software"), to deal
|
||||
|
|
|
@ -14,6 +14,37 @@ let package = Package(
|
|||
.library(name: "llama", targets: ["llama"]),
|
||||
],
|
||||
targets: [
|
||||
.systemLibrary(name: "llama", pkgConfig: "llama"),
|
||||
]
|
||||
.target(
|
||||
name: "llama",
|
||||
path: ".",
|
||||
exclude: [],
|
||||
sources: [
|
||||
"ggml.c",
|
||||
"llama.cpp",
|
||||
"ggml-alloc.c",
|
||||
"ggml-backend.c",
|
||||
"ggml-quants.c",
|
||||
"ggml-metal.m",
|
||||
],
|
||||
resources: [
|
||||
.process("ggml-metal.metal")
|
||||
],
|
||||
publicHeadersPath: "spm-headers",
|
||||
cSettings: [
|
||||
.unsafeFlags(["-Wno-shorten-64-to-32", "-O3", "-DNDEBUG"]),
|
||||
.define("GGML_USE_ACCELERATE"),
|
||||
.unsafeFlags(["-fno-objc-arc"]),
|
||||
.define("GGML_USE_METAL"),
|
||||
// NOTE: NEW_LAPACK will required iOS version 16.4+
|
||||
// We should consider add this in the future when we drop support for iOS 14
|
||||
// (ref: ref: https://developer.apple.com/documentation/accelerate/1513264-cblas_sgemm?language=objc)
|
||||
// .define("ACCELERATE_NEW_LAPACK"),
|
||||
// .define("ACCELERATE_LAPACK_ILP64")
|
||||
],
|
||||
linkerSettings: [
|
||||
.linkedFramework("Accelerate")
|
||||
]
|
||||
)
|
||||
],
|
||||
cxxLanguageStandard: .cxx11
|
||||
)
|
||||
|
|
67
SECURITY.md
67
SECURITY.md
|
@ -1,67 +0,0 @@
|
|||
# Security Policy
|
||||
|
||||
- [**Using llama.cpp securely**](#using-llamacpp-securely)
|
||||
- [Untrusted models](#untrusted-models)
|
||||
- [Untrusted inputs](#untrusted-inputs)
|
||||
- [Data privacy](#data-privacy)
|
||||
- [Untrusted environments or networks](#untrusted-environments-or-networks)
|
||||
- [Multi-Tenant environments](#multi-tenant-environments)
|
||||
- [**Reporting a vulnerability**](#reporting-a-vulnerability)
|
||||
|
||||
## Using llama.cpp securely
|
||||
|
||||
### Untrusted models
|
||||
Be careful when running untrusted models. This classification includes models created by unknown developers or utilizing data obtained from unknown sources.
|
||||
|
||||
*Always execute untrusted models within a secure, isolated environment such as a sandbox* (e.g., containers, virtual machines). This helps protect your system from potentially malicious code.
|
||||
|
||||
> [!NOTE]
|
||||
> The trustworthiness of a model is not binary. You must always determine the proper level of caution depending on the specific model and how it matches your use case and risk tolerance.
|
||||
|
||||
### Untrusted inputs
|
||||
|
||||
Some models accept various input formats (text, images, audio, etc.). The libraries converting these inputs have varying security levels, so it's crucial to isolate the model and carefully pre-process inputs to mitigate script injection risks.
|
||||
|
||||
For maximum security when handling untrusted inputs, you may need to employ the following:
|
||||
|
||||
* Sandboxing: Isolate the environment where the inference happens.
|
||||
* Pre-analysis: Check how the model performs by default when exposed to prompt injection (e.g. using [fuzzing for prompt injection](https://github.com/FonduAI/awesome-prompt-injection?tab=readme-ov-file#tools)). This will give you leads on how hard you will have to work on the next topics.
|
||||
* Updates: Keep both LLaMA C++ and your libraries updated with the latest security patches.
|
||||
* Input Sanitation: Before feeding data to the model, sanitize inputs rigorously. This involves techniques such as:
|
||||
* Validation: Enforce strict rules on allowed characters and data types.
|
||||
* Filtering: Remove potentially malicious scripts or code fragments.
|
||||
* Encoding: Convert special characters into safe representations.
|
||||
* Verification: Run tooling that identifies potential script injections (e.g. [models that detect prompt injection attempts](https://python.langchain.com/docs/guides/safety/hugging_face_prompt_injection)).
|
||||
|
||||
### Data privacy
|
||||
|
||||
To protect sensitive data from potential leaks or unauthorized access, it is crucial to sandbox the model execution. This means running the model in a secure, isolated environment, which helps mitigate many attack vectors.
|
||||
|
||||
### Untrusted environments or networks
|
||||
|
||||
If you can't run your models in a secure and isolated environment or if it must be exposed to an untrusted network, make sure to take the following security precautions:
|
||||
* Confirm the hash of any downloaded artifact (e.g. pre-trained model weights) matches a known-good value
|
||||
* Encrypt your data if sending it over the network.
|
||||
|
||||
### Multi-Tenant environments
|
||||
|
||||
If you intend to run multiple models in parallel with shared memory, it is your responsibility to ensure the models do not interact or access each other's data. The primary areas of concern are tenant isolation, resource allocation, model sharing and hardware attacks.
|
||||
|
||||
1. Tenant Isolation: Models should run separately with strong isolation methods to prevent unwanted data access. Separating networks is crucial for isolation, as it prevents unauthorized access to data or models and malicious users from sending graphs to execute under another tenant's identity.
|
||||
|
||||
2. Resource Allocation: A denial of service caused by one model can impact the overall system health. Implement safeguards like rate limits, access controls, and health monitoring.
|
||||
|
||||
3. Model Sharing: In a multitenant model sharing design, tenants and users must understand the security risks of running code provided by others. Since there are no reliable methods to detect malicious models, sandboxing the model execution is the recommended approach to mitigate the risk.
|
||||
|
||||
4. Hardware Attacks: GPUs or TPUs can also be attacked. [Researches](https://scholar.google.com/scholar?q=gpu+side+channel) has shown that side channel attacks on GPUs are possible, which can make data leak from other models or processes running on the same system at the same time.
|
||||
|
||||
## Reporting a vulnerability
|
||||
|
||||
Beware that none of the topics under [Using llama.cpp securely](#using-llamacpp-securely) are considered vulnerabilities of LLaMA C++.
|
||||
|
||||
<!-- normal version -->
|
||||
However, If you have discovered a security vulnerability in this project, please report it privately. **Do not disclose it as a public issue.** This gives us time to work with you to fix the issue before public exposure, reducing the chance that the exploit will be used before a patch is released.
|
||||
|
||||
Please disclose it as a private [security advisory](https://github.com/ggerganov/llama.cpp/security/advisories/new).
|
||||
|
||||
A team of volunteers on a reasonable-effort basis maintains this project. As such, please give us at least 90 days to work on a fix before public exposure.
|
40
SHA256SUMS
Normal file
40
SHA256SUMS
Normal file
|
@ -0,0 +1,40 @@
|
|||
700df0d3013b703a806d2ae7f1bfb8e59814e3d06ae78be0c66368a50059f33d models/7B/consolidated.00.pth
|
||||
666a4bb533b303bdaf89e1b6a3b6f93535d868de31d903afdc20983dc526c847 models/7B/ggml-model-f16.bin
|
||||
ec2f2d1f0dfb73b72a4cbac7fa121abbe04c37ab327125a38248f930c0f09ddf models/7B/ggml-model-q4_0.bin
|
||||
ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff models/7B/ggml-model-q4_1.bin
|
||||
ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff models/7B/ggml-model-q5_0.bin
|
||||
ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff models/7B/ggml-model-q5_1.bin
|
||||
7e89e242ddc0dd6f060b43ca219ce8b3e8f08959a72cb3c0855df8bb04d46265 models/7B/params.json
|
||||
745bf4e29a4dd6f411e72976d92b452da1b49168a4f41c951cfcc8051823cf08 models/13B/consolidated.00.pth
|
||||
d5ccbcc465c71c0de439a5aeffebe8344c68a519bce70bc7f9f92654ee567085 models/13B/consolidated.01.pth
|
||||
2b206e9b21fb1076f11cafc624e2af97c9e48ea09312a0962153acc20d45f808 models/13B/ggml-model-f16.bin
|
||||
fad169e6f0f575402cf75945961cb4a8ecd824ba4da6be2af831f320c4348fa5 models/13B/ggml-model-q4_0.bin
|
||||
ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff models/13B/ggml-model-q4_1.bin
|
||||
ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff models/13B/ggml-model-q5_0.bin
|
||||
ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff models/13B/ggml-model-q5_1.bin
|
||||
4ab77bec4d4405ccb66a97b282574c89a94417e3c32e5f68f37e2876fc21322f models/13B/params.json
|
||||
e23294a58552d8cdec5b7e8abb87993b97ea6eced4178ff2697c02472539d067 models/30B/consolidated.00.pth
|
||||
4e077b7136c7ae2302e954860cf64930458d3076fcde9443f4d0e939e95903ff models/30B/consolidated.01.pth
|
||||
24a87f01028cbd3a12de551dcedb712346c0b5cbdeff1454e0ddf2df9b675378 models/30B/consolidated.02.pth
|
||||
1adfcef71420886119544949767f6a56cb6339b4d5fcde755d80fe68b49de93b models/30B/consolidated.03.pth
|
||||
7e1b524061a9f4b27c22a12d6d2a5bf13b8ebbea73e99f218809351ed9cf7d37 models/30B/ggml-model-f16.bin
|
||||
d2a441403944819492ec8c2002cc36fa38468149bfb4b7b4c52afc7bd9a7166d models/30B/ggml-model-q4_0.bin
|
||||
ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff models/30B/ggml-model-q4_1.bin
|
||||
ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff models/30B/ggml-model-q5_0.bin
|
||||
ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff models/30B/ggml-model-q5_1.bin
|
||||
2c07118ea98d69dbe7810d88520e30288fa994751b337f8fca02b171955f44cb models/30B/params.json
|
||||
135c563f6b3938114458183afb01adc9a63bef3d8ff7cccc3977e5d3664ecafe models/65B/consolidated.00.pth
|
||||
9a600b37b19d38c7e43809485f70d17d1dc12206c07efa83bc72bb498a568bde models/65B/consolidated.01.pth
|
||||
e7babf7c5606f165a3756f527cb0fedc4f83e67ef1290391e52fb1cce5f26770 models/65B/consolidated.02.pth
|
||||
73176ffb426b40482f2aa67ae1217ef79fbbd1fff5482bae5060cdc5a24ab70e models/65B/consolidated.03.pth
|
||||
882e6431d0b08a8bc66261a0d3607da21cbaeafa96a24e7e59777632dbdac225 models/65B/consolidated.04.pth
|
||||
a287c0dfe49081626567c7fe87f74cce5831f58e459b427b5e05567641f47b78 models/65B/consolidated.05.pth
|
||||
72b4eba67a1a3b18cb67a85b70f8f1640caae9b40033ea943fb166bd80a7b36b models/65B/consolidated.06.pth
|
||||
d27f5b0677d7ff129ceacd73fd461c4d06910ad7787cf217b249948c3f3bc638 models/65B/consolidated.07.pth
|
||||
60758f2384d74e423dffddfd020ffed9d3bb186ebc54506f9c4a787d0f5367b0 models/65B/ggml-model-f16.bin
|
||||
cde053439fa4910ae454407e2717cc46cc2c2b4995c00c93297a2b52e790fa92 models/65B/ggml-model-q4_0.bin
|
||||
ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff models/65B/ggml-model-q4_1.bin
|
||||
ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff models/65B/ggml-model-q5_0.bin
|
||||
ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff models/65B/ggml-model-q5_1.bin
|
||||
999ed1659b469ccc2a941714c0a9656fa571d17c9f7c8c7589817ca90edef51b models/65B/params.json
|
||||
9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347 models/tokenizer.model
|
|
@ -1,4 +0,0 @@
|
|||
#pragma once
|
||||
|
||||
#include <llama.h>
|
||||
|
|
@ -1,5 +0,0 @@
|
|||
module llama [system] {
|
||||
header "llama.h"
|
||||
link "llama"
|
||||
export *
|
||||
}
|
138
build.zig
Normal file
138
build.zig
Normal file
|
@ -0,0 +1,138 @@
|
|||
// Compatible with Zig Version 0.11.0
|
||||
const std = @import("std");
|
||||
const ArrayList = std.ArrayList;
|
||||
const Compile = std.Build.Step.Compile;
|
||||
const ConfigHeader = std.Build.Step.ConfigHeader;
|
||||
const Mode = std.builtin.Mode;
|
||||
const CrossTarget = std.zig.CrossTarget;
|
||||
|
||||
const Maker = struct {
|
||||
builder: *std.build.Builder,
|
||||
target: CrossTarget,
|
||||
optimize: Mode,
|
||||
enable_lto: bool,
|
||||
|
||||
include_dirs: ArrayList([]const u8),
|
||||
cflags: ArrayList([]const u8),
|
||||
cxxflags: ArrayList([]const u8),
|
||||
objs: ArrayList(*Compile),
|
||||
|
||||
fn addInclude(m: *Maker, dir: []const u8) !void {
|
||||
try m.include_dirs.append(dir);
|
||||
}
|
||||
fn addProjectInclude(m: *Maker, path: []const []const u8) !void {
|
||||
try m.addInclude(try m.builder.build_root.join(m.builder.allocator, path));
|
||||
}
|
||||
fn addCFlag(m: *Maker, flag: []const u8) !void {
|
||||
try m.cflags.append(flag);
|
||||
}
|
||||
fn addCxxFlag(m: *Maker, flag: []const u8) !void {
|
||||
try m.cxxflags.append(flag);
|
||||
}
|
||||
fn addFlag(m: *Maker, flag: []const u8) !void {
|
||||
try m.addCFlag(flag);
|
||||
try m.addCxxFlag(flag);
|
||||
}
|
||||
|
||||
fn init(builder: *std.build.Builder) !Maker {
|
||||
const target = builder.standardTargetOptions(.{});
|
||||
const zig_version = @import("builtin").zig_version_string;
|
||||
const commit_hash = try std.ChildProcess.exec(
|
||||
.{ .allocator = builder.allocator, .argv = &.{ "git", "rev-parse", "HEAD" } },
|
||||
);
|
||||
try std.fs.cwd().writeFile("common/build-info.cpp", builder.fmt(
|
||||
\\int LLAMA_BUILD_NUMBER = {};
|
||||
\\char const *LLAMA_COMMIT = "{s}";
|
||||
\\char const *LLAMA_COMPILER = "Zig {s}";
|
||||
\\char const *LLAMA_BUILD_TARGET = "{s}";
|
||||
\\
|
||||
, .{ 0, commit_hash.stdout[0 .. commit_hash.stdout.len - 1], zig_version, try target.allocDescription(builder.allocator) }));
|
||||
var m = Maker{
|
||||
.builder = builder,
|
||||
.target = target,
|
||||
.optimize = builder.standardOptimizeOption(.{}),
|
||||
.enable_lto = false,
|
||||
.include_dirs = ArrayList([]const u8).init(builder.allocator),
|
||||
.cflags = ArrayList([]const u8).init(builder.allocator),
|
||||
.cxxflags = ArrayList([]const u8).init(builder.allocator),
|
||||
.objs = ArrayList(*Compile).init(builder.allocator),
|
||||
};
|
||||
|
||||
try m.addCFlag("-std=c11");
|
||||
try m.addCxxFlag("-std=c++11");
|
||||
try m.addProjectInclude(&.{});
|
||||
try m.addProjectInclude(&.{"common"});
|
||||
return m;
|
||||
}
|
||||
|
||||
fn obj(m: *const Maker, name: []const u8, src: []const u8) *Compile {
|
||||
const o = m.builder.addObject(.{ .name = name, .target = m.target, .optimize = m.optimize });
|
||||
if (o.target.getAbi() != .msvc)
|
||||
o.defineCMacro("_GNU_SOURCE", null);
|
||||
|
||||
if (std.mem.endsWith(u8, src, ".c")) {
|
||||
o.addCSourceFiles(&.{src}, m.cflags.items);
|
||||
o.linkLibC();
|
||||
} else {
|
||||
o.addCSourceFiles(&.{src}, m.cxxflags.items);
|
||||
if (o.target.getAbi() == .msvc) {
|
||||
o.linkLibC(); // need winsdk + crt
|
||||
} else {
|
||||
// linkLibCpp already add (libc++ + libunwind + libc)
|
||||
o.linkLibCpp();
|
||||
}
|
||||
}
|
||||
for (m.include_dirs.items) |i| o.addIncludePath(.{ .path = i });
|
||||
o.want_lto = m.enable_lto;
|
||||
return o;
|
||||
}
|
||||
|
||||
fn exe(m: *const Maker, name: []const u8, src: []const u8, deps: []const *Compile) *Compile {
|
||||
const e = m.builder.addExecutable(.{ .name = name, .target = m.target, .optimize = m.optimize });
|
||||
e.addCSourceFiles(&.{src}, m.cxxflags.items);
|
||||
for (deps) |d| e.addObject(d);
|
||||
for (m.objs.items) |o| e.addObject(o);
|
||||
for (m.include_dirs.items) |i| e.addIncludePath(.{ .path = i });
|
||||
|
||||
// https://github.com/ziglang/zig/issues/15448
|
||||
if (e.target.getAbi() == .msvc) {
|
||||
e.linkLibC(); // need winsdk + crt
|
||||
} else {
|
||||
// linkLibCpp already add (libc++ + libunwind + libc)
|
||||
e.linkLibCpp();
|
||||
}
|
||||
m.builder.installArtifact(e);
|
||||
e.want_lto = m.enable_lto;
|
||||
return e;
|
||||
}
|
||||
};
|
||||
|
||||
pub fn build(b: *std.build.Builder) !void {
|
||||
var make = try Maker.init(b);
|
||||
make.enable_lto = b.option(bool, "lto", "Enable LTO optimization, (default: false)") orelse false;
|
||||
|
||||
const ggml = make.obj("ggml", "ggml.c");
|
||||
const ggml_alloc = make.obj("ggml-alloc", "ggml-alloc.c");
|
||||
const ggml_backend = make.obj("ggml-backend", "ggml-backend.c");
|
||||
const ggml_quants = make.obj("ggml-quants", "ggml-quants.c");
|
||||
const llama = make.obj("llama", "llama.cpp");
|
||||
const buildinfo = make.obj("common", "common/build-info.cpp");
|
||||
const common = make.obj("common", "common/common.cpp");
|
||||
const console = make.obj("console", "common/console.cpp");
|
||||
const sampling = make.obj("sampling", "common/sampling.cpp");
|
||||
const grammar_parser = make.obj("grammar-parser", "common/grammar-parser.cpp");
|
||||
const train = make.obj("train", "common/train.cpp");
|
||||
const clip = make.obj("clip", "examples/llava/clip.cpp");
|
||||
|
||||
_ = make.exe("main", "examples/main/main.cpp", &.{ ggml, ggml_alloc, ggml_backend, ggml_quants, llama, common, buildinfo, sampling, console, grammar_parser });
|
||||
_ = make.exe("quantize", "examples/quantize/quantize.cpp", &.{ ggml, ggml_alloc, ggml_backend, ggml_quants, llama, common, buildinfo });
|
||||
_ = make.exe("perplexity", "examples/perplexity/perplexity.cpp", &.{ ggml, ggml_alloc, ggml_backend, ggml_quants, llama, common, buildinfo });
|
||||
_ = make.exe("embedding", "examples/embedding/embedding.cpp", &.{ ggml, ggml_alloc, ggml_backend, ggml_quants, llama, common, buildinfo });
|
||||
_ = make.exe("finetune", "examples/finetune/finetune.cpp", &.{ ggml, ggml_alloc, ggml_backend, ggml_quants, llama, common, buildinfo, train });
|
||||
_ = make.exe("train-text-from-scratch", "examples/train-text-from-scratch/train-text-from-scratch.cpp", &.{ ggml, ggml_alloc, ggml_backend, ggml_quants, llama, common, buildinfo, train });
|
||||
|
||||
const server = make.exe("server", "examples/server/server.cpp", &.{ ggml, ggml_alloc, ggml_backend, ggml_quants, llama, common, buildinfo, sampling, grammar_parser, clip });
|
||||
if (server.target.isWindows()) {
|
||||
server.linkSystemLibrary("ws2_32");
|
||||
}
|
||||
}
|
|
@ -22,8 +22,4 @@ bash ./ci/run.sh ./tmp/results ./tmp/mnt
|
|||
|
||||
# with CUDA support
|
||||
GG_BUILD_CUDA=1 bash ./ci/run.sh ./tmp/results ./tmp/mnt
|
||||
|
||||
# with SYCL support
|
||||
source /opt/intel/oneapi/setvars.sh
|
||||
GG_BUILD_SYCL=1 bash ./ci/run.sh ./tmp/results ./tmp/mnt
|
||||
```
|
||||
|
|
748
ci/run.sh
748
ci/run.sh
|
@ -1,4 +1,4 @@
|
|||
#!/bin/bash
|
||||
#/bin/bash
|
||||
#
|
||||
# sample usage:
|
||||
#
|
||||
|
@ -10,12 +10,6 @@
|
|||
# # with CUDA support
|
||||
# GG_BUILD_CUDA=1 bash ./ci/run.sh ./tmp/results ./tmp/mnt
|
||||
#
|
||||
# # with SYCL support
|
||||
# GG_BUILD_SYCL=1 bash ./ci/run.sh ./tmp/results ./tmp/mnt
|
||||
#
|
||||
# # with VULKAN support
|
||||
# GG_BUILD_VULKAN=1 bash ./ci/run.sh ./tmp/results ./tmp/mnt
|
||||
#
|
||||
|
||||
if [ -z "$2" ]; then
|
||||
echo "usage: $0 <output-dir> <mnt-dir>"
|
||||
|
@ -28,37 +22,14 @@ mkdir -p "$2"
|
|||
OUT=$(realpath "$1")
|
||||
MNT=$(realpath "$2")
|
||||
|
||||
rm -f "$OUT/*.log"
|
||||
rm -f "$OUT/*.exit"
|
||||
rm -f "$OUT/*.md"
|
||||
rm -v $OUT/*.log
|
||||
rm -v $OUT/*.exit
|
||||
rm -v $OUT/*.md
|
||||
|
||||
sd=`dirname $0`
|
||||
cd $sd/../
|
||||
SRC=`pwd`
|
||||
|
||||
CMAKE_EXTRA="-DLLAMA_FATAL_WARNINGS=ON"
|
||||
|
||||
if [ ! -z ${GG_BUILD_METAL} ]; then
|
||||
CMAKE_EXTRA="${CMAKE_EXTRA} -DGGML_METAL=ON -DGGML_METAL_USE_BF16=ON"
|
||||
fi
|
||||
|
||||
if [ ! -z ${GG_BUILD_CUDA} ]; then
|
||||
CMAKE_EXTRA="${CMAKE_EXTRA} -DGGML_CUDA=ON -DCMAKE_CUDA_ARCHITECTURES=native"
|
||||
fi
|
||||
|
||||
if [ ! -z ${GG_BUILD_SYCL} ]; then
|
||||
if [ -z ${ONEAPI_ROOT} ]; then
|
||||
echo "Not detected ONEAPI_ROOT, please install oneAPI base toolkit and enable it by:"
|
||||
echo "source /opt/intel/oneapi/setvars.sh"
|
||||
exit 1
|
||||
fi
|
||||
|
||||
CMAKE_EXTRA="${CMAKE_EXTRA} -DGGML_SYCL=1 -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx -DGGML_SYCL_F16=ON"
|
||||
fi
|
||||
|
||||
if [ ! -z ${GG_BUILD_VULKAN} ]; then
|
||||
CMAKE_EXTRA="${CMAKE_EXTRA} -DGGML_VULKAN=1"
|
||||
fi
|
||||
## helpers
|
||||
|
||||
# download a file if it does not exist or if it is outdated
|
||||
|
@ -110,13 +81,10 @@ function gg_run_ctest_debug {
|
|||
|
||||
set -e
|
||||
|
||||
# Check cmake, make and ctest are installed
|
||||
gg_check_build_requirements
|
||||
(time cmake -DCMAKE_BUILD_TYPE=Debug .. ) 2>&1 | tee -a $OUT/${ci}-cmake.log
|
||||
(time make -j ) 2>&1 | tee -a $OUT/${ci}-make.log
|
||||
|
||||
(time cmake -DCMAKE_BUILD_TYPE=Debug ${CMAKE_EXTRA} .. ) 2>&1 | tee -a $OUT/${ci}-cmake.log
|
||||
(time make -j$(nproc) ) 2>&1 | tee -a $OUT/${ci}-make.log
|
||||
|
||||
(time ctest --output-on-failure -L main -E test-opt ) 2>&1 | tee -a $OUT/${ci}-ctest.log
|
||||
(time ctest --output-on-failure -E test-opt ) 2>&1 | tee -a $OUT/${ci}-ctest.log
|
||||
|
||||
set +e
|
||||
}
|
||||
|
@ -141,16 +109,13 @@ function gg_run_ctest_release {
|
|||
|
||||
set -e
|
||||
|
||||
# Check cmake, make and ctest are installed
|
||||
gg_check_build_requirements
|
||||
|
||||
(time cmake -DCMAKE_BUILD_TYPE=Release ${CMAKE_EXTRA} .. ) 2>&1 | tee -a $OUT/${ci}-cmake.log
|
||||
(time make -j$(nproc) ) 2>&1 | tee -a $OUT/${ci}-make.log
|
||||
(time cmake -DCMAKE_BUILD_TYPE=Release .. ) 2>&1 | tee -a $OUT/${ci}-cmake.log
|
||||
(time make -j ) 2>&1 | tee -a $OUT/${ci}-make.log
|
||||
|
||||
if [ -z ${GG_BUILD_LOW_PERF} ]; then
|
||||
(time ctest --output-on-failure -L main ) 2>&1 | tee -a $OUT/${ci}-ctest.log
|
||||
(time ctest --output-on-failure ) 2>&1 | tee -a $OUT/${ci}-ctest.log
|
||||
else
|
||||
(time ctest --output-on-failure -L main -E test-opt ) 2>&1 | tee -a $OUT/${ci}-ctest.log
|
||||
(time ctest --output-on-failure -E test-opt ) 2>&1 | tee -a $OUT/${ci}-ctest.log
|
||||
fi
|
||||
|
||||
set +e
|
||||
|
@ -166,140 +131,33 @@ function gg_sum_ctest_release {
|
|||
gg_printf '```\n'
|
||||
}
|
||||
|
||||
# test_scripts_debug
|
||||
# open_llama_3b_v2
|
||||
|
||||
function gg_run_test_scripts_debug {
|
||||
function gg_run_open_llama_3b_v2 {
|
||||
cd ${SRC}
|
||||
|
||||
set -e
|
||||
gg_wget models-mnt/open-llama/3B-v2/ https://huggingface.co/openlm-research/open_llama_3b_v2/raw/main/config.json
|
||||
gg_wget models-mnt/open-llama/3B-v2/ https://huggingface.co/openlm-research/open_llama_3b_v2/resolve/main/tokenizer.model
|
||||
gg_wget models-mnt/open-llama/3B-v2/ https://huggingface.co/openlm-research/open_llama_3b_v2/raw/main/tokenizer_config.json
|
||||
gg_wget models-mnt/open-llama/3B-v2/ https://huggingface.co/openlm-research/open_llama_3b_v2/raw/main/special_tokens_map.json
|
||||
gg_wget models-mnt/open-llama/3B-v2/ https://huggingface.co/openlm-research/open_llama_3b_v2/resolve/main/pytorch_model.bin
|
||||
gg_wget models-mnt/open-llama/3B-v2/ https://huggingface.co/openlm-research/open_llama_3b_v2/raw/main/generation_config.json
|
||||
|
||||
(cd ./examples/gguf-split && time bash tests.sh "$SRC/build-ci-debug/bin" "$MNT/models") 2>&1 | tee -a $OUT/${ci}-scripts.log
|
||||
(cd ./examples/quantize && time bash tests.sh "$SRC/build-ci-debug/bin" "$MNT/models") 2>&1 | tee -a $OUT/${ci}-scripts.log
|
||||
|
||||
set +e
|
||||
}
|
||||
|
||||
function gg_sum_test_scripts_debug {
|
||||
gg_printf '### %s\n\n' "${ci}"
|
||||
|
||||
gg_printf 'Runs test scripts in debug mode\n'
|
||||
gg_printf '- status: %s\n' "$(cat $OUT/${ci}.exit)"
|
||||
gg_printf '```\n'
|
||||
gg_printf '%s\n' "$(cat $OUT/${ci}-scripts.log)"
|
||||
gg_printf '```\n'
|
||||
gg_printf '\n'
|
||||
}
|
||||
|
||||
# test_scripts_release
|
||||
|
||||
function gg_run_test_scripts_release {
|
||||
cd ${SRC}
|
||||
|
||||
set -e
|
||||
|
||||
(cd ./examples/gguf-split && time bash tests.sh "$SRC/build-ci-release/bin" "$MNT/models") 2>&1 | tee -a $OUT/${ci}-scripts.log
|
||||
(cd ./examples/quantize && time bash tests.sh "$SRC/build-ci-release/bin" "$MNT/models") 2>&1 | tee -a $OUT/${ci}-scripts.log
|
||||
|
||||
set +e
|
||||
}
|
||||
|
||||
function gg_sum_test_scripts_release {
|
||||
gg_printf '### %s\n\n' "${ci}"
|
||||
|
||||
gg_printf 'Runs test scripts in release mode\n'
|
||||
gg_printf '- status: %s\n' "$(cat $OUT/${ci}.exit)"
|
||||
gg_printf '```\n'
|
||||
gg_printf '%s\n' "$(cat $OUT/${ci}-scripts.log)"
|
||||
gg_printf '```\n'
|
||||
gg_printf '\n'
|
||||
}
|
||||
|
||||
function gg_get_model {
|
||||
local gguf_0="$MNT/models/pythia/1.4B/ggml-model-f16.gguf"
|
||||
local gguf_1="$MNT/models/pythia/2.8B/ggml-model-f16.gguf"
|
||||
local gguf_2="$MNT/models/open-llama/7B-v2/ggml-model-f16.gguf"
|
||||
if [[ -s $gguf_0 ]]; then
|
||||
echo -n "$gguf_0"
|
||||
elif [[ -s $gguf_1 ]]; then
|
||||
echo -n "$gguf_1"
|
||||
elif [[ -s $gguf_2 ]]; then
|
||||
echo -n "$gguf_2"
|
||||
else
|
||||
echo >&2 "No model found. Can't run gg_run_ctest_with_model."
|
||||
exit 1
|
||||
fi
|
||||
}
|
||||
|
||||
function gg_run_ctest_with_model_debug {
|
||||
cd ${SRC}
|
||||
|
||||
local model; model=$(gg_get_model)
|
||||
cd build-ci-debug
|
||||
set -e
|
||||
(LLAMACPP_TEST_MODELFILE="$model" time ctest --output-on-failure -L model) 2>&1 | tee -a $OUT/${ci}-ctest.log
|
||||
set +e
|
||||
cd ..
|
||||
}
|
||||
|
||||
function gg_run_ctest_with_model_release {
|
||||
cd ${SRC}
|
||||
|
||||
local model; model=$(gg_get_model)
|
||||
cd build-ci-release
|
||||
set -e
|
||||
(LLAMACPP_TEST_MODELFILE="$model" time ctest --output-on-failure -L model) 2>&1 | tee -a $OUT/${ci}-ctest.log
|
||||
set +e
|
||||
cd ..
|
||||
}
|
||||
|
||||
function gg_sum_ctest_with_model_debug {
|
||||
gg_printf '### %s\n\n' "${ci}"
|
||||
|
||||
gg_printf 'Runs ctest with model files in debug mode\n'
|
||||
gg_printf '- status: %s\n' "$(cat $OUT/${ci}.exit)"
|
||||
gg_printf '```\n'
|
||||
gg_printf '%s\n' "$(cat $OUT/${ci}-ctest.log)"
|
||||
gg_printf '```\n'
|
||||
}
|
||||
|
||||
function gg_sum_ctest_with_model_release {
|
||||
gg_printf '### %s\n\n' "${ci}"
|
||||
|
||||
gg_printf 'Runs ctest with model files in release mode\n'
|
||||
gg_printf '- status: %s\n' "$(cat $OUT/${ci}.exit)"
|
||||
gg_printf '```\n'
|
||||
gg_printf '%s\n' "$(cat $OUT/${ci}-ctest.log)"
|
||||
gg_printf '```\n'
|
||||
}
|
||||
|
||||
# open_llama_7b_v2
|
||||
|
||||
function gg_run_open_llama_7b_v2 {
|
||||
cd ${SRC}
|
||||
|
||||
gg_wget models-mnt/open-llama/7B-v2/ https://huggingface.co/openlm-research/open_llama_7b_v2/raw/main/config.json
|
||||
gg_wget models-mnt/open-llama/7B-v2/ https://huggingface.co/openlm-research/open_llama_7b_v2/resolve/main/tokenizer.model
|
||||
gg_wget models-mnt/open-llama/7B-v2/ https://huggingface.co/openlm-research/open_llama_7b_v2/raw/main/tokenizer_config.json
|
||||
gg_wget models-mnt/open-llama/7B-v2/ https://huggingface.co/openlm-research/open_llama_7b_v2/raw/main/special_tokens_map.json
|
||||
gg_wget models-mnt/open-llama/7B-v2/ https://huggingface.co/openlm-research/open_llama_7b_v2/raw/main/pytorch_model.bin.index.json
|
||||
gg_wget models-mnt/open-llama/7B-v2/ https://huggingface.co/openlm-research/open_llama_7b_v2/resolve/main/pytorch_model-00001-of-00002.bin
|
||||
gg_wget models-mnt/open-llama/7B-v2/ https://huggingface.co/openlm-research/open_llama_7b_v2/resolve/main/pytorch_model-00002-of-00002.bin
|
||||
gg_wget models-mnt/open-llama/7B-v2/ https://huggingface.co/openlm-research/open_llama_7b_v2/raw/main/generation_config.json
|
||||
|
||||
gg_wget models-mnt/wikitext/ https://huggingface.co/datasets/ggml-org/ci/resolve/main/wikitext-2-raw-v1.zip
|
||||
gg_wget models-mnt/wikitext/ https://s3.amazonaws.com/research.metamind.io/wikitext/wikitext-2-raw-v1.zip
|
||||
unzip -o models-mnt/wikitext/wikitext-2-raw-v1.zip -d models-mnt/wikitext/
|
||||
head -n 60 models-mnt/wikitext/wikitext-2-raw/wiki.test.raw > models-mnt/wikitext/wikitext-2-raw/wiki.test-60.raw
|
||||
|
||||
path_models="../models-mnt/open-llama/7B-v2"
|
||||
path_models="../models-mnt/open-llama/3B-v2"
|
||||
path_wiki="../models-mnt/wikitext/wikitext-2-raw"
|
||||
|
||||
rm -rf build-ci-release && mkdir build-ci-release && cd build-ci-release
|
||||
|
||||
set -e
|
||||
|
||||
(time cmake -DCMAKE_BUILD_TYPE=Release ${CMAKE_EXTRA} .. ) 2>&1 | tee -a $OUT/${ci}-cmake.log
|
||||
(time make -j$(nproc) ) 2>&1 | tee -a $OUT/${ci}-make.log
|
||||
(time cmake -DCMAKE_BUILD_TYPE=Release -DLLAMA_QKK_64=1 .. ) 2>&1 | tee -a $OUT/${ci}-cmake.log
|
||||
(time make -j ) 2>&1 | tee -a $OUT/${ci}-make.log
|
||||
|
||||
python3 ../examples/convert_legacy_llama.py ${path_models} --outfile ${path_models}/ggml-model-f16.gguf
|
||||
python3 ../convert.py ${path_models}
|
||||
|
||||
model_f16="${path_models}/ggml-model-f16.gguf"
|
||||
model_q8_0="${path_models}/ggml-model-q8_0.gguf"
|
||||
|
@ -313,49 +171,44 @@ function gg_run_open_llama_7b_v2 {
|
|||
model_q5_k="${path_models}/ggml-model-q5_k.gguf"
|
||||
model_q6_k="${path_models}/ggml-model-q6_k.gguf"
|
||||
|
||||
wiki_test="${path_wiki}/wiki.test.raw"
|
||||
wiki_test_60="${path_wiki}/wiki.test-60.raw"
|
||||
|
||||
./bin/llama-quantize ${model_f16} ${model_q8_0} q8_0
|
||||
./bin/llama-quantize ${model_f16} ${model_q4_0} q4_0
|
||||
./bin/llama-quantize ${model_f16} ${model_q4_1} q4_1
|
||||
./bin/llama-quantize ${model_f16} ${model_q5_0} q5_0
|
||||
./bin/llama-quantize ${model_f16} ${model_q5_1} q5_1
|
||||
./bin/llama-quantize ${model_f16} ${model_q2_k} q2_k
|
||||
./bin/llama-quantize ${model_f16} ${model_q3_k} q3_k
|
||||
./bin/llama-quantize ${model_f16} ${model_q4_k} q4_k
|
||||
./bin/llama-quantize ${model_f16} ${model_q5_k} q5_k
|
||||
./bin/llama-quantize ${model_f16} ${model_q6_k} q6_k
|
||||
./bin/quantize ${model_f16} ${model_q8_0} q8_0
|
||||
./bin/quantize ${model_f16} ${model_q4_0} q4_0
|
||||
./bin/quantize ${model_f16} ${model_q4_1} q4_1
|
||||
./bin/quantize ${model_f16} ${model_q5_0} q5_0
|
||||
./bin/quantize ${model_f16} ${model_q5_1} q5_1
|
||||
./bin/quantize ${model_f16} ${model_q2_k} q2_k
|
||||
./bin/quantize ${model_f16} ${model_q3_k} q3_k
|
||||
./bin/quantize ${model_f16} ${model_q4_k} q4_k
|
||||
./bin/quantize ${model_f16} ${model_q5_k} q5_k
|
||||
./bin/quantize ${model_f16} ${model_q6_k} q6_k
|
||||
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_f16} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-f16.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q8_0} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q8_0.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q4_0} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_0.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q4_1} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_1.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q5_0} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_0.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q5_1} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_1.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q2_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q2_k.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q3_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q3_k.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q4_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_k.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q5_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_k.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q6_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q6_k.log
|
||||
(time ./bin/main --model ${model_f16} -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-f16.log
|
||||
(time ./bin/main --model ${model_q8_0} -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q8_0.log
|
||||
(time ./bin/main --model ${model_q4_0} -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_0.log
|
||||
(time ./bin/main --model ${model_q4_1} -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_1.log
|
||||
(time ./bin/main --model ${model_q5_0} -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_0.log
|
||||
(time ./bin/main --model ${model_q5_1} -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_1.log
|
||||
(time ./bin/main --model ${model_q2_k} -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q2_k.log
|
||||
(time ./bin/main --model ${model_q3_k} -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q3_k.log
|
||||
(time ./bin/main --model ${model_q4_k} -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_k.log
|
||||
(time ./bin/main --model ${model_q5_k} -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_k.log
|
||||
(time ./bin/main --model ${model_q6_k} -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q6_k.log
|
||||
|
||||
(time ./bin/llama-perplexity --model ${model_f16} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-f16.log
|
||||
(time ./bin/llama-perplexity --model ${model_q8_0} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q8_0.log
|
||||
(time ./bin/llama-perplexity --model ${model_q4_0} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q4_0.log
|
||||
(time ./bin/llama-perplexity --model ${model_q4_1} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q4_1.log
|
||||
(time ./bin/llama-perplexity --model ${model_q5_0} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q5_0.log
|
||||
(time ./bin/llama-perplexity --model ${model_q5_1} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q5_1.log
|
||||
(time ./bin/llama-perplexity --model ${model_q2_k} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q2_k.log
|
||||
(time ./bin/llama-perplexity --model ${model_q3_k} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q3_k.log
|
||||
(time ./bin/llama-perplexity --model ${model_q4_k} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q4_k.log
|
||||
(time ./bin/llama-perplexity --model ${model_q5_k} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q5_k.log
|
||||
(time ./bin/llama-perplexity --model ${model_q6_k} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q6_k.log
|
||||
(time ./bin/perplexity --model ${model_f16} -f ${wiki_test_60} -c 128 -b 128 --chunks 2 ) 2>&1 | tee -a $OUT/${ci}-tg-f16.log
|
||||
(time ./bin/perplexity --model ${model_q8_0} -f ${wiki_test_60} -c 128 -b 128 --chunks 2 ) 2>&1 | tee -a $OUT/${ci}-tg-q8_0.log
|
||||
(time ./bin/perplexity --model ${model_q4_0} -f ${wiki_test_60} -c 128 -b 128 --chunks 2 ) 2>&1 | tee -a $OUT/${ci}-tg-q4_0.log
|
||||
(time ./bin/perplexity --model ${model_q4_1} -f ${wiki_test_60} -c 128 -b 128 --chunks 2 ) 2>&1 | tee -a $OUT/${ci}-tg-q4_1.log
|
||||
(time ./bin/perplexity --model ${model_q5_0} -f ${wiki_test_60} -c 128 -b 128 --chunks 2 ) 2>&1 | tee -a $OUT/${ci}-tg-q5_0.log
|
||||
(time ./bin/perplexity --model ${model_q5_1} -f ${wiki_test_60} -c 128 -b 128 --chunks 2 ) 2>&1 | tee -a $OUT/${ci}-tg-q5_1.log
|
||||
(time ./bin/perplexity --model ${model_q2_k} -f ${wiki_test_60} -c 128 -b 128 --chunks 2 ) 2>&1 | tee -a $OUT/${ci}-tg-q2_k.log
|
||||
(time ./bin/perplexity --model ${model_q3_k} -f ${wiki_test_60} -c 128 -b 128 --chunks 2 ) 2>&1 | tee -a $OUT/${ci}-tg-q3_k.log
|
||||
(time ./bin/perplexity --model ${model_q4_k} -f ${wiki_test_60} -c 128 -b 128 --chunks 2 ) 2>&1 | tee -a $OUT/${ci}-tg-q4_k.log
|
||||
(time ./bin/perplexity --model ${model_q5_k} -f ${wiki_test_60} -c 128 -b 128 --chunks 2 ) 2>&1 | tee -a $OUT/${ci}-tg-q5_k.log
|
||||
(time ./bin/perplexity --model ${model_q6_k} -f ${wiki_test_60} -c 128 -b 128 --chunks 2 ) 2>&1 | tee -a $OUT/${ci}-tg-q6_k.log
|
||||
|
||||
(time ./bin/llama-imatrix --model ${model_f16} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-imatrix.log
|
||||
|
||||
(time ./bin/llama-save-load-state--model ${model_q4_0} -ngl 10 -c 0 ) 2>&1 | tee -a $OUT/${ci}-save-load-state.log
|
||||
(time ./bin/llama-save-load-state--model ${model_q4_0} -ngl 10 -c 0 -fa ) 2>&1 | tee -a $OUT/${ci}-save-load-state.log
|
||||
(time ./bin/llama-save-load-state--model ${model_q4_0} -ngl 99 -c 0 ) 2>&1 | tee -a $OUT/${ci}-save-load-state.log
|
||||
(time ./bin/llama-save-load-state--model ${model_q4_0} -ngl 99 -c 0 -fa ) 2>&1 | tee -a $OUT/${ci}-save-load-state.log
|
||||
(time ./bin/save-load-state --model ${model_q4_0} ) 2>&1 | tee -a $OUT/${ci}-save-load-state.log
|
||||
|
||||
function check_ppl {
|
||||
qnt="$1"
|
||||
|
@ -382,150 +235,58 @@ function gg_run_open_llama_7b_v2 {
|
|||
check_ppl "q5_k" "$(cat $OUT/${ci}-tg-q5_k.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log
|
||||
check_ppl "q6_k" "$(cat $OUT/${ci}-tg-q6_k.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log
|
||||
|
||||
cat $OUT/${ci}-imatrix.log | grep "Final" >> $OUT/${ci}-imatrix-sum.log
|
||||
|
||||
set +e
|
||||
}
|
||||
|
||||
function gg_sum_open_llama_7b_v2 {
|
||||
gg_printf '### %s\n\n' "${ci}"
|
||||
|
||||
gg_printf 'OpenLLaMA 7B-v2:\n'
|
||||
gg_printf '- status: %s\n' "$(cat $OUT/${ci}.exit)"
|
||||
gg_printf '- perplexity:\n%s\n' "$(cat $OUT/${ci}-ppl.log)"
|
||||
gg_printf '- imatrix:\n```\n%s\n```\n' "$(cat $OUT/${ci}-imatrix-sum.log)"
|
||||
gg_printf '- f16: \n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-f16.log)"
|
||||
gg_printf '- q8_0:\n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-q8_0.log)"
|
||||
gg_printf '- q4_0:\n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-q4_0.log)"
|
||||
gg_printf '- q4_1:\n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-q4_1.log)"
|
||||
gg_printf '- q5_0:\n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-q5_0.log)"
|
||||
gg_printf '- q5_1:\n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-q5_1.log)"
|
||||
gg_printf '- q2_k:\n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-q2_k.log)"
|
||||
gg_printf '- q3_k:\n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-q3_k.log)"
|
||||
gg_printf '- q4_k:\n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-q4_k.log)"
|
||||
gg_printf '- q5_k:\n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-q5_k.log)"
|
||||
gg_printf '- q6_k:\n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-q6_k.log)"
|
||||
gg_printf '- save-load-state: \n```\n%s\n```\n' "$(cat $OUT/${ci}-save-load-state.log)"
|
||||
}
|
||||
|
||||
# pythia_1.4b
|
||||
|
||||
function gg_run_pythia_1_4b {
|
||||
cd ${SRC}
|
||||
|
||||
gg_wget models-mnt/pythia/1.4B/ https://huggingface.co/EleutherAI/pythia-1.4b/raw/main/config.json
|
||||
gg_wget models-mnt/pythia/1.4B/ https://huggingface.co/EleutherAI/pythia-1.4b/raw/main/tokenizer.json
|
||||
gg_wget models-mnt/pythia/1.4B/ https://huggingface.co/EleutherAI/pythia-1.4b/raw/main/tokenizer_config.json
|
||||
gg_wget models-mnt/pythia/1.4B/ https://huggingface.co/EleutherAI/pythia-1.4b/raw/main/special_tokens_map.json
|
||||
gg_wget models-mnt/pythia/1.4B/ https://huggingface.co/EleutherAI/pythia-1.4b/resolve/main/pytorch_model.bin
|
||||
|
||||
gg_wget models-mnt/wikitext/ https://huggingface.co/datasets/ggml-org/ci/resolve/main/wikitext-2-raw-v1.zip
|
||||
unzip -o models-mnt/wikitext/wikitext-2-raw-v1.zip -d models-mnt/wikitext/
|
||||
head -n 60 models-mnt/wikitext/wikitext-2-raw/wiki.test.raw > models-mnt/wikitext/wikitext-2-raw/wiki.test-60.raw
|
||||
|
||||
path_models="../models-mnt/pythia/1.4B"
|
||||
path_wiki="../models-mnt/wikitext/wikitext-2-raw"
|
||||
|
||||
rm -rf build-ci-release && mkdir build-ci-release && cd build-ci-release
|
||||
|
||||
set -e
|
||||
|
||||
(time cmake -DCMAKE_BUILD_TYPE=Release ${CMAKE_EXTRA} .. ) 2>&1 | tee -a $OUT/${ci}-cmake.log
|
||||
(time make -j$(nproc) ) 2>&1 | tee -a $OUT/${ci}-make.log
|
||||
|
||||
python3 ../convert_hf_to_gguf.py ${path_models} --outfile ${path_models}/ggml-model-f16.gguf
|
||||
|
||||
model_f16="${path_models}/ggml-model-f16.gguf"
|
||||
model_q8_0="${path_models}/ggml-model-q8_0.gguf"
|
||||
model_q4_0="${path_models}/ggml-model-q4_0.gguf"
|
||||
model_q4_1="${path_models}/ggml-model-q4_1.gguf"
|
||||
model_q5_0="${path_models}/ggml-model-q5_0.gguf"
|
||||
model_q5_1="${path_models}/ggml-model-q5_1.gguf"
|
||||
model_q2_k="${path_models}/ggml-model-q2_k.gguf"
|
||||
model_q3_k="${path_models}/ggml-model-q3_k.gguf"
|
||||
model_q4_k="${path_models}/ggml-model-q4_k.gguf"
|
||||
model_q5_k="${path_models}/ggml-model-q5_k.gguf"
|
||||
model_q6_k="${path_models}/ggml-model-q6_k.gguf"
|
||||
|
||||
wiki_test_60="${path_wiki}/wiki.test-60.raw"
|
||||
|
||||
./bin/llama-quantize ${model_f16} ${model_q8_0} q8_0
|
||||
./bin/llama-quantize ${model_f16} ${model_q4_0} q4_0
|
||||
./bin/llama-quantize ${model_f16} ${model_q4_1} q4_1
|
||||
./bin/llama-quantize ${model_f16} ${model_q5_0} q5_0
|
||||
./bin/llama-quantize ${model_f16} ${model_q5_1} q5_1
|
||||
./bin/llama-quantize ${model_f16} ${model_q2_k} q2_k
|
||||
./bin/llama-quantize ${model_f16} ${model_q3_k} q3_k
|
||||
./bin/llama-quantize ${model_f16} ${model_q4_k} q4_k
|
||||
./bin/llama-quantize ${model_f16} ${model_q5_k} q5_k
|
||||
./bin/llama-quantize ${model_f16} ${model_q6_k} q6_k
|
||||
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_f16} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-f16.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q8_0} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q8_0.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q4_0} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_0.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q4_1} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_1.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q5_0} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_0.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q5_1} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_1.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q2_k} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q2_k.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q3_k} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q3_k.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q4_k} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_k.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q5_k} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_k.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q6_k} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q6_k.log
|
||||
|
||||
(time ./bin/llama-perplexity --model ${model_f16} -f ${wiki_test_60} -ngl 99 -c 128 -b 128 --chunks 1 ) 2>&1 | tee -a $OUT/${ci}-tg-f16.log
|
||||
(time ./bin/llama-perplexity --model ${model_q8_0} -f ${wiki_test_60} -ngl 99 -c 128 -b 128 --chunks 1 ) 2>&1 | tee -a $OUT/${ci}-tg-q8_0.log
|
||||
(time ./bin/llama-perplexity --model ${model_q4_0} -f ${wiki_test_60} -ngl 99 -c 128 -b 128 --chunks 1 ) 2>&1 | tee -a $OUT/${ci}-tg-q4_0.log
|
||||
(time ./bin/llama-perplexity --model ${model_q4_1} -f ${wiki_test_60} -ngl 99 -c 128 -b 128 --chunks 1 ) 2>&1 | tee -a $OUT/${ci}-tg-q4_1.log
|
||||
(time ./bin/llama-perplexity --model ${model_q5_0} -f ${wiki_test_60} -ngl 99 -c 128 -b 128 --chunks 1 ) 2>&1 | tee -a $OUT/${ci}-tg-q5_0.log
|
||||
(time ./bin/llama-perplexity --model ${model_q5_1} -f ${wiki_test_60} -ngl 99 -c 128 -b 128 --chunks 1 ) 2>&1 | tee -a $OUT/${ci}-tg-q5_1.log
|
||||
(time ./bin/llama-perplexity --model ${model_q2_k} -f ${wiki_test_60} -ngl 99 -c 128 -b 128 --chunks 1 ) 2>&1 | tee -a $OUT/${ci}-tg-q2_k.log
|
||||
(time ./bin/llama-perplexity --model ${model_q3_k} -f ${wiki_test_60} -ngl 99 -c 128 -b 128 --chunks 1 ) 2>&1 | tee -a $OUT/${ci}-tg-q3_k.log
|
||||
(time ./bin/llama-perplexity --model ${model_q4_k} -f ${wiki_test_60} -ngl 99 -c 128 -b 128 --chunks 1 ) 2>&1 | tee -a $OUT/${ci}-tg-q4_k.log
|
||||
(time ./bin/llama-perplexity --model ${model_q5_k} -f ${wiki_test_60} -ngl 99 -c 128 -b 128 --chunks 1 ) 2>&1 | tee -a $OUT/${ci}-tg-q5_k.log
|
||||
(time ./bin/llama-perplexity --model ${model_q6_k} -f ${wiki_test_60} -ngl 99 -c 128 -b 128 --chunks 1 ) 2>&1 | tee -a $OUT/${ci}-tg-q6_k.log
|
||||
|
||||
(time ./bin/llama-imatrix --model ${model_f16} -f ${wiki_test_60} -ngl 99 -c 128 -b 128 --chunks 1 ) 2>&1 | tee -a $OUT/${ci}-imatrix.log
|
||||
|
||||
(time ./bin/llama-save-load-state --model ${model_q4_0} -ngl 99 -c 0 ) 2>&1 | tee -a $OUT/${ci}-save-load-state.log
|
||||
(time ./bin/llama-save-load-state --model ${model_q4_0} -ngl 99 -c 0 -fa ) 2>&1 | tee -a $OUT/${ci}-save-load-state.log
|
||||
|
||||
function check_ppl {
|
||||
# lora
|
||||
function compare_ppl {
|
||||
qnt="$1"
|
||||
ppl=$(echo "$2" | grep -oE "[0-9]+\.[0-9]+" | tail -n 1)
|
||||
ppl1=$(echo "$2" | grep -oE "[0-9]+\.[0-9]+" | tail -n 1)
|
||||
ppl2=$(echo "$3" | grep -oE "[0-9]+\.[0-9]+" | tail -n 1)
|
||||
|
||||
if [ $(echo "$ppl > 20.0" | bc) -eq 1 ]; then
|
||||
printf ' - %s @ %s (FAIL: ppl > 20.0)\n' "$qnt" "$ppl"
|
||||
if [ $(echo "$ppl1 < $ppl2" | bc) -eq 1 ]; then
|
||||
printf ' - %s @ %s (FAIL: %s > %s)\n' "$qnt" "$ppl" "$ppl1" "$ppl2"
|
||||
return 20
|
||||
fi
|
||||
|
||||
printf ' - %s @ %s OK\n' "$qnt" "$ppl"
|
||||
printf ' - %s @ %s %s OK\n' "$qnt" "$ppl1" "$ppl2"
|
||||
return 0
|
||||
}
|
||||
|
||||
check_ppl "f16" "$(cat $OUT/${ci}-tg-f16.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log
|
||||
check_ppl "q8_0" "$(cat $OUT/${ci}-tg-q8_0.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log
|
||||
check_ppl "q4_0" "$(cat $OUT/${ci}-tg-q4_0.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log
|
||||
check_ppl "q4_1" "$(cat $OUT/${ci}-tg-q4_1.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log
|
||||
check_ppl "q5_0" "$(cat $OUT/${ci}-tg-q5_0.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log
|
||||
check_ppl "q5_1" "$(cat $OUT/${ci}-tg-q5_1.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log
|
||||
#check_ppl "q2_k" "$(cat $OUT/${ci}-tg-q2_k.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log # note: ppl > 20.0 for this quant and model
|
||||
check_ppl "q3_k" "$(cat $OUT/${ci}-tg-q3_k.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log
|
||||
check_ppl "q4_k" "$(cat $OUT/${ci}-tg-q4_k.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log
|
||||
check_ppl "q5_k" "$(cat $OUT/${ci}-tg-q5_k.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log
|
||||
check_ppl "q6_k" "$(cat $OUT/${ci}-tg-q6_k.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log
|
||||
path_lora="../models-mnt/open-llama/3B-v2/lora"
|
||||
path_shakespeare="../models-mnt/shakespeare"
|
||||
|
||||
shakespeare="${path_shakespeare}/shakespeare.txt"
|
||||
lora_shakespeare="${path_lora}/ggml-adapter-model.bin"
|
||||
|
||||
gg_wget ${path_lora} https://huggingface.co/slaren/open_llama_3b_v2_shakespeare_lora/resolve/main/adapter_config.json
|
||||
gg_wget ${path_lora} https://huggingface.co/slaren/open_llama_3b_v2_shakespeare_lora/resolve/main/adapter_model.bin
|
||||
gg_wget ${path_shakespeare} https://huggingface.co/slaren/open_llama_3b_v2_shakespeare_lora/resolve/main/shakespeare.txt
|
||||
|
||||
python3 ../convert-lora-to-ggml.py ${path_lora}
|
||||
|
||||
# f16
|
||||
(time ./bin/perplexity --model ${model_f16} -f ${shakespeare} -c 128 -b 128 --chunks 2 ) 2>&1 | tee -a $OUT/${ci}-ppl-shakespeare-f16.log
|
||||
(time ./bin/perplexity --model ${model_f16} -f ${shakespeare} --lora ${lora_shakespeare} -c 128 -b 128 --chunks 2 ) 2>&1 | tee -a $OUT/${ci}-ppl-shakespeare-lora-f16.log
|
||||
compare_ppl "f16 shakespeare" "$(cat $OUT/${ci}-ppl-shakespeare-f16.log | grep "^\[1\]")" "$(cat $OUT/${ci}-ppl-shakespeare-lora-f16.log | grep "^\[1\]")" | tee -a $OUT/${ci}-lora-ppl.log
|
||||
|
||||
# q8_0
|
||||
(time ./bin/perplexity --model ${model_q8_0} -f ${shakespeare} -c 128 -b 128 --chunks 2 ) 2>&1 | tee -a $OUT/${ci}-ppl-shakespeare-q8_0.log
|
||||
(time ./bin/perplexity --model ${model_q8_0} -f ${shakespeare} --lora ${lora_shakespeare} -c 128 -b 128 --chunks 2 ) 2>&1 | tee -a $OUT/${ci}-ppl-shakespeare-lora-q8_0.log
|
||||
compare_ppl "q8_0 shakespeare" "$(cat $OUT/${ci}-ppl-shakespeare-q8_0.log | grep "^\[1\]")" "$(cat $OUT/${ci}-ppl-shakespeare-lora-q8_0.log | grep "^\[1\]")" | tee -a $OUT/${ci}-lora-ppl.log
|
||||
|
||||
# q8_0 + f16 lora-base
|
||||
(time ./bin/perplexity --model ${model_q8_0} -f ${shakespeare} --lora ${lora_shakespeare} --lora-base ${model_f16} -c 128 -b 128 --chunks 2 ) 2>&1 | tee -a $OUT/${ci}-ppl-shakespeare-lora-q8_0-f16.log
|
||||
compare_ppl "q8_0 / f16 base shakespeare" "$(cat $OUT/${ci}-ppl-shakespeare-q8_0.log | grep "^\[1\]")" "$(cat $OUT/${ci}-ppl-shakespeare-lora-q8_0-f16.log | grep "^\[1\]")" | tee -a $OUT/${ci}-lora-ppl.log
|
||||
|
||||
cat $OUT/${ci}-imatrix.log | grep "Final" >> $OUT/${ci}-imatrix-sum.log
|
||||
|
||||
set +e
|
||||
}
|
||||
|
||||
function gg_sum_pythia_1_4b {
|
||||
function gg_sum_open_llama_3b_v2 {
|
||||
gg_printf '### %s\n\n' "${ci}"
|
||||
|
||||
gg_printf 'Pythia 1.4B:\n'
|
||||
gg_printf 'OpenLLaMA 3B-v2:\n'
|
||||
gg_printf '- status: %s\n' "$(cat $OUT/${ci}.exit)"
|
||||
gg_printf '- perplexity:\n%s\n' "$(cat $OUT/${ci}-ppl.log)"
|
||||
gg_printf '- imatrix:\n```\n%s\n```\n' "$(cat $OUT/${ci}-imatrix-sum.log)"
|
||||
gg_printf '- lora:\n%s\n' "$(cat $OUT/${ci}-lora-ppl.log)"
|
||||
gg_printf '- f16: \n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-f16.log)"
|
||||
gg_printf '- q8_0:\n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-q8_0.log)"
|
||||
gg_printf '- q4_0:\n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-q4_0.log)"
|
||||
|
@ -538,33 +299,42 @@ function gg_sum_pythia_1_4b {
|
|||
gg_printf '- q5_k:\n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-q5_k.log)"
|
||||
gg_printf '- q6_k:\n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-q6_k.log)"
|
||||
gg_printf '- save-load-state: \n```\n%s\n```\n' "$(cat $OUT/${ci}-save-load-state.log)"
|
||||
gg_printf '- shakespeare (f16):\n```\n%s\n```\n' "$(cat $OUT/${ci}-ppl-shakespeare-f16.log)"
|
||||
gg_printf '- shakespeare (f16 lora):\n```\n%s\n```\n' "$(cat $OUT/${ci}-ppl-shakespeare-lora-f16.log)"
|
||||
gg_printf '- shakespeare (q8_0):\n```\n%s\n```\n' "$(cat $OUT/${ci}-ppl-shakespeare-q8_0.log)"
|
||||
gg_printf '- shakespeare (q8_0 lora):\n```\n%s\n```\n' "$(cat $OUT/${ci}-ppl-shakespeare-lora-q8_0.log)"
|
||||
gg_printf '- shakespeare (q8_0 / f16 base lora):\n```\n%s\n```\n' "$(cat $OUT/${ci}-ppl-shakespeare-lora-q8_0-f16.log)"
|
||||
}
|
||||
|
||||
# pythia_2_8b
|
||||
# open_llama_7b_v2
|
||||
# requires: GG_BUILD_CUDA
|
||||
|
||||
function gg_run_pythia_2_8b {
|
||||
function gg_run_open_llama_7b_v2 {
|
||||
cd ${SRC}
|
||||
|
||||
gg_wget models-mnt/pythia/2.8B/ https://huggingface.co/EleutherAI/pythia-2.8b/raw/main/config.json
|
||||
gg_wget models-mnt/pythia/2.8B/ https://huggingface.co/EleutherAI/pythia-2.8b/raw/main/tokenizer.json
|
||||
gg_wget models-mnt/pythia/2.8B/ https://huggingface.co/EleutherAI/pythia-2.8b/raw/main/tokenizer_config.json
|
||||
gg_wget models-mnt/pythia/2.8B/ https://huggingface.co/EleutherAI/pythia-2.8b/raw/main/special_tokens_map.json
|
||||
gg_wget models-mnt/pythia/2.8B/ https://huggingface.co/EleutherAI/pythia-2.8b/resolve/main/pytorch_model.bin
|
||||
gg_wget models-mnt/open-llama/7B-v2/ https://huggingface.co/openlm-research/open_llama_7b_v2/raw/main/config.json
|
||||
gg_wget models-mnt/open-llama/7B-v2/ https://huggingface.co/openlm-research/open_llama_7b_v2/resolve/main/tokenizer.model
|
||||
gg_wget models-mnt/open-llama/7B-v2/ https://huggingface.co/openlm-research/open_llama_7b_v2/raw/main/tokenizer_config.json
|
||||
gg_wget models-mnt/open-llama/7B-v2/ https://huggingface.co/openlm-research/open_llama_7b_v2/raw/main/special_tokens_map.json
|
||||
gg_wget models-mnt/open-llama/7B-v2/ https://huggingface.co/openlm-research/open_llama_7b_v2/raw/main/pytorch_model.bin.index.json
|
||||
gg_wget models-mnt/open-llama/7B-v2/ https://huggingface.co/openlm-research/open_llama_7b_v2/resolve/main/pytorch_model-00001-of-00002.bin
|
||||
gg_wget models-mnt/open-llama/7B-v2/ https://huggingface.co/openlm-research/open_llama_7b_v2/resolve/main/pytorch_model-00002-of-00002.bin
|
||||
gg_wget models-mnt/open-llama/7B-v2/ https://huggingface.co/openlm-research/open_llama_7b_v2/raw/main/generation_config.json
|
||||
|
||||
gg_wget models-mnt/wikitext/ https://huggingface.co/datasets/ggml-org/ci/resolve/main/wikitext-2-raw-v1.zip
|
||||
gg_wget models-mnt/wikitext/ https://s3.amazonaws.com/research.metamind.io/wikitext/wikitext-2-raw-v1.zip
|
||||
unzip -o models-mnt/wikitext/wikitext-2-raw-v1.zip -d models-mnt/wikitext/
|
||||
|
||||
path_models="../models-mnt/pythia/2.8B"
|
||||
path_models="../models-mnt/open-llama/7B-v2"
|
||||
path_wiki="../models-mnt/wikitext/wikitext-2-raw"
|
||||
|
||||
rm -rf build-ci-release && mkdir build-ci-release && cd build-ci-release
|
||||
|
||||
set -e
|
||||
|
||||
(time cmake -DCMAKE_BUILD_TYPE=Release ${CMAKE_EXTRA} .. ) 2>&1 | tee -a $OUT/${ci}-cmake.log
|
||||
(time make -j$(nproc) ) 2>&1 | tee -a $OUT/${ci}-make.log
|
||||
(time cmake -DCMAKE_BUILD_TYPE=Release -DLLAMA_CUBLAS=1 .. ) 2>&1 | tee -a $OUT/${ci}-cmake.log
|
||||
(time make -j ) 2>&1 | tee -a $OUT/${ci}-make.log
|
||||
|
||||
python3 ../convert_hf_to_gguf.py ${path_models} --outfile ${path_models}/ggml-model-f16.gguf
|
||||
python3 ../convert.py ${path_models}
|
||||
|
||||
model_f16="${path_models}/ggml-model-f16.gguf"
|
||||
model_q8_0="${path_models}/ggml-model-q8_0.gguf"
|
||||
|
@ -580,47 +350,42 @@ function gg_run_pythia_2_8b {
|
|||
|
||||
wiki_test="${path_wiki}/wiki.test.raw"
|
||||
|
||||
./bin/llama-quantize ${model_f16} ${model_q8_0} q8_0
|
||||
./bin/llama-quantize ${model_f16} ${model_q4_0} q4_0
|
||||
./bin/llama-quantize ${model_f16} ${model_q4_1} q4_1
|
||||
./bin/llama-quantize ${model_f16} ${model_q5_0} q5_0
|
||||
./bin/llama-quantize ${model_f16} ${model_q5_1} q5_1
|
||||
./bin/llama-quantize ${model_f16} ${model_q2_k} q2_k
|
||||
./bin/llama-quantize ${model_f16} ${model_q3_k} q3_k
|
||||
./bin/llama-quantize ${model_f16} ${model_q4_k} q4_k
|
||||
./bin/llama-quantize ${model_f16} ${model_q5_k} q5_k
|
||||
./bin/llama-quantize ${model_f16} ${model_q6_k} q6_k
|
||||
./bin/quantize ${model_f16} ${model_q8_0} q8_0
|
||||
./bin/quantize ${model_f16} ${model_q4_0} q4_0
|
||||
./bin/quantize ${model_f16} ${model_q4_1} q4_1
|
||||
./bin/quantize ${model_f16} ${model_q5_0} q5_0
|
||||
./bin/quantize ${model_f16} ${model_q5_1} q5_1
|
||||
./bin/quantize ${model_f16} ${model_q2_k} q2_k
|
||||
./bin/quantize ${model_f16} ${model_q3_k} q3_k
|
||||
./bin/quantize ${model_f16} ${model_q4_k} q4_k
|
||||
./bin/quantize ${model_f16} ${model_q5_k} q5_k
|
||||
./bin/quantize ${model_f16} ${model_q6_k} q6_k
|
||||
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_f16} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-f16.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q8_0} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q8_0.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q4_0} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_0.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q4_1} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_1.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q5_0} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_0.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q5_1} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_1.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q2_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q2_k.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q3_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q3_k.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q4_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_k.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q5_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_k.log
|
||||
(time ./bin/llama-cli -no-cnv --model ${model_q6_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q6_k.log
|
||||
(time ./bin/main --model ${model_f16} -t 1 -ngl 999 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-f16.log
|
||||
(time ./bin/main --model ${model_q8_0} -t 1 -ngl 999 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q8_0.log
|
||||
(time ./bin/main --model ${model_q4_0} -t 1 -ngl 999 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_0.log
|
||||
(time ./bin/main --model ${model_q4_1} -t 1 -ngl 999 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_1.log
|
||||
(time ./bin/main --model ${model_q5_0} -t 1 -ngl 999 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_0.log
|
||||
(time ./bin/main --model ${model_q5_1} -t 1 -ngl 999 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_1.log
|
||||
(time ./bin/main --model ${model_q2_k} -t 1 -ngl 999 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q2_k.log
|
||||
(time ./bin/main --model ${model_q3_k} -t 1 -ngl 999 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q3_k.log
|
||||
(time ./bin/main --model ${model_q4_k} -t 1 -ngl 999 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_k.log
|
||||
(time ./bin/main --model ${model_q5_k} -t 1 -ngl 999 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_k.log
|
||||
(time ./bin/main --model ${model_q6_k} -t 1 -ngl 999 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q6_k.log
|
||||
|
||||
(time ./bin/llama-perplexity --model ${model_f16} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-f16.log
|
||||
(time ./bin/llama-perplexity --model ${model_q8_0} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q8_0.log
|
||||
(time ./bin/llama-perplexity --model ${model_q4_0} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q4_0.log
|
||||
(time ./bin/llama-perplexity --model ${model_q4_1} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q4_1.log
|
||||
(time ./bin/llama-perplexity --model ${model_q5_0} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q5_0.log
|
||||
(time ./bin/llama-perplexity --model ${model_q5_1} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q5_1.log
|
||||
(time ./bin/llama-perplexity --model ${model_q2_k} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q2_k.log
|
||||
(time ./bin/llama-perplexity --model ${model_q3_k} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q3_k.log
|
||||
(time ./bin/llama-perplexity --model ${model_q4_k} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q4_k.log
|
||||
(time ./bin/llama-perplexity --model ${model_q5_k} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q5_k.log
|
||||
(time ./bin/llama-perplexity --model ${model_q6_k} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q6_k.log
|
||||
(time ./bin/perplexity --model ${model_f16} -f ${wiki_test} -t 1 -ngl 999 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-f16.log
|
||||
(time ./bin/perplexity --model ${model_q8_0} -f ${wiki_test} -t 1 -ngl 999 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q8_0.log
|
||||
(time ./bin/perplexity --model ${model_q4_0} -f ${wiki_test} -t 1 -ngl 999 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q4_0.log
|
||||
(time ./bin/perplexity --model ${model_q4_1} -f ${wiki_test} -t 1 -ngl 999 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q4_1.log
|
||||
(time ./bin/perplexity --model ${model_q5_0} -f ${wiki_test} -t 1 -ngl 999 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q5_0.log
|
||||
(time ./bin/perplexity --model ${model_q5_1} -f ${wiki_test} -t 1 -ngl 999 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q5_1.log
|
||||
(time ./bin/perplexity --model ${model_q2_k} -f ${wiki_test} -t 1 -ngl 999 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q2_k.log
|
||||
(time ./bin/perplexity --model ${model_q3_k} -f ${wiki_test} -t 1 -ngl 999 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q3_k.log
|
||||
(time ./bin/perplexity --model ${model_q4_k} -f ${wiki_test} -t 1 -ngl 999 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q4_k.log
|
||||
(time ./bin/perplexity --model ${model_q5_k} -f ${wiki_test} -t 1 -ngl 999 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q5_k.log
|
||||
(time ./bin/perplexity --model ${model_q6_k} -f ${wiki_test} -t 1 -ngl 999 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q6_k.log
|
||||
|
||||
(time ./bin/llama-imatrix --model ${model_f16} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-imatrix.log
|
||||
|
||||
(time ./bin/llama-save-load-state --model ${model_q4_0} -ngl 10 -c 0 ) 2>&1 | tee -a $OUT/${ci}-save-load-state.log
|
||||
(time ./bin/llama-save-load-state --model ${model_q4_0} -ngl 10 -c 0 -fa ) 2>&1 | tee -a $OUT/${ci}-save-load-state.log
|
||||
(time ./bin/llama-save-load-state --model ${model_q4_0} -ngl 99 -c 0 ) 2>&1 | tee -a $OUT/${ci}-save-load-state.log
|
||||
(time ./bin/llama-save-load-state --model ${model_q4_0} -ngl 99 -c 0 -fa ) 2>&1 | tee -a $OUT/${ci}-save-load-state.log
|
||||
(time ./bin/save-load-state --model ${model_q4_0} ) 2>&1 | tee -a $OUT/${ci}-save-load-state.log
|
||||
|
||||
function check_ppl {
|
||||
qnt="$1"
|
||||
|
@ -641,24 +406,64 @@ function gg_run_pythia_2_8b {
|
|||
check_ppl "q4_1" "$(cat $OUT/${ci}-tg-q4_1.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log
|
||||
check_ppl "q5_0" "$(cat $OUT/${ci}-tg-q5_0.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log
|
||||
check_ppl "q5_1" "$(cat $OUT/${ci}-tg-q5_1.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log
|
||||
#check_ppl "q2_k" "$(cat $OUT/${ci}-tg-q2_k.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log # note: ppl > 20.0 for this quant and model
|
||||
check_ppl "q2_k" "$(cat $OUT/${ci}-tg-q2_k.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log
|
||||
check_ppl "q3_k" "$(cat $OUT/${ci}-tg-q3_k.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log
|
||||
check_ppl "q4_k" "$(cat $OUT/${ci}-tg-q4_k.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log
|
||||
check_ppl "q5_k" "$(cat $OUT/${ci}-tg-q5_k.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log
|
||||
check_ppl "q6_k" "$(cat $OUT/${ci}-tg-q6_k.log | grep "^\[1\]")" | tee -a $OUT/${ci}-ppl.log
|
||||
|
||||
cat $OUT/${ci}-imatrix.log | grep "Final" >> $OUT/${ci}-imatrix-sum.log
|
||||
# lora
|
||||
function compare_ppl {
|
||||
qnt="$1"
|
||||
ppl1=$(echo "$2" | grep -oE "[0-9]+\.[0-9]+" | tail -n 1)
|
||||
ppl2=$(echo "$3" | grep -oE "[0-9]+\.[0-9]+" | tail -n 1)
|
||||
|
||||
if [ $(echo "$ppl1 < $ppl2" | bc) -eq 1 ]; then
|
||||
printf ' - %s @ %s (FAIL: %s > %s)\n' "$qnt" "$ppl" "$ppl1" "$ppl2"
|
||||
return 20
|
||||
fi
|
||||
|
||||
printf ' - %s @ %s %s OK\n' "$qnt" "$ppl1" "$ppl2"
|
||||
return 0
|
||||
}
|
||||
|
||||
path_lora="../models-mnt/open-llama/7B-v2/lora"
|
||||
path_shakespeare="../models-mnt/shakespeare"
|
||||
|
||||
shakespeare="${path_shakespeare}/shakespeare.txt"
|
||||
lora_shakespeare="${path_lora}/ggml-adapter-model.bin"
|
||||
|
||||
gg_wget ${path_lora} https://huggingface.co/slaren/open_llama_7b_v2_shakespeare_lora/resolve/main/adapter_config.json
|
||||
gg_wget ${path_lora} https://huggingface.co/slaren/open_llama_7b_v2_shakespeare_lora/resolve/main/adapter_model.bin
|
||||
gg_wget ${path_shakespeare} https://huggingface.co/slaren/open_llama_7b_v2_shakespeare_lora/resolve/main/shakespeare.txt
|
||||
|
||||
python3 ../convert-lora-to-ggml.py ${path_lora}
|
||||
|
||||
# f16
|
||||
(time ./bin/perplexity --model ${model_f16} -f ${shakespeare} -t 1 -ngl 999 -c 2048 -b 512 --chunks 3 ) 2>&1 | tee -a $OUT/${ci}-ppl-shakespeare-f16.log
|
||||
(time ./bin/perplexity --model ${model_f16} -f ${shakespeare} --lora ${lora_shakespeare} -t 1 -ngl 999 -c 2048 -b 512 --chunks 3 ) 2>&1 | tee -a $OUT/${ci}-ppl-shakespeare-lora-f16.log
|
||||
compare_ppl "f16 shakespeare" "$(cat $OUT/${ci}-ppl-shakespeare-f16.log | grep "^\[1\]")" "$(cat $OUT/${ci}-ppl-shakespeare-lora-f16.log | grep "^\[1\]")" | tee -a $OUT/${ci}-lora-ppl.log
|
||||
|
||||
# currently not supported by the CUDA backend
|
||||
# q8_0
|
||||
#(time ./bin/perplexity --model ${model_q8_0} -f ${shakespeare} -t 1 -ngl 999 -c 2048 -b 512 --chunks 3 ) 2>&1 | tee -a $OUT/${ci}-ppl-shakespeare-q8_0.log
|
||||
#(time ./bin/perplexity --model ${model_q8_0} -f ${shakespeare} --lora ${lora_shakespeare} -t 1 -ngl 999 -c 2048 -b 512 --chunks 3 ) 2>&1 | tee -a $OUT/${ci}-ppl-shakespeare-lora-q8_0.log
|
||||
#compare_ppl "q8_0 shakespeare" "$(cat $OUT/${ci}-ppl-shakespeare-q8_0.log | grep "^\[1\]")" "$(cat $OUT/${ci}-ppl-shakespeare-lora-q8_0.log | grep "^\[1\]")" | tee -a $OUT/${ci}-lora-ppl.log
|
||||
|
||||
# q8_0 + f16 lora-base
|
||||
#(time ./bin/perplexity --model ${model_q8_0} -f ${shakespeare} --lora ${lora_shakespeare} --lora-base ${model_f16} -t 1 -ngl 999 -c 2048 -b 512 --chunks 3 ) 2>&1 | tee -a $OUT/${ci}-ppl-shakespeare-lora-q8_0-f16.log
|
||||
#compare_ppl "q8_0 / f16 shakespeare" "$(cat $OUT/${ci}-ppl-shakespeare-q8_0.log | grep "^\[1\]")" "$(cat $OUT/${ci}-ppl-shakespeare-lora-q8_0-f16.log | grep "^\[1\]")" | tee -a $OUT/${ci}-lora-ppl.log
|
||||
|
||||
set +e
|
||||
}
|
||||
|
||||
function gg_sum_pythia_2_8b {
|
||||
function gg_sum_open_llama_7b_v2 {
|
||||
gg_printf '### %s\n\n' "${ci}"
|
||||
|
||||
gg_printf 'Pythia 2.8B:\n'
|
||||
gg_printf 'OpenLLaMA 7B-v2:\n'
|
||||
gg_printf '- status: %s\n' "$(cat $OUT/${ci}.exit)"
|
||||
gg_printf '- perplexity:\n%s\n' "$(cat $OUT/${ci}-ppl.log)"
|
||||
gg_printf '- imatrix:\n```\n%s\n```\n' "$(cat $OUT/${ci}-imatrix-sum.log)"
|
||||
gg_printf '- lora:\n%s\n' "$(cat $OUT/${ci}-lora-ppl.log)"
|
||||
gg_printf '- f16: \n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-f16.log)"
|
||||
gg_printf '- q8_0:\n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-q8_0.log)"
|
||||
gg_printf '- q4_0:\n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-q4_0.log)"
|
||||
|
@ -671,158 +476,24 @@ function gg_sum_pythia_2_8b {
|
|||
gg_printf '- q5_k:\n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-q5_k.log)"
|
||||
gg_printf '- q6_k:\n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-q6_k.log)"
|
||||
gg_printf '- save-load-state: \n```\n%s\n```\n' "$(cat $OUT/${ci}-save-load-state.log)"
|
||||
}
|
||||
|
||||
# bge-small
|
||||
|
||||
function gg_run_embd_bge_small {
|
||||
cd ${SRC}
|
||||
|
||||
gg_wget models-mnt/bge-small/ https://huggingface.co/BAAI/bge-small-en-v1.5/raw/main/config.json
|
||||
gg_wget models-mnt/bge-small/ https://huggingface.co/BAAI/bge-small-en-v1.5/raw/main/tokenizer.json
|
||||
gg_wget models-mnt/bge-small/ https://huggingface.co/BAAI/bge-small-en-v1.5/raw/main/tokenizer_config.json
|
||||
gg_wget models-mnt/bge-small/ https://huggingface.co/BAAI/bge-small-en-v1.5/raw/main/special_tokens_map.json
|
||||
gg_wget models-mnt/bge-small/ https://huggingface.co/BAAI/bge-small-en-v1.5/resolve/main/pytorch_model.bin
|
||||
gg_wget models-mnt/bge-small/ https://huggingface.co/BAAI/bge-small-en-v1.5/raw/main/sentence_bert_config.json
|
||||
gg_wget models-mnt/bge-small/ https://huggingface.co/BAAI/bge-small-en-v1.5/raw/main/vocab.txt
|
||||
gg_wget models-mnt/bge-small/ https://huggingface.co/BAAI/bge-small-en-v1.5/raw/main/modules.json
|
||||
gg_wget models-mnt/bge-small/ https://huggingface.co/BAAI/bge-small-en-v1.5/raw/main/config.json
|
||||
|
||||
gg_wget models-mnt/bge-small/1_Pooling https://huggingface.co/BAAI/bge-small-en-v1.5/raw/main/1_Pooling/config.json
|
||||
|
||||
path_models="../models-mnt/bge-small"
|
||||
|
||||
rm -rf build-ci-release && mkdir build-ci-release && cd build-ci-release
|
||||
|
||||
set -e
|
||||
|
||||
(time cmake -DCMAKE_BUILD_TYPE=Release ${CMAKE_EXTRA} .. ) 2>&1 | tee -a $OUT/${ci}-cmake.log
|
||||
(time make -j$(nproc) ) 2>&1 | tee -a $OUT/${ci}-make.log
|
||||
|
||||
python3 ../convert_hf_to_gguf.py ${path_models} --outfile ${path_models}/ggml-model-f16.gguf
|
||||
|
||||
model_f16="${path_models}/ggml-model-f16.gguf"
|
||||
model_q8_0="${path_models}/ggml-model-q8_0.gguf"
|
||||
|
||||
./bin/llama-quantize ${model_f16} ${model_q8_0} q8_0
|
||||
|
||||
(time ./bin/llama-embedding --model ${model_f16} -p "I believe the meaning of life is" -ngl 99 -c 0 ) 2>&1 | tee -a $OUT/${ci}-tg-f16.log
|
||||
(time ./bin/llama-embedding --model ${model_q8_0} -p "I believe the meaning of life is" -ngl 99 -c 0 ) 2>&1 | tee -a $OUT/${ci}-tg-q8_0.log
|
||||
|
||||
set +e
|
||||
}
|
||||
|
||||
function gg_sum_embd_bge_small {
|
||||
gg_printf '### %s\n\n' "${ci}"
|
||||
|
||||
gg_printf 'BGE Small (BERT):\n'
|
||||
gg_printf '- status: %s\n' "$(cat $OUT/${ci}.exit)"
|
||||
gg_printf '- f16: \n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-f16.log)"
|
||||
gg_printf '- q8_0:\n```\n%s\n```\n' "$(cat $OUT/${ci}-tg-q8_0.log)"
|
||||
}
|
||||
|
||||
# rerank_tiny
|
||||
|
||||
function gg_run_rerank_tiny {
|
||||
cd ${SRC}
|
||||
|
||||
gg_wget models-mnt/rerank-tiny/ https://huggingface.co/jinaai/jina-reranker-v1-tiny-en/raw/main/config.json
|
||||
gg_wget models-mnt/rerank-tiny/ https://huggingface.co/jinaai/jina-reranker-v1-tiny-en/raw/main/tokenizer.json
|
||||
gg_wget models-mnt/rerank-tiny/ https://huggingface.co/jinaai/jina-reranker-v1-tiny-en/raw/main/tokenizer_config.json
|
||||
gg_wget models-mnt/rerank-tiny/ https://huggingface.co/jinaai/jina-reranker-v1-tiny-en/raw/main/special_tokens_map.json
|
||||
gg_wget models-mnt/rerank-tiny/ https://huggingface.co/jinaai/jina-reranker-v1-tiny-en/resolve/main/pytorch_model.bin
|
||||
gg_wget models-mnt/rerank-tiny/ https://huggingface.co/jinaai/jina-reranker-v1-tiny-en/raw/main/sentence_bert_config.json
|
||||
gg_wget models-mnt/rerank-tiny/ https://huggingface.co/jinaai/jina-reranker-v1-tiny-en/raw/main/vocab.txt
|
||||
gg_wget models-mnt/rerank-tiny/ https://huggingface.co/jinaai/jina-reranker-v1-tiny-en/raw/main/modules.json
|
||||
gg_wget models-mnt/rerank-tiny/ https://huggingface.co/jinaai/jina-reranker-v1-tiny-en/raw/main/config.json
|
||||
|
||||
gg_wget models-mnt/rerank-tiny/1_Pooling https://huggingface.co/jinaai/jina-reranker-v1-tiny-en/raw/main/1_Pooling/config.json
|
||||
|
||||
path_models="../models-mnt/rerank-tiny"
|
||||
|
||||
rm -rf build-ci-release && mkdir build-ci-release && cd build-ci-release
|
||||
|
||||
set -e
|
||||
|
||||
(time cmake -DCMAKE_BUILD_TYPE=Release ${CMAKE_EXTRA} .. ) 2>&1 | tee -a $OUT/${ci}-cmake.log
|
||||
(time make -j$(nproc) ) 2>&1 | tee -a $OUT/${ci}-make.log
|
||||
|
||||
python3 ../convert_hf_to_gguf.py ${path_models} --outfile ${path_models}/ggml-model-f16.gguf
|
||||
|
||||
model_f16="${path_models}/ggml-model-f16.gguf"
|
||||
|
||||
# for this model, the SEP token is "</s>"
|
||||
(time ./bin/llama-embedding --model ${model_f16} -p "what is panda?</s></s>hi\nwhat is panda?</s></s>it's a bear\nwhat is panda?</s></s>The giant panda (Ailuropoda melanoleuca), sometimes called a panda bear or simply panda, is a bear species endemic to China." -ngl 99 -c 0 --pooling rank --embd-normalize -1 --verbose-prompt) 2>&1 | tee -a $OUT/${ci}-rk-f16.log
|
||||
|
||||
# sample output
|
||||
# rerank score 0: 0.029
|
||||
# rerank score 1: 0.029
|
||||
# rerank score 2: 0.135
|
||||
|
||||
# check that the score is in the range [$3, $4]
|
||||
function check_score {
|
||||
qnt="$1"
|
||||
score=$(echo "$2" | grep -oE "[0-9]+\.[0-9]+" | tail -n 1)
|
||||
|
||||
if [ $(echo "$score < $3" | bc) -eq 1 ] || [ $(echo "$score > $4" | bc) -eq 1 ]; then
|
||||
printf ' - %s @ %s (FAIL: score not in range [%s, %s])\n' "$qnt" "$score" "$3" "$4"
|
||||
return 20
|
||||
fi
|
||||
|
||||
printf ' - %s @ %s OK\n' "$qnt" "$score"
|
||||
return 0
|
||||
}
|
||||
|
||||
check_score "rerank score 0" "$(cat $OUT/${ci}-rk-f16.log | grep "rerank score 0")" "0.00" "0.05" | tee -a $OUT/${ci}-rk-f16.log
|
||||
check_score "rerank score 1" "$(cat $OUT/${ci}-rk-f16.log | grep "rerank score 1")" "0.00" "0.05" | tee -a $OUT/${ci}-rk-f16.log
|
||||
check_score "rerank score 2" "$(cat $OUT/${ci}-rk-f16.log | grep "rerank score 2")" "0.10" "0.30" | tee -a $OUT/${ci}-rk-f16.log
|
||||
|
||||
set +e
|
||||
}
|
||||
|
||||
function gg_sum_rerank_tiny {
|
||||
gg_printf '### %s\n\n' "${ci}"
|
||||
|
||||
gg_printf 'Rerank Tiny (Jina):\n'
|
||||
gg_printf '- status: %s\n' "$(cat $OUT/${ci}.exit)"
|
||||
gg_printf '- f16: \n```\n%s\n```\n' "$(cat $OUT/${ci}-rk-f16.log)"
|
||||
}
|
||||
|
||||
function gg_check_build_requirements {
|
||||
if ! command -v cmake &> /dev/null; then
|
||||
gg_printf 'cmake not found, please install'
|
||||
fi
|
||||
|
||||
if ! command -v make &> /dev/null; then
|
||||
gg_printf 'make not found, please install'
|
||||
fi
|
||||
|
||||
if ! command -v ctest &> /dev/null; then
|
||||
gg_printf 'ctest not found, please install'
|
||||
fi
|
||||
gg_printf '- shakespeare (f16):\n```\n%s\n```\n' "$(cat $OUT/${ci}-ppl-shakespeare-f16.log)"
|
||||
gg_printf '- shakespeare (f16 lora):\n```\n%s\n```\n' "$(cat $OUT/${ci}-ppl-shakespeare-lora-f16.log)"
|
||||
#gg_printf '- shakespeare (q8_0):\n```\n%s\n```\n' "$(cat $OUT/${ci}-ppl-shakespeare-q8_0.log)"
|
||||
#gg_printf '- shakespeare (q8_0 lora):\n```\n%s\n```\n' "$(cat $OUT/${ci}-ppl-shakespeare-lora-q8_0.log)"
|
||||
#gg_printf '- shakespeare (q8_0 / f16 base lora):\n```\n%s\n```\n' "$(cat $OUT/${ci}-ppl-shakespeare-lora-q8_0-f16.log)"
|
||||
}
|
||||
|
||||
## main
|
||||
|
||||
export LLAMA_LOG_PREFIX=1
|
||||
export LLAMA_LOG_TIMESTAMPS=1
|
||||
|
||||
if [ -z ${GG_BUILD_LOW_PERF} ]; then
|
||||
# Create symlink: ./llama.cpp/models-mnt -> $MNT/models/models-mnt
|
||||
rm -rf ${SRC}/models-mnt
|
||||
|
||||
mnt_models=${MNT}/models
|
||||
mkdir -p ${mnt_models}
|
||||
ln -sfn ${mnt_models} ${SRC}/models-mnt
|
||||
|
||||
# Create a fresh python3 venv and enter it
|
||||
if ! python3 -m venv "$MNT/venv"; then
|
||||
echo "Error: Failed to create Python virtual environment at $MNT/venv."
|
||||
exit 1
|
||||
fi
|
||||
source "$MNT/venv/bin/activate"
|
||||
|
||||
pip install -r ${SRC}/requirements.txt --disable-pip-version-check
|
||||
pip install --editable gguf-py --disable-pip-version-check
|
||||
python3 -m pip install -r ${SRC}/requirements.txt
|
||||
python3 -m pip install --editable gguf-py
|
||||
fi
|
||||
|
||||
ret=0
|
||||
|
@ -831,23 +502,12 @@ test $ret -eq 0 && gg_run ctest_debug
|
|||
test $ret -eq 0 && gg_run ctest_release
|
||||
|
||||
if [ -z ${GG_BUILD_LOW_PERF} ]; then
|
||||
test $ret -eq 0 && gg_run embd_bge_small
|
||||
test $ret -eq 0 && gg_run rerank_tiny
|
||||
|
||||
if [ -z ${GG_BUILD_CLOUD} ] || [ ${GG_BUILD_EXTRA_TESTS_0} ]; then
|
||||
test $ret -eq 0 && gg_run test_scripts_debug
|
||||
test $ret -eq 0 && gg_run test_scripts_release
|
||||
fi
|
||||
|
||||
if [ -z ${GG_BUILD_VRAM_GB} ] || [ ${GG_BUILD_VRAM_GB} -ge 8 ]; then
|
||||
if [ -z ${GG_BUILD_CUDA} ] && [ -z ${GG_BUILD_VULKAN} ]; then
|
||||
test $ret -eq 0 && gg_run pythia_1_4b
|
||||
if [ -z ${GG_BUILD_CUDA} ]; then
|
||||
test $ret -eq 0 && gg_run open_llama_3b_v2
|
||||
else
|
||||
test $ret -eq 0 && gg_run pythia_2_8b
|
||||
#test $ret -eq 0 && gg_run open_llama_7b_v2
|
||||
test $ret -eq 0 && gg_run open_llama_7b_v2
|
||||
fi
|
||||
test $ret -eq 0 && gg_run ctest_with_model_debug
|
||||
test $ret -eq 0 && gg_run ctest_with_model_release
|
||||
fi
|
||||
fi
|
||||
|
||||
|
|
|
@ -79,22 +79,22 @@ endmacro()
|
|||
# flags are for MSVC only!
|
||||
check_sse("AVX" " ;/arch:AVX")
|
||||
if (NOT ${AVX_FOUND})
|
||||
set(GGML_AVX OFF)
|
||||
set(LLAMA_AVX OFF)
|
||||
else()
|
||||
set(GGML_AVX ON)
|
||||
set(LLAMA_AVX ON)
|
||||
endif()
|
||||
|
||||
check_sse("AVX2" " ;/arch:AVX2")
|
||||
check_sse("FMA" " ;/arch:AVX2")
|
||||
if ((NOT ${AVX2_FOUND}) OR (NOT ${FMA_FOUND}))
|
||||
set(GGML_AVX2 OFF)
|
||||
set(LLAMA_AVX2 OFF)
|
||||
else()
|
||||
set(GGML_AVX2 ON)
|
||||
set(LLAMA_AVX2 ON)
|
||||
endif()
|
||||
|
||||
check_sse("AVX512" " ;/arch:AVX512")
|
||||
if (NOT ${AVX512_FOUND})
|
||||
set(GGML_AVX512 OFF)
|
||||
set(LLAMA_AVX512 OFF)
|
||||
else()
|
||||
set(GGML_AVX512 ON)
|
||||
set(LLAMA_AVX512 ON)
|
||||
endif()
|
|
@ -1,16 +0,0 @@
|
|||
set( CMAKE_SYSTEM_NAME Darwin )
|
||||
set( CMAKE_SYSTEM_PROCESSOR arm64 )
|
||||
|
||||
set( target arm64-apple-darwin-macho )
|
||||
|
||||
set( CMAKE_C_COMPILER clang )
|
||||
set( CMAKE_CXX_COMPILER clang++ )
|
||||
|
||||
set( CMAKE_C_COMPILER_TARGET ${target} )
|
||||
set( CMAKE_CXX_COMPILER_TARGET ${target} )
|
||||
|
||||
set( arch_c_flags "-march=armv8.4-a -fvectorize -ffp-model=fast -fno-finite-math-only" )
|
||||
set( warn_c_flags "-Wno-format -Wno-unused-variable -Wno-unused-function" )
|
||||
|
||||
set( CMAKE_C_FLAGS_INIT "${arch_c_flags} ${warn_c_flags}" )
|
||||
set( CMAKE_CXX_FLAGS_INIT "${arch_c_flags} ${warn_c_flags}" )
|
|
@ -1,16 +0,0 @@
|
|||
set( CMAKE_SYSTEM_NAME Windows )
|
||||
set( CMAKE_SYSTEM_PROCESSOR arm64 )
|
||||
|
||||
set( target arm64-pc-windows-msvc )
|
||||
|
||||
set( CMAKE_C_COMPILER clang )
|
||||
set( CMAKE_CXX_COMPILER clang++ )
|
||||
|
||||
set( CMAKE_C_COMPILER_TARGET ${target} )
|
||||
set( CMAKE_CXX_COMPILER_TARGET ${target} )
|
||||
|
||||
set( arch_c_flags "-march=armv8.7-a -fvectorize -ffp-model=fast -fno-finite-math-only" )
|
||||
set( warn_c_flags "-Wno-format -Wno-unused-variable -Wno-unused-function -Wno-gnu-zero-variadic-macro-arguments" )
|
||||
|
||||
set( CMAKE_C_FLAGS_INIT "${arch_c_flags} ${warn_c_flags}" )
|
||||
set( CMAKE_CXX_FLAGS_INIT "${arch_c_flags} ${warn_c_flags}" )
|
|
@ -1,6 +0,0 @@
|
|||
set( CMAKE_SYSTEM_NAME Windows )
|
||||
set( CMAKE_SYSTEM_PROCESSOR arm64 )
|
||||
|
||||
set( target arm64-pc-windows-msvc )
|
||||
set( CMAKE_C_COMPILER_TARGET ${target} )
|
||||
set( CMAKE_CXX_COMPILER_TARGET ${target} )
|
|
@ -1,33 +0,0 @@
|
|||
function(llama_add_compile_flags)
|
||||
if (LLAMA_FATAL_WARNINGS)
|
||||
if (CMAKE_CXX_COMPILER_ID MATCHES "GNU" OR CMAKE_CXX_COMPILER_ID MATCHES "Clang")
|
||||
list(APPEND C_FLAGS -Werror)
|
||||
list(APPEND CXX_FLAGS -Werror)
|
||||
elseif (CMAKE_CXX_COMPILER_ID STREQUAL "MSVC")
|
||||
add_compile_options(/WX)
|
||||
endif()
|
||||
endif()
|
||||
|
||||
if (LLAMA_ALL_WARNINGS)
|
||||
if (NOT MSVC)
|
||||
list(APPEND C_FLAGS -Wshadow -Wstrict-prototypes -Wpointer-arith -Wmissing-prototypes
|
||||
-Werror=implicit-int -Werror=implicit-function-declaration)
|
||||
|
||||
list(APPEND CXX_FLAGS -Wmissing-declarations -Wmissing-noreturn)
|
||||
|
||||
list(APPEND WARNING_FLAGS -Wall -Wextra -Wpedantic -Wcast-qual -Wno-unused-function)
|
||||
|
||||
list(APPEND C_FLAGS ${WARNING_FLAGS})
|
||||
list(APPEND CXX_FLAGS ${WARNING_FLAGS})
|
||||
|
||||
ggml_get_flags(${CMAKE_CXX_COMPILER_ID} ${CMAKE_CXX_COMPILER_VERSION})
|
||||
|
||||
add_compile_options("$<$<COMPILE_LANGUAGE:C>:${C_FLAGS};${GF_C_FLAGS}>"
|
||||
"$<$<COMPILE_LANGUAGE:CXX>:${CXX_FLAGS};${GF_CXX_FLAGS}>")
|
||||
else()
|
||||
# todo : msvc
|
||||
set(C_FLAGS "" PARENT_SCOPE)
|
||||
set(CXX_FLAGS "" PARENT_SCOPE)
|
||||
endif()
|
||||
endif()
|
||||
endfunction()
|
|
@ -1,22 +0,0 @@
|
|||
find_package(Git)
|
||||
|
||||
# the commit's SHA1
|
||||
execute_process(COMMAND
|
||||
"${GIT_EXECUTABLE}" describe --match=NeVeRmAtCh --always --abbrev=8
|
||||
WORKING_DIRECTORY "${CMAKE_SOURCE_DIR}"
|
||||
OUTPUT_VARIABLE GIT_SHA1
|
||||
ERROR_QUIET OUTPUT_STRIP_TRAILING_WHITESPACE)
|
||||
|
||||
# the date of the commit
|
||||
execute_process(COMMAND
|
||||
"${GIT_EXECUTABLE}" log -1 --format=%ad --date=local
|
||||
WORKING_DIRECTORY "${CMAKE_SOURCE_DIR}"
|
||||
OUTPUT_VARIABLE GIT_DATE
|
||||
ERROR_QUIET OUTPUT_STRIP_TRAILING_WHITESPACE)
|
||||
|
||||
# the subject of the commit
|
||||
execute_process(COMMAND
|
||||
"${GIT_EXECUTABLE}" log -1 --format=%s
|
||||
WORKING_DIRECTORY "${CMAKE_SOURCE_DIR}"
|
||||
OUTPUT_VARIABLE GIT_COMMIT_SUBJECT
|
||||
ERROR_QUIET OUTPUT_STRIP_TRAILING_WHITESPACE)
|
|
@ -1,30 +0,0 @@
|
|||
set(LLAMA_VERSION @LLAMA_INSTALL_VERSION@)
|
||||
set(LLAMA_BUILD_COMMIT @LLAMA_BUILD_COMMIT@)
|
||||
set(LLAMA_BUILD_NUMBER @LLAMA_BUILD_NUMBER@)
|
||||
set(LLAMA_SHARED_LIB @BUILD_SHARED_LIBS@)
|
||||
|
||||
@PACKAGE_INIT@
|
||||
|
||||
set_and_check(LLAMA_INCLUDE_DIR "@PACKAGE_LLAMA_INCLUDE_INSTALL_DIR@")
|
||||
set_and_check(LLAMA_LIB_DIR "@PACKAGE_LLAMA_LIB_INSTALL_DIR@")
|
||||
set_and_check(LLAMA_BIN_DIR "@PACKAGE_LLAMA_BIN_INSTALL_DIR@")
|
||||
|
||||
find_package(ggml REQUIRED HINTS ${LLAMA_LIB_DIR}/cmake)
|
||||
|
||||
find_library(llama_LIBRARY llama
|
||||
REQUIRED
|
||||
HINTS ${LLAMA_LIB_DIR}
|
||||
NO_CMAKE_FIND_ROOT_PATH
|
||||
)
|
||||
|
||||
add_library(llama UNKNOWN IMPORTED)
|
||||
set_target_properties(llama
|
||||
PROPERTIES
|
||||
INTERFACE_INCLUDE_DIRECTORIES "${LLAMA_INCLUDE_DIR}"
|
||||
INTERFACE_LINK_LIBRARIES "ggml::ggml;ggml::ggml-base;"
|
||||
IMPORTED_LINK_INTERFACE_LANGUAGES "CXX"
|
||||
IMPORTED_LOCATION "${llama_LIBRARY}"
|
||||
INTERFACE_COMPILE_FEATURES c_std_90
|
||||
POSITION_INDEPENDENT_CODE ON)
|
||||
|
||||
check_required_components(Llama)
|
|
@ -1,10 +0,0 @@
|
|||
prefix=@CMAKE_INSTALL_PREFIX@
|
||||
exec_prefix=@CMAKE_INSTALL_PREFIX@
|
||||
libdir=@CMAKE_INSTALL_FULL_LIBDIR@
|
||||
includedir=@CMAKE_INSTALL_FULL_INCLUDEDIR@
|
||||
|
||||
Name: llama
|
||||
Description: Port of Facebook's LLaMA model in C/C++
|
||||
Version: @LLAMA_INSTALL_VERSION@
|
||||
Libs: -L${libdir} -lggml -lggml-base -lllama
|
||||
Cflags: -I${includedir}
|
|
@ -1,11 +0,0 @@
|
|||
set( CMAKE_SYSTEM_NAME Windows )
|
||||
set( CMAKE_SYSTEM_PROCESSOR x86_64 )
|
||||
|
||||
set( CMAKE_C_COMPILER clang )
|
||||
set( CMAKE_CXX_COMPILER clang++ )
|
||||
|
||||
set( arch_c_flags "-march=native" )
|
||||
|
||||
set( CMAKE_C_FLAGS_INIT "${arch_c_flags}" )
|
||||
set( CMAKE_CXX_FLAGS_INIT "${arch_c_flags}" )
|
||||
|
14
codecov.yml
Normal file
14
codecov.yml
Normal file
|
@ -0,0 +1,14 @@
|
|||
comment: off
|
||||
|
||||
coverage:
|
||||
status:
|
||||
project:
|
||||
default:
|
||||
target: auto
|
||||
threshold: 0
|
||||
base: auto
|
||||
patch:
|
||||
default:
|
||||
target: auto
|
||||
threshold: 0
|
||||
base: auto
|
|
@ -1,8 +1,5 @@
|
|||
# common
|
||||
|
||||
find_package(Threads REQUIRED)
|
||||
|
||||
llama_add_compile_flags()
|
||||
|
||||
# Build info header
|
||||
#
|
||||
|
@ -22,12 +19,7 @@ if(EXISTS "${CMAKE_CURRENT_SOURCE_DIR}/../.git")
|
|||
endif()
|
||||
endif()
|
||||
|
||||
if(EXISTS "${GIT_DIR}/index")
|
||||
set(GIT_INDEX "${GIT_DIR}/index")
|
||||
else()
|
||||
message(WARNING "Git index not found in git repository.")
|
||||
set(GIT_INDEX "")
|
||||
endif()
|
||||
set(GIT_INDEX "${GIT_DIR}/index")
|
||||
else()
|
||||
message(WARNING "Git repository not found; to enable automatic generation of build info, make sure Git is installed and the project is a Git repository.")
|
||||
set(GIT_INDEX "")
|
||||
|
@ -39,7 +31,7 @@ add_custom_command(
|
|||
COMMENT "Generating build details from Git"
|
||||
COMMAND ${CMAKE_COMMAND} -DMSVC=${MSVC} -DCMAKE_C_COMPILER_VERSION=${CMAKE_C_COMPILER_VERSION}
|
||||
-DCMAKE_C_COMPILER_ID=${CMAKE_C_COMPILER_ID} -DCMAKE_VS_PLATFORM_NAME=${CMAKE_VS_PLATFORM_NAME}
|
||||
-DCMAKE_C_COMPILER=${CMAKE_C_COMPILER} -P "${CMAKE_CURRENT_SOURCE_DIR}/cmake/build-info-gen-cpp.cmake"
|
||||
-DCMAKE_C_COMPILER=${CMAKE_C_COMPILER} -P "${CMAKE_CURRENT_SOURCE_DIR}/../scripts/gen-build-info-cpp.cmake"
|
||||
WORKING_DIRECTORY "${CMAKE_CURRENT_SOURCE_DIR}/.."
|
||||
DEPENDS "${CMAKE_CURRENT_SOURCE_DIR}/build-info.cpp.in" ${GIT_INDEX}
|
||||
VERBATIM
|
||||
|
@ -50,75 +42,27 @@ if (BUILD_SHARED_LIBS)
|
|||
set_target_properties(${TARGET} PROPERTIES POSITION_INDEPENDENT_CODE ON)
|
||||
endif()
|
||||
|
||||
|
||||
set(TARGET common)
|
||||
|
||||
add_library(${TARGET} STATIC
|
||||
arg.cpp
|
||||
arg.h
|
||||
base64.hpp
|
||||
chat.cpp
|
||||
chat.hpp
|
||||
chat-template.hpp
|
||||
common.cpp
|
||||
common.h
|
||||
console.cpp
|
||||
console.h
|
||||
json-schema-to-grammar.cpp
|
||||
json.hpp
|
||||
llguidance.cpp
|
||||
log.cpp
|
||||
log.h
|
||||
minja.hpp
|
||||
ngram-cache.cpp
|
||||
ngram-cache.h
|
||||
sampling.cpp
|
||||
common.cpp
|
||||
sampling.h
|
||||
speculative.cpp
|
||||
speculative.h
|
||||
sampling.cpp
|
||||
console.h
|
||||
console.cpp
|
||||
grammar-parser.h
|
||||
grammar-parser.cpp
|
||||
train.h
|
||||
train.cpp
|
||||
)
|
||||
|
||||
if (BUILD_SHARED_LIBS)
|
||||
set_target_properties(${TARGET} PROPERTIES POSITION_INDEPENDENT_CODE ON)
|
||||
endif()
|
||||
|
||||
set(LLAMA_COMMON_EXTRA_LIBS build_info)
|
||||
|
||||
# Use curl to download model url
|
||||
if (LLAMA_CURL)
|
||||
find_package(CURL REQUIRED)
|
||||
target_compile_definitions(${TARGET} PUBLIC LLAMA_USE_CURL)
|
||||
include_directories(${CURL_INCLUDE_DIRS})
|
||||
find_library(CURL_LIBRARY curl REQUIRED)
|
||||
set(LLAMA_COMMON_EXTRA_LIBS ${LLAMA_COMMON_EXTRA_LIBS} ${CURL_LIBRARY})
|
||||
endif ()
|
||||
|
||||
if (LLAMA_LLGUIDANCE)
|
||||
include(ExternalProject)
|
||||
set(LLGUIDANCE_SRC ${CMAKE_BINARY_DIR}/llguidance/source)
|
||||
set(LLGUIDANCE_PATH ${LLGUIDANCE_SRC}/target/release)
|
||||
ExternalProject_Add(llguidance_ext
|
||||
GIT_REPOSITORY https://github.com/guidance-ai/llguidance
|
||||
# v0.6.12:
|
||||
GIT_TAG ced1c9023d47ec194fa977932d35ce65c2ebfc09
|
||||
PREFIX ${CMAKE_BINARY_DIR}/llguidance
|
||||
SOURCE_DIR ${LLGUIDANCE_SRC}
|
||||
BUILD_IN_SOURCE TRUE
|
||||
CONFIGURE_COMMAND ""
|
||||
BUILD_COMMAND cargo build --release
|
||||
INSTALL_COMMAND ""
|
||||
BUILD_BYPRODUCTS ${LLGUIDANCE_PATH}/libllguidance.a ${LLGUIDANCE_PATH}/llguidance.h
|
||||
UPDATE_COMMAND ""
|
||||
)
|
||||
target_compile_definitions(${TARGET} PUBLIC LLAMA_USE_LLGUIDANCE)
|
||||
|
||||
add_library(llguidance STATIC IMPORTED)
|
||||
set_target_properties(llguidance PROPERTIES IMPORTED_LOCATION ${LLGUIDANCE_PATH}/libllguidance.a)
|
||||
add_dependencies(llguidance llguidance_ext)
|
||||
|
||||
target_include_directories(${TARGET} PRIVATE ${LLGUIDANCE_PATH})
|
||||
set(LLAMA_COMMON_EXTRA_LIBS ${LLAMA_COMMON_EXTRA_LIBS} llguidance)
|
||||
endif ()
|
||||
|
||||
target_include_directories(${TARGET} PUBLIC .)
|
||||
target_compile_features (${TARGET} PUBLIC cxx_std_17)
|
||||
target_link_libraries (${TARGET} PRIVATE ${LLAMA_COMMON_EXTRA_LIBS} PUBLIC llama Threads::Threads)
|
||||
target_compile_features(${TARGET} PUBLIC cxx_std_11)
|
||||
target_link_libraries(${TARGET} PRIVATE llama build_info)
|
||||
|
|
2370
common/arg.cpp
2370
common/arg.cpp
File diff suppressed because it is too large
Load diff
80
common/arg.h
80
common/arg.h
|
@ -1,80 +0,0 @@
|
|||
#pragma once
|
||||
|
||||
#include "common.h"
|
||||
|
||||
#include <set>
|
||||
#include <string>
|
||||
#include <vector>
|
||||
|
||||
//
|
||||
// CLI argument parsing
|
||||
//
|
||||
|
||||
struct common_arg {
|
||||
std::set<enum llama_example> examples = {LLAMA_EXAMPLE_COMMON};
|
||||
std::set<enum llama_example> excludes = {};
|
||||
std::vector<const char *> args;
|
||||
const char * value_hint = nullptr; // help text or example for arg value
|
||||
const char * value_hint_2 = nullptr; // for second arg value
|
||||
const char * env = nullptr;
|
||||
std::string help;
|
||||
bool is_sparam = false; // is current arg a sampling param?
|
||||
void (*handler_void) (common_params & params) = nullptr;
|
||||
void (*handler_string) (common_params & params, const std::string &) = nullptr;
|
||||
void (*handler_str_str)(common_params & params, const std::string &, const std::string &) = nullptr;
|
||||
void (*handler_int) (common_params & params, int) = nullptr;
|
||||
|
||||
common_arg(
|
||||
const std::initializer_list<const char *> & args,
|
||||
const char * value_hint,
|
||||
const std::string & help,
|
||||
void (*handler)(common_params & params, const std::string &)
|
||||
) : args(args), value_hint(value_hint), help(help), handler_string(handler) {}
|
||||
|
||||
common_arg(
|
||||
const std::initializer_list<const char *> & args,
|
||||
const char * value_hint,
|
||||
const std::string & help,
|
||||
void (*handler)(common_params & params, int)
|
||||
) : args(args), value_hint(value_hint), help(help), handler_int(handler) {}
|
||||
|
||||
common_arg(
|
||||
const std::initializer_list<const char *> & args,
|
||||
const std::string & help,
|
||||
void (*handler)(common_params & params)
|
||||
) : args(args), help(help), handler_void(handler) {}
|
||||
|
||||
// support 2 values for arg
|
||||
common_arg(
|
||||
const std::initializer_list<const char *> & args,
|
||||
const char * value_hint,
|
||||
const char * value_hint_2,
|
||||
const std::string & help,
|
||||
void (*handler)(common_params & params, const std::string &, const std::string &)
|
||||
) : args(args), value_hint(value_hint), value_hint_2(value_hint_2), help(help), handler_str_str(handler) {}
|
||||
|
||||
common_arg & set_examples(std::initializer_list<enum llama_example> examples);
|
||||
common_arg & set_excludes(std::initializer_list<enum llama_example> excludes);
|
||||
common_arg & set_env(const char * env);
|
||||
common_arg & set_sparam();
|
||||
bool in_example(enum llama_example ex);
|
||||
bool is_exclude(enum llama_example ex);
|
||||
bool get_value_from_env(std::string & output);
|
||||
bool has_value_from_env();
|
||||
std::string to_string();
|
||||
};
|
||||
|
||||
struct common_params_context {
|
||||
enum llama_example ex = LLAMA_EXAMPLE_COMMON;
|
||||
common_params & params;
|
||||
std::vector<common_arg> options;
|
||||
void(*print_usage)(int, char **) = nullptr;
|
||||
common_params_context(common_params & params) : params(params) {}
|
||||
};
|
||||
|
||||
// parse input arguments from CLI
|
||||
// if one argument has invalid value, it will automatically display usage of the specific argument (and not the full usage message)
|
||||
bool common_params_parse(int argc, char ** argv, common_params & params, llama_example ex, void(*print_usage)(int, char **) = nullptr);
|
||||
|
||||
// function to be used by test-arg-parser
|
||||
common_params_context common_params_parser_init(common_params & params, llama_example ex, void(*print_usage)(int, char **) = nullptr);
|
|
@ -1,529 +0,0 @@
|
|||
/*
|
||||
Copyright 2024 Google LLC
|
||||
|
||||
Use of this source code is governed by an MIT-style
|
||||
license that can be found in the LICENSE file or at
|
||||
https://opensource.org/licenses/MIT.
|
||||
*/
|
||||
// SPDX-License-Identifier: MIT
|
||||
#pragma once
|
||||
|
||||
#include "minja.hpp"
|
||||
#include <json.hpp>
|
||||
#include <string>
|
||||
#include <vector>
|
||||
|
||||
using json = nlohmann::ordered_json;
|
||||
|
||||
namespace minja {
|
||||
|
||||
struct chat_template_caps {
|
||||
bool supports_tools = false;
|
||||
bool supports_tool_calls = false;
|
||||
bool supports_tool_responses = false;
|
||||
bool supports_system_role = false;
|
||||
bool supports_parallel_tool_calls = false;
|
||||
bool supports_tool_call_id = false;
|
||||
// meta-llama/Llama-3.1-8B-Instruct expects arguments to be an object.
|
||||
// Most other templates (and OpenAI's API) expect the arguments object to be stringified.
|
||||
bool requires_object_arguments = false;
|
||||
// CohereForAI/c4ai-command-r-plus simple variant
|
||||
bool requires_non_null_content = false;
|
||||
// MiniMaxAI/MiniMax-Text-01 special
|
||||
bool requires_typed_content = false;
|
||||
};
|
||||
|
||||
struct chat_template_inputs {
|
||||
nlohmann::ordered_json messages;
|
||||
nlohmann::ordered_json tools;
|
||||
bool add_generation_prompt = true;
|
||||
nlohmann::ordered_json extra_context;
|
||||
std::chrono::system_clock::time_point now = std::chrono::system_clock::now();
|
||||
};
|
||||
|
||||
struct chat_template_options {
|
||||
bool apply_polyfills = true;
|
||||
bool use_bos_token = true;
|
||||
bool use_eos_token = true;
|
||||
bool define_strftime_now = true;
|
||||
|
||||
bool polyfill_tools = true;
|
||||
bool polyfill_tool_call_examples = true;
|
||||
bool polyfill_tool_calls = true;
|
||||
bool polyfill_tool_responses = true;
|
||||
bool polyfill_system_role = true;
|
||||
bool polyfill_object_arguments = true;
|
||||
bool polyfill_typed_content = true;
|
||||
};
|
||||
|
||||
class chat_template {
|
||||
|
||||
private:
|
||||
chat_template_caps caps_;
|
||||
std::string source_;
|
||||
std::string bos_token_;
|
||||
std::string eos_token_;
|
||||
std::shared_ptr<minja::TemplateNode> template_root_;
|
||||
std::string tool_call_example_;
|
||||
|
||||
std::string try_raw_render(
|
||||
const nlohmann::ordered_json & messages,
|
||||
const nlohmann::ordered_json & tools,
|
||||
bool add_generation_prompt,
|
||||
const nlohmann::ordered_json & extra_context = nlohmann::ordered_json()) const
|
||||
{
|
||||
try {
|
||||
chat_template_inputs inputs;
|
||||
inputs.messages = messages;
|
||||
inputs.tools = tools;
|
||||
inputs.add_generation_prompt = add_generation_prompt;
|
||||
inputs.extra_context = extra_context;
|
||||
// Use fixed date for tests
|
||||
inputs.now = std::chrono::system_clock::from_time_t(0);
|
||||
|
||||
chat_template_options opts;
|
||||
opts.apply_polyfills = false;
|
||||
|
||||
auto prompt = apply(inputs, opts);
|
||||
// fprintf(stderr, "try_raw_render: %s\n", prompt.c_str());
|
||||
return prompt;
|
||||
} catch (const std::exception & e) {
|
||||
// fprintf(stderr, "try_raw_render error: %s\n", e.what());
|
||||
return "";
|
||||
}
|
||||
}
|
||||
|
||||
public:
|
||||
|
||||
chat_template(const std::string & source, const std::string & bos_token, const std::string & eos_token)
|
||||
: source_(source), bos_token_(bos_token), eos_token_(eos_token)
|
||||
{
|
||||
template_root_ = minja::Parser::parse(source_, {
|
||||
/* .trim_blocks = */ true,
|
||||
/* .lstrip_blocks = */ true,
|
||||
/* .keep_trailing_newline = */ false,
|
||||
});
|
||||
|
||||
auto contains = [](const std::string & haystack, const std::string & needle) {
|
||||
return haystack.find(needle) != std::string::npos;
|
||||
};
|
||||
|
||||
const std::string user_needle = "<User Needle>";
|
||||
const std::string sys_needle = "<System Needle>";
|
||||
const json dummy_str_user_msg = {{"role", "user"}, {"content", user_needle}};
|
||||
const json dummy_typed_user_msg = {{"role", "user"}, {"content", json::array({{{"type", "text"}, {"text", user_needle}}})}};
|
||||
|
||||
caps_.requires_typed_content =
|
||||
!contains(try_raw_render(json::array({dummy_str_user_msg}), {}, false), user_needle)
|
||||
&& contains(try_raw_render(json::array({dummy_typed_user_msg}), {}, false), user_needle);
|
||||
|
||||
const auto dummy_user_msg = caps_.requires_typed_content
|
||||
? dummy_typed_user_msg
|
||||
: dummy_str_user_msg;
|
||||
const json needle_system_msg = {
|
||||
{"role", "system"},
|
||||
{"content", caps_.requires_typed_content ? json::array({{{"type", "text"}, {"text", sys_needle}}}) : json(sys_needle)},
|
||||
};
|
||||
|
||||
caps_.supports_system_role = contains(try_raw_render({needle_system_msg, dummy_user_msg,}, {}, false), sys_needle);
|
||||
|
||||
auto out = try_raw_render(json::array({
|
||||
dummy_user_msg
|
||||
}), json::array({
|
||||
{
|
||||
{"name", "some_tool"},
|
||||
{"type", "function"},
|
||||
{"function", {
|
||||
{"name", "some_tool"},
|
||||
{"description", "Some tool."},
|
||||
{"parameters", {
|
||||
{"type", "object"},
|
||||
{"properties", {
|
||||
{"arg", {
|
||||
{"type", "string"},
|
||||
{"description", "Some argument."},
|
||||
}},
|
||||
}},
|
||||
{"required", json::array({ "arg" })},
|
||||
}},
|
||||
}},
|
||||
},
|
||||
}), false);
|
||||
caps_.supports_tools = contains(out, "some_tool");
|
||||
|
||||
auto make_tool_calls_msg = [&](const json & tool_calls) {
|
||||
return json {
|
||||
{"role", "assistant"},
|
||||
{"content", nullptr},
|
||||
{"tool_calls", tool_calls},
|
||||
};
|
||||
};
|
||||
auto make_tool_call = [](const std::string & tool_name, const json & arguments) {
|
||||
return json {
|
||||
{"id", "call_1___"},
|
||||
{"type", "function"},
|
||||
{"function", {
|
||||
{"arguments", arguments},
|
||||
{"name", tool_name},
|
||||
}},
|
||||
};
|
||||
};
|
||||
const json dummy_args_obj {{"argument_needle", "print('Hello, World!')"}};
|
||||
|
||||
// Note: the arguments are rendered in both cases, but may be double-escaped, which we don't want.
|
||||
out = try_raw_render(json::array({
|
||||
dummy_user_msg,
|
||||
make_tool_calls_msg(json::array({make_tool_call("ipython", dummy_args_obj.dump())})),
|
||||
}), {}, false);
|
||||
auto tool_call_renders_str_arguments = contains(out, "\"argument_needle\":") || contains(out, "'argument_needle':");
|
||||
out = try_raw_render(json::array({
|
||||
dummy_user_msg,
|
||||
make_tool_calls_msg(json::array({make_tool_call("ipython", dummy_args_obj)})),
|
||||
}), {}, false);
|
||||
auto tool_call_renders_obj_arguments = contains(out, "\"argument_needle\":") || contains(out, "'argument_needle':");
|
||||
|
||||
caps_.supports_tool_calls = tool_call_renders_str_arguments || tool_call_renders_obj_arguments;
|
||||
caps_.requires_object_arguments = !tool_call_renders_str_arguments && tool_call_renders_obj_arguments;
|
||||
auto out_empty = try_raw_render(json::array({dummy_user_msg, {{"role", "assistant"}, {"content", ""}}}), {}, false);
|
||||
auto out_null = try_raw_render(json::array({dummy_user_msg, {{"role", "assistant"}, {"content", nullptr}}}), {}, false);
|
||||
caps_.requires_non_null_content = contains(out_empty, user_needle) && !contains(out_null, user_needle);
|
||||
|
||||
if (caps_.supports_tool_calls) {
|
||||
auto dummy_args = caps_.requires_object_arguments ? dummy_args_obj : json(dummy_args_obj.dump());
|
||||
auto tc1 = make_tool_call("test_tool1", dummy_args);
|
||||
auto tc2 = make_tool_call("test_tool2", dummy_args);
|
||||
auto out = try_raw_render(json::array({
|
||||
dummy_user_msg,
|
||||
make_tool_calls_msg(json::array({tc1, tc2})),
|
||||
}), {}, false);
|
||||
caps_.supports_parallel_tool_calls = contains(out, "test_tool1") && contains(out, "test_tool2");
|
||||
|
||||
out = try_raw_render(json::array({
|
||||
dummy_user_msg,
|
||||
make_tool_calls_msg(json::array({tc1})),
|
||||
{
|
||||
{"role", "tool"},
|
||||
{"name", "test_tool1"},
|
||||
{"content", "Some response!"},
|
||||
{"tool_call_id", "call_911_"},
|
||||
}
|
||||
}), {}, false);
|
||||
caps_.supports_tool_responses = contains(out, "Some response!");
|
||||
caps_.supports_tool_call_id = contains(out, "call_911_");
|
||||
}
|
||||
|
||||
try {
|
||||
if (!caps_.supports_tools) {
|
||||
const json user_msg {
|
||||
{"role", "user"},
|
||||
{"content", "Hey"},
|
||||
};
|
||||
const json args {
|
||||
{"arg1", "some_value"},
|
||||
};
|
||||
const json tool_call_msg {
|
||||
{"role", "assistant"},
|
||||
{"content", nullptr},
|
||||
{"tool_calls", json::array({
|
||||
{
|
||||
// TODO: detect if requires numerical id or fixed length == 6 like Nemo
|
||||
{"id", "call_1___"},
|
||||
{"type", "function"},
|
||||
{"function", {
|
||||
{"name", "tool_name"},
|
||||
{"arguments", (caps_.requires_object_arguments ? args : json(minja::Value(args).dump(-1, /* to_json= */ true)))},
|
||||
}},
|
||||
},
|
||||
})},
|
||||
};
|
||||
std::string prefix, full;
|
||||
{
|
||||
chat_template_inputs inputs;
|
||||
inputs.messages = json::array({user_msg});
|
||||
inputs.add_generation_prompt = true;
|
||||
prefix = apply(inputs);
|
||||
}
|
||||
{
|
||||
chat_template_inputs inputs;
|
||||
inputs.messages = json::array({user_msg, tool_call_msg});
|
||||
inputs.add_generation_prompt = false;
|
||||
full = apply(inputs);
|
||||
}
|
||||
auto eos_pos_last = full.rfind(eos_token_);
|
||||
if (eos_pos_last == prefix.size() - eos_token_.size() ||
|
||||
(full[full.size() - 1] == '\n' && (eos_pos_last == full.size() - eos_token_.size() - 1))) {
|
||||
full = full.substr(0, eos_pos_last);
|
||||
}
|
||||
size_t common_prefix_length = 0;
|
||||
for (size_t i = 0; i < prefix.size() && i < full.size(); ++i) {
|
||||
if (prefix[i] != full[i]) {
|
||||
break;
|
||||
}
|
||||
if (prefix[i] == '<') {
|
||||
// DeepSeek R1's template (as of 20250209) adds a trailing <think> if add_generation_prompt,
|
||||
// but it removes thinking tags for past messages.
|
||||
// The prefix and full strings diverge at <think> vs. <|tool▁calls▁begin|>, we avoid consuming the leading <.
|
||||
continue;
|
||||
}
|
||||
common_prefix_length = i + 1;
|
||||
}
|
||||
auto example = full.substr(common_prefix_length);
|
||||
if (example.find("tool_name") == std::string::npos && example.find("some_value") == std::string::npos) {
|
||||
fprintf(stderr, "Failed to infer a tool call example (possible template bug)\n");
|
||||
} else {
|
||||
tool_call_example_ = example;
|
||||
}
|
||||
}
|
||||
} catch (const std::exception & e) {
|
||||
fprintf(stderr, "Failed to generate tool call example: %s\n", e.what());
|
||||
}
|
||||
}
|
||||
|
||||
const std::string & source() const { return source_; }
|
||||
const std::string & bos_token() const { return bos_token_; }
|
||||
const std::string & eos_token() const { return eos_token_; }
|
||||
const chat_template_caps & original_caps() const { return caps_; }
|
||||
|
||||
// Deprecated, please use the form with chat_template_inputs and chat_template_options
|
||||
std::string apply(
|
||||
const nlohmann::ordered_json & messages,
|
||||
const nlohmann::ordered_json & tools,
|
||||
bool add_generation_prompt,
|
||||
const nlohmann::ordered_json & extra_context = nlohmann::ordered_json(),
|
||||
bool apply_polyfills = true)
|
||||
{
|
||||
fprintf(stderr, "[%s] Deprecated!\n", __func__);
|
||||
chat_template_inputs inputs;
|
||||
inputs.messages = messages;
|
||||
inputs.tools = tools;
|
||||
inputs.add_generation_prompt = add_generation_prompt;
|
||||
inputs.extra_context = extra_context;
|
||||
inputs.now = std::chrono::system_clock::now();
|
||||
|
||||
chat_template_options opts;
|
||||
opts.apply_polyfills = apply_polyfills;
|
||||
|
||||
return apply(inputs, opts);
|
||||
}
|
||||
|
||||
std::string apply(
|
||||
const chat_template_inputs & inputs,
|
||||
const chat_template_options & opts = chat_template_options()) const
|
||||
{
|
||||
json actual_messages;
|
||||
|
||||
auto has_tools = inputs.tools.is_array() && !inputs.tools.empty();
|
||||
auto has_tool_calls = false;
|
||||
auto has_tool_responses = false;
|
||||
auto has_string_content = false;
|
||||
for (const auto & message : inputs.messages) {
|
||||
if (message.contains("tool_calls") && !message["tool_calls"].is_null()) {
|
||||
has_tool_calls = true;
|
||||
}
|
||||
if (message.contains("role") && message["role"] == "tool") {
|
||||
has_tool_responses = true;
|
||||
}
|
||||
if (message.contains("content") && message["content"].is_string()) {
|
||||
has_string_content = true;
|
||||
}
|
||||
}
|
||||
|
||||
auto polyfill_system_role = opts.polyfill_system_role && !caps_.supports_system_role;
|
||||
auto polyfill_tools = opts.polyfill_tools && has_tools && !caps_.supports_tools;
|
||||
auto polyfill_tool_call_example = polyfill_tools && opts.polyfill_tool_call_examples;
|
||||
auto polyfill_tool_calls = opts.polyfill_tool_calls && has_tool_calls && !caps_.supports_tool_calls;
|
||||
auto polyfill_tool_responses = opts.polyfill_tool_responses && has_tool_responses && !caps_.supports_tool_responses;
|
||||
auto polyfill_object_arguments = opts.polyfill_object_arguments && has_tool_calls && caps_.requires_object_arguments;
|
||||
auto polyfill_typed_content = opts.polyfill_typed_content && has_string_content && caps_.requires_typed_content;
|
||||
|
||||
auto needs_polyfills = opts.apply_polyfills && (false
|
||||
|| polyfill_system_role
|
||||
|| polyfill_tools
|
||||
|| polyfill_tool_calls
|
||||
|| polyfill_tool_responses
|
||||
|| polyfill_object_arguments
|
||||
|| polyfill_typed_content
|
||||
);
|
||||
|
||||
if (needs_polyfills) {
|
||||
actual_messages = json::array();
|
||||
|
||||
auto add_message = [&](const json & msg) {
|
||||
if (polyfill_typed_content && msg.contains("content") && !msg.at("content").is_null() && msg.at("content").is_string()) {
|
||||
actual_messages.push_back({
|
||||
{"role", msg.at("role")},
|
||||
{"content", {{
|
||||
{"type", "text"},
|
||||
{"text", msg.at("content")},
|
||||
}}},
|
||||
});
|
||||
} else {
|
||||
actual_messages.push_back(msg);
|
||||
}
|
||||
};
|
||||
|
||||
std::string pending_system;
|
||||
auto flush_sys = [&]() {
|
||||
if (!pending_system.empty()) {
|
||||
add_message({
|
||||
{"role", "user"},
|
||||
{"content", pending_system},
|
||||
});
|
||||
pending_system.clear();
|
||||
}
|
||||
};
|
||||
|
||||
json adjusted_messages;
|
||||
if (polyfill_tools) {
|
||||
adjusted_messages = add_system(inputs.messages,
|
||||
"You can call any of the following tools to satisfy the user's requests: " + minja::Value(inputs.tools).dump(2, /* to_json= */ true) +
|
||||
(!polyfill_tool_call_example || tool_call_example_.empty() ? "" : "\n\nExample tool call syntax:\n\n" + tool_call_example_ + "\n\n"));
|
||||
} else {
|
||||
adjusted_messages = inputs.messages;
|
||||
}
|
||||
|
||||
for (const auto & message_ : adjusted_messages) {
|
||||
auto message = message_;
|
||||
if (!message.contains("role") || !message.contains("content")) {
|
||||
throw std::runtime_error("message must have 'role' and 'content' fields: " + message.dump());
|
||||
}
|
||||
std::string role = message.at("role");
|
||||
|
||||
if (message.contains("tool_calls")) {
|
||||
if (polyfill_object_arguments || polyfill_tool_calls) {
|
||||
for (auto & tool_call : message.at("tool_calls")) {
|
||||
if (tool_call["type"] == "function") {
|
||||
auto & function = tool_call.at("function");
|
||||
auto & arguments = function.at("arguments");
|
||||
if (arguments.is_string()) {
|
||||
try {
|
||||
arguments = json::parse(arguments.get<std::string>());
|
||||
} catch (const std::exception & ecvt) {
|
||||
fprintf(stderr, "Failed to parse arguments: %s\n", ecvt.what());
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
if (polyfill_tool_calls) {
|
||||
auto content = message.at("content");
|
||||
auto tool_calls = json::array();
|
||||
for (const auto & tool_call : message.at("tool_calls")) {
|
||||
if (tool_call.at("type") != "function") {
|
||||
continue;
|
||||
}
|
||||
const auto & function = tool_call.at("function");
|
||||
auto tc = json {
|
||||
{"name", function.at("name")},
|
||||
{"arguments", function.at("arguments")},
|
||||
};
|
||||
if (tool_call.contains("id")) {
|
||||
tc["id"] = tool_call["id"];
|
||||
}
|
||||
tool_calls.push_back(tc);
|
||||
}
|
||||
auto obj = json {
|
||||
{"tool_calls", tool_calls},
|
||||
};
|
||||
if (!content.is_null() && content != "") {
|
||||
obj["content"] = content;
|
||||
}
|
||||
message["content"] = obj.dump(2);
|
||||
message.erase("tool_calls");
|
||||
}
|
||||
}
|
||||
if (polyfill_tool_responses && role == "tool") {
|
||||
message["role"] = "user";
|
||||
auto obj = json {
|
||||
{"tool_response", {
|
||||
{"content", message.at("content")},
|
||||
}},
|
||||
};
|
||||
if (message.contains("name")) {
|
||||
obj["tool_response"]["name"] = message.at("name");
|
||||
}
|
||||
if (message.contains("tool_call_id")) {
|
||||
obj["tool_response"]["tool_call_id"] = message.at("tool_call_id");
|
||||
}
|
||||
message["content"] = obj.dump(2);
|
||||
message.erase("name");
|
||||
}
|
||||
|
||||
if (!message["content"].is_null() && polyfill_system_role) {
|
||||
std::string content = message.at("content");
|
||||
if (role == "system") {
|
||||
if (!pending_system.empty()) pending_system += "\n";
|
||||
pending_system += content;
|
||||
continue;
|
||||
} else {
|
||||
if (role == "user") {
|
||||
if (!pending_system.empty()) {
|
||||
message["content"] = pending_system + (content.empty() ? "" : "\n" + content);
|
||||
pending_system.clear();
|
||||
}
|
||||
} else {
|
||||
flush_sys();
|
||||
}
|
||||
}
|
||||
}
|
||||
add_message(message);
|
||||
}
|
||||
flush_sys();
|
||||
} else {
|
||||
actual_messages = inputs.messages;
|
||||
}
|
||||
|
||||
auto context = minja::Context::make(json({
|
||||
{"messages", actual_messages},
|
||||
{"add_generation_prompt", inputs.add_generation_prompt},
|
||||
}));
|
||||
context->set("bos_token", opts.use_bos_token ? bos_token_ : "");
|
||||
context->set("eos_token", opts.use_eos_token ? eos_token_ : "");
|
||||
if (opts.define_strftime_now) {
|
||||
auto now = inputs.now;
|
||||
context->set("strftime_now", Value::callable([now](const std::shared_ptr<minja::Context> &, minja::ArgumentsValue & args) {
|
||||
args.expectArgs("strftime_now", {1, 1}, {0, 0});
|
||||
auto format = args.args[0].get<std::string>();
|
||||
|
||||
auto time = std::chrono::system_clock::to_time_t(now);
|
||||
auto local_time = *std::localtime(&time);
|
||||
std::ostringstream ss;
|
||||
ss << std::put_time(&local_time, format.c_str());
|
||||
return ss.str();
|
||||
}));
|
||||
}
|
||||
if (!inputs.tools.is_null()) {
|
||||
context->set("tools", minja::Value(inputs.tools));
|
||||
}
|
||||
if (!inputs.extra_context.is_null()) {
|
||||
for (auto & kv : inputs.extra_context.items()) {
|
||||
context->set(kv.key(), minja::Value(kv.value()));
|
||||
}
|
||||
}
|
||||
|
||||
auto ret = template_root_->render(context);
|
||||
// fprintf(stderr, "actual_messages: %s\n", actual_messages.dump(2).c_str());
|
||||
// fprintf(stderr, "apply: %s\n\n", ret.c_str());
|
||||
return ret;
|
||||
}
|
||||
|
||||
static nlohmann::ordered_json add_system(const nlohmann::ordered_json & messages, const std::string & system_prompt) {
|
||||
json messages_with_system = messages;
|
||||
|
||||
if (messages_with_system.size() > 0 && messages_with_system[0].at("role") == "system") {
|
||||
std::string existing_system = messages_with_system.at(0).at("content");
|
||||
messages_with_system[0] = json {
|
||||
{"role", "system"},
|
||||
{"content", existing_system + "\n\n" + system_prompt},
|
||||
};
|
||||
} else {
|
||||
messages_with_system.insert(messages_with_system.begin(), json {
|
||||
{"role", "system"},
|
||||
{"content", system_prompt},
|
||||
});
|
||||
}
|
||||
return messages_with_system;
|
||||
}
|
||||
};
|
||||
|
||||
} // namespace minja
|
966
common/chat.cpp
966
common/chat.cpp
|
@ -1,966 +0,0 @@
|
|||
#include "chat.hpp"
|
||||
#include "chat-template.hpp"
|
||||
#include "json-schema-to-grammar.h"
|
||||
#include "log.h"
|
||||
#include "minja.hpp"
|
||||
|
||||
std::string common_chat_format_name(common_chat_format format) {
|
||||
switch (format) {
|
||||
case COMMON_CHAT_FORMAT_CONTENT_ONLY: return "Content-only";
|
||||
case COMMON_CHAT_FORMAT_GENERIC: return "Generic";
|
||||
case COMMON_CHAT_FORMAT_MISTRAL_NEMO: return "Mistral Nemo";
|
||||
case COMMON_CHAT_FORMAT_LLAMA_3_X: return "Llama 3.x";
|
||||
case COMMON_CHAT_FORMAT_LLAMA_3_X_WITH_BUILTIN_TOOLS: return "Llama 3.x with builtin tools";
|
||||
case COMMON_CHAT_FORMAT_DEEPSEEK_R1: return "DeepSeek R1";
|
||||
case COMMON_CHAT_FORMAT_FIREFUNCTION_V2: return "FireFunction v2";
|
||||
case COMMON_CHAT_FORMAT_FUNCTIONARY_V3_2: return "Functionary v3.2";
|
||||
case COMMON_CHAT_FORMAT_FUNCTIONARY_V3_1_LLAMA_3_1: return "Functionary v3.1 Llama 3.1";
|
||||
case COMMON_CHAT_FORMAT_HERMES_2_PRO: return "Hermes 2 Pro";
|
||||
case COMMON_CHAT_FORMAT_COMMAND_R7B: return "Command R7B";
|
||||
default:
|
||||
throw std::runtime_error("Unknown chat format");
|
||||
}
|
||||
}
|
||||
|
||||
const common_grammar_options grammar_options {
|
||||
/* .dotall = */ false,
|
||||
/* .compact_spaces = */ false,
|
||||
// /* .compact_spaces = */ true,
|
||||
};
|
||||
|
||||
static bool parse_json(std::string::const_iterator & it, const std::string::const_iterator & end, json & out) {
|
||||
// // https://json.nlohmann.me/features/parsing/sax_interface/
|
||||
struct json_error_locator : public nlohmann::json_sax<json> {
|
||||
std::size_t position;
|
||||
bool found_error;
|
||||
|
||||
json_error_locator() : position(0), found_error(false) {}
|
||||
|
||||
bool parse_error(std::size_t position, const std::string &, const json::exception &) override {
|
||||
this->position = position - 1;
|
||||
this->found_error = true;
|
||||
return false;
|
||||
}
|
||||
bool null() override { return true; }
|
||||
bool boolean(bool) override { return true; }
|
||||
bool number_integer(number_integer_t) override { return true; }
|
||||
bool number_unsigned(number_unsigned_t) override { return true; }
|
||||
bool number_float(number_float_t, const string_t &) override { return true; }
|
||||
bool string(string_t &) override { return true; }
|
||||
bool binary(binary_t &) override { return true; }
|
||||
bool start_object(std::size_t) override { return true; }
|
||||
bool key(string_t &) override { return true; }
|
||||
bool end_object() override { return true; }
|
||||
bool start_array(std::size_t) override { return true; }
|
||||
bool end_array() override { return true; }
|
||||
};
|
||||
json_error_locator err_loc;
|
||||
json::sax_parse(it, end, &err_loc);
|
||||
|
||||
std::string::const_iterator temptative_end;
|
||||
if (err_loc.found_error) {
|
||||
temptative_end = it + err_loc.position;
|
||||
} else {
|
||||
temptative_end = end;
|
||||
}
|
||||
std::string json_sub {it, temptative_end};
|
||||
try {
|
||||
out = json::parse(json_sub);
|
||||
it = temptative_end;
|
||||
return true;
|
||||
} catch (const std::exception &) {
|
||||
return false;
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
/**
|
||||
* Takes a prefix regex that must have 1 group to capture the function name, a closing suffix, and expects json parameters in between.
|
||||
* Aggregates the prefix, suffix and in-between text into the content.
|
||||
*/
|
||||
static common_chat_msg parse_json_tool_calls(
|
||||
const std::string& input,
|
||||
const std::optional<std::regex> & trigger_opt,
|
||||
const std::regex & function_regex,
|
||||
const std::regex & close_regex) {
|
||||
std::smatch match;
|
||||
|
||||
common_chat_msg result;
|
||||
result.role = "assistant";
|
||||
|
||||
|
||||
auto end = input.end();
|
||||
auto it = input.begin();
|
||||
|
||||
if (trigger_opt) {
|
||||
if (!std::regex_search(it, end, match, *trigger_opt)) {
|
||||
result.content = input;
|
||||
return result;
|
||||
}
|
||||
result.content = match.prefix().str();
|
||||
it = match.suffix().first;
|
||||
}
|
||||
|
||||
while (it != end) {
|
||||
std::sregex_iterator rend;
|
||||
std::sregex_iterator rit(it, end, function_regex);
|
||||
if (rit == rend) {
|
||||
fprintf(stderr, "No more tool calls found\n");
|
||||
result.content += std::string(it, end);
|
||||
break;
|
||||
}
|
||||
auto name = rit->str(1);
|
||||
result.content += std::string(it, rit->prefix().second);
|
||||
it = rit->suffix().first;
|
||||
|
||||
json arguments;
|
||||
if (!parse_json(it, end, arguments)) {
|
||||
throw std::runtime_error("Failed to parse json tool call arguments");
|
||||
}
|
||||
if (!std::regex_search(it, end, match, close_regex)) {
|
||||
throw std::runtime_error("Malformed input, missing closing pattern");
|
||||
}
|
||||
it = match.suffix().first;
|
||||
result.tool_calls.push_back({name, arguments.is_string() ? arguments.get<std::string>() : arguments.dump(), /* id= */ ""});
|
||||
}
|
||||
return result;
|
||||
}
|
||||
|
||||
static common_chat_msg parse_prefixed_json_tool_call_array(const std::string& input, const std::string & prefix, size_t rstrip_prefix = 0) {
|
||||
auto content_end = input.find(prefix);
|
||||
size_t tc_start = std::string::npos;
|
||||
|
||||
common_chat_msg result;
|
||||
result.role = "assistant";
|
||||
const auto process_tool_calls = [&](const json & tool_calls) {
|
||||
for (const auto & tool_call : tool_calls) {
|
||||
const auto & arguments = tool_call["arguments"];
|
||||
result.tool_calls.push_back({
|
||||
tool_call["name"],
|
||||
arguments.is_string() ? arguments.get<std::string>() : arguments.dump(),
|
||||
tool_call.contains("id") ? tool_call["id"] : "",
|
||||
});
|
||||
}
|
||||
};
|
||||
if (content_end == std::string::npos) {
|
||||
result.content = input;
|
||||
} else {
|
||||
tc_start = content_end + prefix.size() - rstrip_prefix;
|
||||
result.content = input.substr(0, content_end);
|
||||
auto tool_calls = json::parse(input.substr(tc_start));
|
||||
process_tool_calls(tool_calls);
|
||||
}
|
||||
return result;
|
||||
}
|
||||
|
||||
static void foreach_function(const json & tools, const std::function<void(const json &)> & fn) {
|
||||
for (const auto & tool : tools) {
|
||||
if (!tool.contains("type") || tool["type"] != "function" || !tool.contains("function")) {
|
||||
LOG_INF("Skipping tool without function: %s", tool.dump(2).c_str());
|
||||
continue;
|
||||
}
|
||||
fn(tool);
|
||||
}
|
||||
}
|
||||
|
||||
static std::string apply(
|
||||
const common_chat_template & tmpl,
|
||||
const nlohmann::ordered_json & messages,
|
||||
const nlohmann::ordered_json & tools,
|
||||
bool add_generation_prompt,
|
||||
const nlohmann::ordered_json & extra_context = nlohmann::ordered_json())
|
||||
{
|
||||
minja::chat_template_inputs tmpl_inputs;
|
||||
tmpl_inputs.messages = messages;
|
||||
tmpl_inputs.tools = tools;
|
||||
tmpl_inputs.add_generation_prompt = add_generation_prompt;
|
||||
tmpl_inputs.extra_context = extra_context;
|
||||
// TODO: add flag to control date/time, if only for testing purposes.
|
||||
// tmpl_inputs.now = std::chrono::system_clock::now();
|
||||
|
||||
minja::chat_template_options tmpl_opts;
|
||||
tmpl_opts.use_bos_token = false;
|
||||
tmpl_opts.use_eos_token = false;
|
||||
|
||||
return tmpl.apply(tmpl_inputs, tmpl_opts);
|
||||
}
|
||||
|
||||
static common_chat_params common_chat_params_init_generic(const common_chat_template & tmpl, const struct common_chat_inputs & inputs) {
|
||||
common_chat_params data;
|
||||
|
||||
auto tool_call_schemas = json::array();
|
||||
foreach_function(inputs.tools, [&](const json & tool) {
|
||||
const auto & function = tool["function"];
|
||||
auto tool_schema = json {
|
||||
{"type", "object"},
|
||||
{"properties", {
|
||||
{"name", {
|
||||
{"type", "string"},
|
||||
{"const", function["name"]},
|
||||
}},
|
||||
{"arguments", function["parameters"]},
|
||||
}},
|
||||
{"required", json::array({"name", "arguments"})},
|
||||
};
|
||||
if (function.contains("description")) {
|
||||
tool_schema["description"] = function["description"];
|
||||
}
|
||||
if (inputs.parallel_tool_calls) {
|
||||
tool_schema["properties"]["id"] = {
|
||||
{"type", "string"},
|
||||
{"minLength", 4},
|
||||
};
|
||||
tool_schema["required"].push_back("id");
|
||||
}
|
||||
tool_call_schemas.emplace_back(tool_schema);
|
||||
});
|
||||
const auto tool_call =
|
||||
inputs.parallel_tool_calls
|
||||
? json {
|
||||
{"type", "object"},
|
||||
{"properties", {
|
||||
{"tool_calls", {
|
||||
{"type", "array"},
|
||||
{"items", tool_call_schemas.size() == 1 ? tool_call_schemas[0] : json {
|
||||
{"anyOf", tool_call_schemas},
|
||||
}},
|
||||
{"minItems", 1},
|
||||
}},
|
||||
}},
|
||||
{"required", json::array({"tool_calls"})},
|
||||
}
|
||||
: json {
|
||||
{"type", "object"},
|
||||
{"properties", {
|
||||
{"tool_call", tool_call_schemas.size() == 1 ? tool_call_schemas[0] : json {
|
||||
{"anyOf", tool_call_schemas},
|
||||
}},
|
||||
}},
|
||||
{"required", json::array({"tool_call"})},
|
||||
};
|
||||
const auto schema =
|
||||
inputs.tool_choice != "required"
|
||||
? json {
|
||||
{"anyOf", json::array({
|
||||
tool_call,
|
||||
{
|
||||
{"type", "object"},
|
||||
{"properties", {
|
||||
{"response", inputs.json_schema.is_null()
|
||||
? json {{"type", "string"}}
|
||||
: inputs.json_schema
|
||||
},
|
||||
}},
|
||||
{"required", json::array({"response"})},
|
||||
},
|
||||
})}
|
||||
}
|
||||
: tool_call;
|
||||
|
||||
data.grammar_lazy = false;
|
||||
data.grammar = build_grammar([&](const common_grammar_builder & builder) {
|
||||
builder.add_schema("root", schema);
|
||||
}, grammar_options);
|
||||
|
||||
auto tweaked_messages = common_chat_template::add_system(
|
||||
inputs.messages,
|
||||
"Respond in JSON format, either with `tool_call` (a request to call tools) or with `response` reply to the user's request");
|
||||
|
||||
data.prompt = apply(tmpl, tweaked_messages, inputs.tools.empty() ? json() : inputs.tools, inputs.add_generation_prompt);
|
||||
data.format = COMMON_CHAT_FORMAT_GENERIC;
|
||||
return data;
|
||||
}
|
||||
static common_chat_msg common_chat_parse_generic(const std::string & input) {
|
||||
json data = json::parse(input);
|
||||
common_chat_msg result;
|
||||
result.role = "assistant";
|
||||
if (data.contains("tool_calls")) {
|
||||
for (const auto & tool_call : data["tool_calls"]) {
|
||||
result.tool_calls.push_back({
|
||||
tool_call["name"],
|
||||
tool_call["arguments"].dump(),
|
||||
tool_call.contains("id") ? tool_call["id"] : "",
|
||||
});
|
||||
}
|
||||
} else if (data.contains("tool_call")) {
|
||||
result.tool_calls.push_back({
|
||||
data["tool_call"]["name"],
|
||||
data["tool_call"]["arguments"].dump(),
|
||||
/* id= */ "",
|
||||
});
|
||||
} else if (data.contains("response")) {
|
||||
const auto & response = data["response"];
|
||||
result.content = response.is_string() ? response.get<std::string>() : response.dump(2);
|
||||
}
|
||||
return result;
|
||||
}
|
||||
|
||||
static common_chat_params common_chat_params_init_mistral_nemo(const common_chat_template & tmpl, const struct common_chat_inputs & inputs) {
|
||||
common_chat_params data;
|
||||
data.grammar_lazy = inputs.tool_choice != "required";
|
||||
data.grammar = build_grammar([&](const common_grammar_builder & builder) {
|
||||
auto schemas = json::array();
|
||||
foreach_function(inputs.tools, [&](const json & tool) {
|
||||
const auto & function = tool["function"];
|
||||
schemas.push_back({
|
||||
{"type", "object"},
|
||||
{"properties", {
|
||||
// Important note: the model is probably trained to take a JSON stringified arguments value.
|
||||
// It's hard to constrain that for now (while reusing the JSON schema conversion), so we're just expecting a plain object.
|
||||
{"name", {
|
||||
{"type", "string"},
|
||||
{"const", function["name"]},
|
||||
}},
|
||||
{"arguments", function["parameters"]},
|
||||
{"id", {
|
||||
{"type", "string"},
|
||||
// Nemo's template expects a 9-character alphanumeric ID.
|
||||
{"pattern", "^[a-zA-Z0-9]{9}$"},
|
||||
}},
|
||||
}},
|
||||
{"required", json::array({"name", "arguments", "id"})},
|
||||
});
|
||||
});
|
||||
auto schema = json {
|
||||
{"type", "array"},
|
||||
{"items", schemas.size() == 1 ? schemas[0] : json {{"anyOf", schemas}}},
|
||||
{"minItems", 1},
|
||||
};
|
||||
if (!inputs.parallel_tool_calls) {
|
||||
schema["maxItems"] = 1;
|
||||
}
|
||||
builder.add_rule("root", "\"[TOOL_CALLS]\" " + builder.add_schema("tool_calls", schema));
|
||||
}, grammar_options);
|
||||
data.grammar_triggers.push_back({"[TOOL_CALLS]", /* .at_start = */ true});
|
||||
data.prompt = apply(tmpl, inputs.messages, inputs.tools.empty() ? json() : inputs.tools, inputs.add_generation_prompt);
|
||||
data.format = COMMON_CHAT_FORMAT_MISTRAL_NEMO;
|
||||
return data;
|
||||
}
|
||||
static common_chat_msg common_chat_parse_mistral_nemo(const std::string & input) {
|
||||
return parse_prefixed_json_tool_call_array(input, "[TOOL_CALLS]");
|
||||
}
|
||||
|
||||
static common_chat_params common_chat_params_init_command_r7b(const common_chat_template & tmpl, const struct common_chat_inputs & inputs) {
|
||||
common_chat_params data;
|
||||
data.grammar_lazy = inputs.tool_choice != "required";
|
||||
data.grammar = build_grammar([&](const common_grammar_builder & builder) {
|
||||
auto schemas = json::array();
|
||||
foreach_function(inputs.tools, [&](const json & tool) {
|
||||
const auto & function = tool["function"];
|
||||
schemas.push_back({
|
||||
{"type", "object"},
|
||||
{"properties", {
|
||||
{"tool_call_id", {
|
||||
{"type", "string"},
|
||||
// Command-R's template expects an integer string.
|
||||
{"pattern", "^[0-9]{1,10}$"},
|
||||
}},
|
||||
{"tool_name", {
|
||||
{"type", "string"},
|
||||
{"const", function["name"]},
|
||||
}},
|
||||
{"parameters", function["parameters"]},
|
||||
}},
|
||||
{"required", json::array({"tool_call_id", "tool_name", "parameters"})},
|
||||
});
|
||||
});
|
||||
auto schema = json {
|
||||
{"type", "array"},
|
||||
{"items", schemas.size() == 1 ? schemas[0] : json {{"anyOf", schemas}}},
|
||||
{"minItems", 1},
|
||||
};
|
||||
if (!inputs.parallel_tool_calls) {
|
||||
schema["maxItems"] = 1;
|
||||
}
|
||||
builder.add_rule("root", "\"<|START_ACTION|>\" " + builder.add_schema("tool_calls", schema) + " \"<|END_ACTION|>\"");
|
||||
}, grammar_options);
|
||||
data.grammar_triggers.push_back({"<|START_ACTION|>", /* .at_start = */ false});
|
||||
data.preserved_tokens = {
|
||||
"<|START_RESPONSE|>",
|
||||
"<|END_RESPONSE|>",
|
||||
"<|START_THINKING|>",
|
||||
"<|END_THINKING|>",
|
||||
"<|END_ACTION|>",
|
||||
};
|
||||
data.prompt = apply(tmpl, inputs.messages, inputs.tools.empty() ? json() : inputs.tools, inputs.add_generation_prompt);
|
||||
data.format = COMMON_CHAT_FORMAT_COMMAND_R7B;
|
||||
return data;
|
||||
}
|
||||
static common_chat_msg common_chat_parse_command_r7b(const std::string & input) {
|
||||
static std::regex response_regex("<\\|START_RESPONSE\\|>([\\s\\S\\n\\r]*?)<\\|END_RESPONSE\\|>");
|
||||
static std::regex thought_action_regex("<\\|START_THINKING\\|>([\\s\\S\\n\\r]*?)<\\|END_THINKING\\|><\\|START_ACTION\\|>([\\s\\S\\n\\r]*?)<\\|END_ACTION\\|>");
|
||||
std::smatch match;
|
||||
|
||||
common_chat_msg result;
|
||||
result.role = "assistant";
|
||||
if (std::regex_match(input, match, response_regex)) {
|
||||
result.content = match[1].str();
|
||||
} else if (std::regex_match(input, match, thought_action_regex)) {
|
||||
result.tool_plan = match[1].str();
|
||||
auto actions_str = match[2].str();
|
||||
auto actions = json::parse(actions_str);
|
||||
for (const auto & action : actions) {
|
||||
result.tool_calls.push_back({
|
||||
/* .name = */ action["tool_name"],
|
||||
/* .arguments = */ action["parameters"].dump(),
|
||||
/* .id = */ action["tool_call_id"],
|
||||
});
|
||||
}
|
||||
} else {
|
||||
LOG_ERR("Failed to parse command_r output");
|
||||
result.content = input;
|
||||
}
|
||||
return result;
|
||||
}
|
||||
|
||||
static void expect_tool_parameters(const std::string & name, const json & parameters, const std::vector<std::string> & expected_properties) {
|
||||
if (!parameters.is_object() || !parameters.contains("type") || parameters["type"] != "object" || !parameters.contains("properties") || !parameters.contains("required")) {
|
||||
throw std::runtime_error("Parameters of tool " + name + " must be an object w/ required properties");
|
||||
}
|
||||
const auto & parameters_properties = parameters.at("properties");
|
||||
const auto & parameters_required = parameters.at("required");
|
||||
for (const auto & prop : expected_properties) {
|
||||
if (!parameters_properties.contains(prop)) {
|
||||
throw std::runtime_error("Parameters of tool " + name + " is missing property: " + prop);
|
||||
}
|
||||
if (std::find(parameters_required.begin(), parameters_required.end(), json(prop)) == parameters_required.end()) {
|
||||
throw std::runtime_error("Parameters of tool " + name + " must have property marked as required: " + prop);
|
||||
}
|
||||
}
|
||||
if (parameters_properties.size() != expected_properties.size()) {
|
||||
throw std::runtime_error("Parameters of tool " + name + " must only have these properties:" + string_join(expected_properties, ", "));
|
||||
}
|
||||
}
|
||||
|
||||
static common_chat_params common_chat_params_init_llama_3_1_tool_calls(const common_chat_template & tmpl, const struct common_chat_inputs & inputs, bool allow_python_tag_builtin_tools) {
|
||||
auto builtin_tools = json::array();
|
||||
common_chat_params data;
|
||||
data.grammar_lazy = inputs.tool_choice != "required";
|
||||
data.grammar = build_grammar([&](const common_grammar_builder & builder) {
|
||||
std::vector<std::string> tool_rules;
|
||||
|
||||
auto handle_builtin_tool = [&](const std::string & name, const json & parameters) {
|
||||
if (name == "wolfram_alpha") {
|
||||
// https://github.com/meta-llama/llama-stack/blob/main/llama_stack/providers/remote/tool_runtime/wolfram_alpha/wolfram_alpha.py
|
||||
expect_tool_parameters(name, parameters, {"query"});
|
||||
} else if (name == "web_search" || name == "brave_search") {
|
||||
// https://github.com/meta-llama/llama-stack/blob/main/llama_stack/providers/remote/tool_runtime/brave_search/brave_search.py
|
||||
expect_tool_parameters(name, parameters, {"query"});
|
||||
} else if (name == "python" || name == "code_interpreter") {
|
||||
// https://github.com/meta-llama/llama-stack/blob/main/llama_stack/providers/inline/tool_runtime/code_interpreter/code_interpreter.py
|
||||
expect_tool_parameters(name, parameters, {"code"});
|
||||
} else {
|
||||
return false;
|
||||
}
|
||||
|
||||
std::vector<std::string> kvs;
|
||||
for (const auto & [key, value] : parameters.at("properties").items()) {
|
||||
kvs.push_back("\"" + key + "=\" " + builder.add_schema(name + "-args-" + key, value));
|
||||
}
|
||||
|
||||
tool_rules.push_back(
|
||||
builder.add_rule(
|
||||
name + "-call",
|
||||
"\"<|python_tag|>" + name + ".call(\" " + string_join(kvs, " \", \" ") + " \")\""));
|
||||
builtin_tools.push_back(name);
|
||||
|
||||
return true;
|
||||
};
|
||||
|
||||
foreach_function(inputs.tools, [&](const json & tool) {
|
||||
const auto & function = tool["function"];
|
||||
std::string name = function["name"];
|
||||
auto parameters = function["parameters"];
|
||||
builder.resolve_refs(parameters);
|
||||
|
||||
// https://github.com/meta-llama/llama-stack/tree/main/llama_stack/providers/remote/tool_runtime
|
||||
if (allow_python_tag_builtin_tools) {
|
||||
handle_builtin_tool(name, parameters);
|
||||
}
|
||||
tool_rules.push_back(
|
||||
builder.add_rule(
|
||||
name + "-call",
|
||||
"\"{\" space "
|
||||
"( \"\\\"type\\\":\" space \"\\\"function\\\",\" space )? "
|
||||
"\"\\\"name\\\": \\\"" + name + "\\\", \\\"parameters\\\": \" " +
|
||||
builder.add_schema(name + "-args", parameters) +
|
||||
" \"}\""));
|
||||
data.grammar_triggers.push_back({"{\"name\": \"" + name + "\"", /* .at_start = */ true});
|
||||
});
|
||||
data.grammar_triggers.push_back({"{\"name\":", /* .at_start = */ true});
|
||||
data.grammar_triggers.push_back({"{\n \"name\":", /* .at_start = */ true});
|
||||
data.grammar_triggers.push_back({"{\n \"name\":", /* .at_start = */ true});
|
||||
data.grammar_triggers.push_back({"{\"type\": \"function\"", /* .at_start = */ true});
|
||||
data.grammar_triggers.push_back({"{\n \"type\": \"function\"", /* .at_start = */ true});
|
||||
data.grammar_triggers.push_back({"{\n \"type\": \"function\"", /* .at_start = */ true});
|
||||
if (!builtin_tools.empty()) {
|
||||
data.grammar_triggers.push_back({"<|python_tag|>", /* .at_start = */ false});
|
||||
}
|
||||
builder.add_rule("root", string_join(tool_rules, " | "));
|
||||
}, grammar_options);
|
||||
data.additional_stops.push_back("<|eom_id|>");
|
||||
data.prompt = apply(tmpl, inputs.messages, inputs.tools.empty() ? json() : inputs.tools, inputs.add_generation_prompt, {
|
||||
{"tools_in_user_message", false},
|
||||
{"builtin_tools", builtin_tools.empty() ? json() : builtin_tools},
|
||||
});
|
||||
data.format = allow_python_tag_builtin_tools && !builtin_tools.empty()
|
||||
? COMMON_CHAT_FORMAT_LLAMA_3_X_WITH_BUILTIN_TOOLS
|
||||
: COMMON_CHAT_FORMAT_LLAMA_3_X;
|
||||
return data;
|
||||
}
|
||||
static common_chat_msg common_chat_parse_llama_3_1(const std::string & input, bool with_builtin_tools = false) {
|
||||
// TODO: tighten & simplify the parser, don't accept leading text context.
|
||||
static std::regex function_regex("\\{[\\s\\n\\r]*(?:\"type\"[\\s\\n\\r]*:[\\s\\n\\r]*\"function\"[\\s\\n\\r]*,[\\s\\n\\r]*|[\\s\\n\\r]*)\"name\"[\\s\\n\\r]*:[\\s\\n\\r]*\"([^\"]+)\"[\\s\\n\\r]*,[\\s\\n\\r]*\"parameters\": ");
|
||||
static std::regex close_regex("\\}");
|
||||
static std::regex builtin_call_regex("<\\|python_tag\\|>([^.(]+)\\.call\\((.*)\\)");
|
||||
|
||||
if (with_builtin_tools) {
|
||||
std::smatch match;
|
||||
if (std::regex_match(input, match, builtin_call_regex)) {
|
||||
auto name = match[1].str();
|
||||
auto raw_args = match[2].str();
|
||||
|
||||
// TODO: if/when builtin tools start accepting more than 1 argument, use parse_json for real parsing.
|
||||
auto it_eq = raw_args.find('=');
|
||||
auto arg_name = raw_args.substr(0, it_eq);
|
||||
auto arg_value_str = raw_args.substr(it_eq + 1);
|
||||
auto arg_value = json::parse(arg_value_str);
|
||||
|
||||
return {
|
||||
/* .role = */ "assistant",
|
||||
/* .content = */ match.prefix().str(),
|
||||
/* .tool_calls = */ {
|
||||
{
|
||||
/* .name = */ match[1],
|
||||
/* .arguments = */ (json {
|
||||
{arg_name, arg_value},
|
||||
}).dump(),
|
||||
/* .id = */ "",
|
||||
},
|
||||
},
|
||||
};
|
||||
}
|
||||
}
|
||||
return parse_json_tool_calls(input, std::nullopt, function_regex, close_regex);
|
||||
}
|
||||
|
||||
static common_chat_params common_chat_params_init_deepseek_r1(const common_chat_template & tmpl, const struct common_chat_inputs & inputs) {
|
||||
common_chat_params data;
|
||||
data.grammar_lazy = inputs.tool_choice != "required";
|
||||
data.grammar = build_grammar([&](const common_grammar_builder & builder) {
|
||||
std::vector<std::string> tool_rules;
|
||||
foreach_function(inputs.tools, [&](const json & tool) {
|
||||
const auto & function = tool["function"];
|
||||
std::string name = function["name"];
|
||||
auto parameters = function["parameters"];
|
||||
auto args_rule = builder.add_schema(name + "-args", parameters);
|
||||
tool_rules.push_back(builder.add_rule(name + "-call",
|
||||
"\"<|tool▁call▁begin|>function<|tool▁sep|>" + name + "\\n```json\\n\" " + args_rule + " \"```<|tool▁call▁end|>\""));
|
||||
});
|
||||
data.grammar_triggers.push_back({"<|tool▁calls▁begin|>", /* .at_start = */ false});
|
||||
data.preserved_tokens = {
|
||||
"<|tool▁sep|>",
|
||||
"<|tool▁call▁end|>",
|
||||
};
|
||||
builder.add_rule("root", "\"<|tool▁calls▁begin|>\" (" + string_join(tool_rules, " | ") + ")" + (inputs.parallel_tool_calls ? "*" : "") + " space");
|
||||
}, grammar_options);
|
||||
auto prompt = apply(tmpl, inputs.messages, inputs.tools.empty() ? json() : inputs.tools, inputs.add_generation_prompt);
|
||||
data.prompt = prompt;
|
||||
data.format = COMMON_CHAT_FORMAT_DEEPSEEK_R1;
|
||||
return data;
|
||||
}
|
||||
static common_chat_msg common_chat_parse_deepseek_r1(const std::string & input) {
|
||||
static std::regex trigger_regex("<|tool▁calls▁begin|>");
|
||||
static std::regex function_regex("<|tool▁call▁begin|>function<|tool▁sep|>([^\n]+)\n```json\n");
|
||||
static std::regex close_regex("```<|tool▁call▁end|>");
|
||||
return parse_json_tool_calls(input, trigger_regex, function_regex, close_regex);
|
||||
}
|
||||
|
||||
static common_chat_params common_chat_params_init_firefunction_v2(const common_chat_template & tmpl, const struct common_chat_inputs & inputs) {
|
||||
fprintf(stderr, "%s\n", __func__);
|
||||
common_chat_params data;
|
||||
data.prompt = apply(tmpl, inputs.messages, /* tools= */ nullptr, inputs.add_generation_prompt, {
|
||||
{"datetime", "Jan 29 2025 13:00:00 GMT"},
|
||||
{"functions", json(inputs.tools.empty() ? "" : inputs.tools.dump(2))},
|
||||
});
|
||||
if (!inputs.tools.is_null() && !inputs.tools.empty()) {
|
||||
data.grammar_lazy = inputs.tool_choice != "required";
|
||||
data.grammar = build_grammar([&](const common_grammar_builder & builder) {
|
||||
auto schemas = json::array();
|
||||
foreach_function(inputs.tools, [&](const json & tool) {
|
||||
const auto & function = tool["function"];
|
||||
schemas.push_back({
|
||||
{"type", "object"},
|
||||
{"properties", {
|
||||
{"name", {
|
||||
{"type", "string"},
|
||||
{"const", function["name"]},
|
||||
}},
|
||||
{"arguments", function["parameters"]},
|
||||
}},
|
||||
{"required", json::array({"name", "arguments", "id"})},
|
||||
});
|
||||
});
|
||||
auto schema = json {
|
||||
{"type", "array"},
|
||||
{"items", schemas.size() == 1 ? schemas[0] : json {{"anyOf", schemas}}},
|
||||
{"minItems", 1},
|
||||
};
|
||||
if (!inputs.parallel_tool_calls) {
|
||||
schema["maxItems"] = 1;
|
||||
}
|
||||
builder.add_rule("root", "\" functools\"? " + builder.add_schema("tool_calls", schema));
|
||||
}, grammar_options);
|
||||
data.grammar_triggers.push_back({" functools[", /* .at_start = */ false});
|
||||
data.format = COMMON_CHAT_FORMAT_FIREFUNCTION_V2;
|
||||
} else {
|
||||
data.format = COMMON_CHAT_FORMAT_CONTENT_ONLY;
|
||||
}
|
||||
return data;
|
||||
}
|
||||
static common_chat_msg common_chat_parse_firefunction_v2(const std::string & input) {
|
||||
return parse_prefixed_json_tool_call_array(input, " functools[", /* rstrip_prefix= */ 1);
|
||||
}
|
||||
|
||||
static common_chat_params common_chat_params_init_functionary_v3_2(const common_chat_template & tmpl, const struct common_chat_inputs & inputs) {
|
||||
// >>>all\nlet's call functions>>>fn1\n{"arg1": 1...}\n>>>fn2\n{"arg1": 1...}...
|
||||
// Using ">>>f1\n", ">>>f2\n"... as trigger words for the grammar
|
||||
common_chat_params data;
|
||||
data.prompt = apply(tmpl, inputs.messages, inputs.tools.empty() ? json() : inputs.tools, inputs.add_generation_prompt);
|
||||
data.format = COMMON_CHAT_FORMAT_FUNCTIONARY_V3_2;
|
||||
if (!inputs.tools.is_null() && !inputs.tools.empty()) {
|
||||
data.grammar_lazy = inputs.tool_choice != "required";
|
||||
data.grammar = build_grammar([&](const common_grammar_builder & builder) {
|
||||
std::vector<std::string> first_tool_rules;
|
||||
std::vector<std::string> subsequent_tool_rules;
|
||||
foreach_function(inputs.tools, [&](const json & tool) {
|
||||
const auto & function = tool["function"];
|
||||
std::string name = function["name"];
|
||||
auto parameters = function["parameters"];
|
||||
auto args_rule = builder.add_schema(name + "-args", parameters);
|
||||
first_tool_rules.push_back(builder.add_rule(name + "-call", "\"" + name + "\\n\" " + args_rule));
|
||||
subsequent_tool_rules.push_back(builder.add_rule(name + "-call2", "\">>>" + name + "\\n\" " + args_rule));
|
||||
data.grammar_triggers.push_back({name, /* .at_start = */ true});
|
||||
data.grammar_triggers.push_back({">>>" + name, /* .at_start = */ false});
|
||||
});
|
||||
auto first_rule = first_tool_rules.empty() ? "" : builder.add_rule("first_tool_call", string_join(first_tool_rules, " | ")) + " space";
|
||||
if (inputs.parallel_tool_calls) {
|
||||
auto subsequent_rule = builder.add_rule("subsequent_tool_call", string_join(subsequent_tool_rules, " | ")) + " space";
|
||||
builder.add_rule("root", first_rule + " (" + subsequent_rule + ")*");
|
||||
} else {
|
||||
builder.add_rule("root", first_rule);
|
||||
}
|
||||
|
||||
}, grammar_options);
|
||||
}
|
||||
return data;
|
||||
}
|
||||
|
||||
static bool consume(std::string::const_iterator & it, const std::string::const_iterator & end, const std::string & expected) {
|
||||
auto expected_it = expected.begin();
|
||||
auto tmp_it = it;
|
||||
while (tmp_it != end && expected_it != expected.end() && *tmp_it == *expected_it) {
|
||||
++tmp_it;
|
||||
++expected_it;
|
||||
}
|
||||
if (expected_it == expected.end()) {
|
||||
it = tmp_it;
|
||||
return true;
|
||||
}
|
||||
return false;
|
||||
}
|
||||
|
||||
static common_chat_msg common_chat_parse_functionary_v3_2(const std::string & input) {
|
||||
static std::regex function_regex(R"((?:>>>)?(\w+)\n)");
|
||||
static std::regex close_regex(R"($|(?=>>>))");
|
||||
|
||||
std::string content;
|
||||
auto it = input.begin();
|
||||
const auto end = input.end();
|
||||
|
||||
if (consume(it, end, "all\n")) {
|
||||
std::smatch match;
|
||||
if (std::regex_search(it, end, match, function_regex)) {
|
||||
auto fun_it = match.prefix().second;
|
||||
content = std::string(it, fun_it);
|
||||
it = fun_it;
|
||||
} else {
|
||||
common_chat_msg res;
|
||||
res.role = "assistant";
|
||||
res.content = std::string(it, end);
|
||||
return res;
|
||||
}
|
||||
}
|
||||
// TODO: tighten & simplify.
|
||||
try {
|
||||
auto res = parse_json_tool_calls(std::string(it, end), std::nullopt, function_regex, close_regex);
|
||||
res.content = content + res.content;
|
||||
return res;
|
||||
} catch (const std::exception & e) {
|
||||
LOG_ERR("Failed to parse functionary v3.2 input: %s\n", e.what());
|
||||
common_chat_msg res;
|
||||
res.role = "assistant";
|
||||
res.content = input;
|
||||
return res;
|
||||
}
|
||||
}
|
||||
|
||||
static common_chat_params common_chat_params_init_functionary_v3_1_llama_3_1(const common_chat_template & tmpl, const struct common_chat_inputs & inputs) {
|
||||
// https://github.com/MeetKai/functionary/blob/main/tests/prompt_test_v3-llama3.1.txt
|
||||
common_chat_params data;
|
||||
json tools = inputs.tools.is_null() ? inputs.tools : json::array();
|
||||
std::string python_code_argument_name;
|
||||
auto has_raw_python = false;
|
||||
|
||||
data.grammar_lazy = inputs.tool_choice != "required";
|
||||
data.grammar = build_grammar([&](const common_grammar_builder & builder) {
|
||||
std::vector<std::string> tool_rules;
|
||||
foreach_function(inputs.tools, [&](const json & tool) {
|
||||
const auto & function = tool["function"];
|
||||
const auto & parameters = function["parameters"];
|
||||
std::string name = function["name"];
|
||||
if (name == "python" || name == "ipython") {
|
||||
if (!parameters.contains("type")) {
|
||||
throw std::runtime_error("Missing type in python tool");
|
||||
}
|
||||
has_raw_python = true;
|
||||
auto type = parameters.at("type");
|
||||
if (type == "object") {
|
||||
auto properties = parameters.at("properties");
|
||||
for (auto it = properties.begin(); it != properties.end(); ++it) {
|
||||
if (it.value().at("type") == "string") {
|
||||
if (!python_code_argument_name.empty()) {
|
||||
throw std::runtime_error("Multiple string arguments found in python tool");
|
||||
}
|
||||
python_code_argument_name = it.key();
|
||||
}
|
||||
}
|
||||
if (python_code_argument_name.empty()) {
|
||||
throw std::runtime_error("No string argument found in python tool");
|
||||
}
|
||||
} else if (type != "string") {
|
||||
throw std::runtime_error("Invalid type in python tool: " + type.dump());
|
||||
}
|
||||
}
|
||||
tool_rules.push_back(builder.add_rule(name + "-call", "\"<function=" + name + ">\" " + builder.add_schema(name + "-args", parameters) + " \"</function>\" space"));
|
||||
});
|
||||
if (has_raw_python) {
|
||||
tool_rules.push_back(builder.add_rule("python-call", "\"<|python_tag|>\" .*"));
|
||||
data.grammar_triggers.push_back({"<|python_tag|>", /* .at_start = */ false});
|
||||
}
|
||||
auto tool_call = builder.add_rule("tool_call", string_join(tool_rules, " | ")) + " space";
|
||||
builder.add_rule("root", inputs.parallel_tool_calls ? "(" + tool_call + ")+" : tool_call);
|
||||
data.grammar_triggers.push_back({"<function=", /* .at_start = */ false});
|
||||
}, grammar_options);
|
||||
|
||||
data.prompt = apply(tmpl, inputs.messages, inputs.tools.empty() ? json() : inputs.tools, inputs.add_generation_prompt);
|
||||
// TODO: if (has_raw_python)
|
||||
data.format = COMMON_CHAT_FORMAT_FUNCTIONARY_V3_1_LLAMA_3_1;
|
||||
return data;
|
||||
}
|
||||
static common_chat_msg common_chat_parse_functionary_v3_1_llama_3_1(const std::string & input) {
|
||||
// This version of Functionary still supports the llama 3.1 tool call format for the python tool.
|
||||
static std::regex python_tag_regex(R"(<\|python_tag\|>([\s\S\n]*)$)");
|
||||
std::smatch match;
|
||||
if (std::regex_search(input, match, python_tag_regex)) {
|
||||
auto code = match[1].str();
|
||||
return {
|
||||
/* .role = */ "assistant",
|
||||
/* .content = */ match.prefix().str(),
|
||||
/* .tool_calls = */ {
|
||||
{
|
||||
/* .name = */ "python",
|
||||
/* .arguments = */ (json {{"code", code}}).dump(),
|
||||
/* .id = */ "",
|
||||
},
|
||||
}
|
||||
};
|
||||
}
|
||||
static std::regex function_regex(R"(<function=(\w+)>)");
|
||||
static std::regex close_regex(R"(</function>)");
|
||||
// TODO: tighten & simplify.
|
||||
return parse_json_tool_calls(input, std::nullopt, function_regex, close_regex);
|
||||
}
|
||||
|
||||
static common_chat_params common_chat_params_init_hermes_2_pro(const common_chat_template & tmpl, const struct common_chat_inputs & inputs) {
|
||||
common_chat_params data;
|
||||
// (content)?(<tool_call>{"name": "foo", "arguments": {"a": 1}}</tool_call>)*
|
||||
data.grammar_lazy = inputs.tool_choice != "required";
|
||||
data.grammar = build_grammar([&](const common_grammar_builder & builder) {
|
||||
std::vector<std::string> tool_rules;
|
||||
foreach_function(inputs.tools, [&](const json & tool) {
|
||||
const auto & function = tool["function"];
|
||||
std::string name = function["name"];
|
||||
auto parameters = function["parameters"];
|
||||
builder.resolve_refs(parameters);
|
||||
tool_rules.push_back(builder.add_schema(name + "-call", {
|
||||
{"type", "object"},
|
||||
{"properties", json {
|
||||
{"name", json {{"const", name}}},
|
||||
{"arguments", parameters},
|
||||
}},
|
||||
{"required", json::array({"name", "arguments"})},
|
||||
}));
|
||||
});
|
||||
auto tool_call = "\"<tool_call>\" space " + builder.add_rule("tool_call", string_join(tool_rules, " | ")) + " \"</tool_call>\" space";
|
||||
builder.add_rule("root", inputs.parallel_tool_calls ? "(" + tool_call + ")+" : tool_call);
|
||||
data.grammar_triggers.push_back({"<tool_call>", /* .at_start = */ false});
|
||||
data.preserved_tokens = { "</tool_call>" };
|
||||
}, grammar_options);
|
||||
|
||||
data.prompt = apply(tmpl, inputs.messages, inputs.tools.empty() ? json() : inputs.tools, inputs.add_generation_prompt);
|
||||
data.format = COMMON_CHAT_FORMAT_HERMES_2_PRO;
|
||||
return data;
|
||||
}
|
||||
static common_chat_msg common_chat_parse_hermes_2_pro(const std::string & input) {
|
||||
try {
|
||||
std::regex start_pattern(R"([\n\s]*<tool_call>)");
|
||||
std::regex middle_pattern(R"([\n\s]*</tool_call>[\n\s]*<tool_call>)");
|
||||
std::regex end_pattern(R"([\n\s]*</tool_call>[\n\s]*$)");
|
||||
|
||||
auto end = input.end();
|
||||
std::sregex_iterator rend;
|
||||
std::sregex_iterator rit(input.begin(), end, start_pattern);
|
||||
if (rit == rend) {
|
||||
return {
|
||||
/* .role = */ "assistant",
|
||||
/* .content = */ input,
|
||||
/* .tool_calls = */ {},
|
||||
};
|
||||
}
|
||||
|
||||
common_chat_msg result;
|
||||
result.role = "assistant";
|
||||
result.content = rit->prefix();
|
||||
|
||||
auto it = rit->suffix().first;
|
||||
while (it != end) {
|
||||
json call;
|
||||
if (!parse_json(it, end, call)) {
|
||||
throw std::runtime_error("Failed to parse json tool call");
|
||||
}
|
||||
const auto & arguments = call["arguments"];
|
||||
result.tool_calls.push_back({
|
||||
call["name"],
|
||||
arguments.dump(),
|
||||
// arguments.is_string() ? arguments.get<std::string>() : arguments.dump(),
|
||||
/* id= */ "",
|
||||
});
|
||||
rit = {it, end, middle_pattern};
|
||||
if (rit != rend) {
|
||||
it = rit->suffix().first;
|
||||
} else {
|
||||
rit = {it, end, end_pattern};
|
||||
if (rit == rend) {
|
||||
throw std::runtime_error("Malformed input, missing </tool_call>");
|
||||
}
|
||||
break;
|
||||
}
|
||||
}
|
||||
return result;
|
||||
} catch (const std::exception & e) {
|
||||
return {
|
||||
/* .role = */ "assistant",
|
||||
/* .content = */ input,
|
||||
/* .tool_calls = */ {},
|
||||
};
|
||||
}
|
||||
}
|
||||
|
||||
static common_chat_params common_chat_params_init_without_tools(const common_chat_template & tmpl, const struct common_chat_inputs & inputs) {
|
||||
common_chat_params data;
|
||||
data.prompt = apply(tmpl, inputs.messages, inputs.tools.empty() ? json() : inputs.tools, inputs.add_generation_prompt);
|
||||
data.format = COMMON_CHAT_FORMAT_CONTENT_ONLY;
|
||||
data.grammar_lazy = false;
|
||||
if (!inputs.json_schema.is_null()) {
|
||||
if (!inputs.grammar.empty()) {
|
||||
throw std::runtime_error("Either \"json_schema\" or \"grammar\" can be specified, but not both");
|
||||
}
|
||||
data.grammar = json_schema_to_grammar(inputs.json_schema);
|
||||
} else {
|
||||
data.grammar = inputs.grammar.empty();
|
||||
}
|
||||
return data;
|
||||
}
|
||||
|
||||
common_chat_params common_chat_params_init(const common_chat_template & tmpl, const struct common_chat_inputs & inputs) {
|
||||
auto has_tools = !inputs.tools.is_null() && inputs.tool_choice != "none";
|
||||
LOG_DBG("[%s] has_tools=%s\n", __func__, has_tools ? "true" : "false");
|
||||
|
||||
if (has_tools && !inputs.grammar.empty()) {
|
||||
throw std::runtime_error("Cannot specify grammar with tools");
|
||||
}
|
||||
|
||||
const auto & src = tmpl.source();
|
||||
if (src.find(">>>all") != std::string::npos) {
|
||||
// Functionary prepends "all\n" to plain content outputs, so we use the parser no matter when
|
||||
return common_chat_params_init_functionary_v3_2(tmpl, inputs);
|
||||
}
|
||||
if (src.find(" functools[") != std::string::npos) {
|
||||
// Firefunction v2 requires datetime and functions in the context, even w/o tools.
|
||||
return common_chat_params_init_firefunction_v2(tmpl, inputs);
|
||||
}
|
||||
|
||||
if (!has_tools) {
|
||||
return common_chat_params_init_without_tools(tmpl, inputs);
|
||||
}
|
||||
|
||||
if (src.find("<tool_call>") != std::string::npos) {
|
||||
return common_chat_params_init_hermes_2_pro(tmpl, inputs);
|
||||
}
|
||||
if (src.find("<|start_header_id|>") != std::string::npos
|
||||
&& src.find("<function=") != std::string::npos) {
|
||||
return common_chat_params_init_functionary_v3_1_llama_3_1(tmpl, inputs);
|
||||
}
|
||||
if (src.find("<|start_header_id|>ipython<|end_header_id|>") != std::string::npos) {
|
||||
auto allow_python_tag_builtin_tools = src.find("<|python_tag|>") != std::string::npos;
|
||||
return common_chat_params_init_llama_3_1_tool_calls(tmpl, inputs, allow_python_tag_builtin_tools);
|
||||
}
|
||||
if (src.find("<|tool▁calls▁begin|>") != std::string::npos) {
|
||||
return common_chat_params_init_deepseek_r1(tmpl, inputs);
|
||||
}
|
||||
if (src.find("[TOOL_CALLS]") != std::string::npos) {
|
||||
return common_chat_params_init_mistral_nemo(tmpl, inputs);
|
||||
}
|
||||
if (src.find("<|END_THINKING|><|START_ACTION|>") != std::string::npos) {
|
||||
return common_chat_params_init_command_r7b(tmpl, inputs);
|
||||
}
|
||||
return common_chat_params_init_generic(tmpl, inputs);
|
||||
}
|
||||
|
||||
static common_chat_msg common_chat_parse_content_only(const std::string & input) {
|
||||
return {
|
||||
/* .role = */ "assistant",
|
||||
/* .content = */ input,
|
||||
/* .tool_calls = */ {},
|
||||
};
|
||||
}
|
||||
|
||||
common_chat_msg common_chat_parse(const std::string & input, common_chat_format format) {
|
||||
switch (format) {
|
||||
case COMMON_CHAT_FORMAT_CONTENT_ONLY:
|
||||
return common_chat_parse_content_only(input);
|
||||
case COMMON_CHAT_FORMAT_GENERIC:
|
||||
return common_chat_parse_generic(input);
|
||||
case COMMON_CHAT_FORMAT_MISTRAL_NEMO:
|
||||
return common_chat_parse_mistral_nemo(input);
|
||||
case COMMON_CHAT_FORMAT_LLAMA_3_X:
|
||||
return common_chat_parse_llama_3_1(input);
|
||||
case COMMON_CHAT_FORMAT_LLAMA_3_X_WITH_BUILTIN_TOOLS:
|
||||
return common_chat_parse_llama_3_1(input, /* with_builtin_tools= */ true);
|
||||
case COMMON_CHAT_FORMAT_DEEPSEEK_R1:
|
||||
return common_chat_parse_deepseek_r1(input);
|
||||
case COMMON_CHAT_FORMAT_FUNCTIONARY_V3_2:
|
||||
return common_chat_parse_functionary_v3_2(input);
|
||||
case COMMON_CHAT_FORMAT_FUNCTIONARY_V3_1_LLAMA_3_1:
|
||||
return common_chat_parse_functionary_v3_1_llama_3_1(input);
|
||||
case COMMON_CHAT_FORMAT_HERMES_2_PRO:
|
||||
return common_chat_parse_hermes_2_pro(input);
|
||||
case COMMON_CHAT_FORMAT_FIREFUNCTION_V2:
|
||||
return common_chat_parse_firefunction_v2(input);
|
||||
case COMMON_CHAT_FORMAT_COMMAND_R7B:
|
||||
return common_chat_parse_command_r7b(input);
|
||||
default:
|
||||
throw std::runtime_error("Unsupported format: " + common_chat_format_name(format));
|
||||
}
|
||||
}
|
|
@ -1,52 +0,0 @@
|
|||
// Chat support (incl. tool call grammar constraining & output parsing) w/ generic & custom template handlers.
|
||||
|
||||
#pragma once
|
||||
|
||||
#include "common.h"
|
||||
#include <json.hpp>
|
||||
#include <optional>
|
||||
#include <string>
|
||||
#include <vector>
|
||||
|
||||
using json = nlohmann::ordered_json;
|
||||
|
||||
struct common_chat_inputs {
|
||||
json messages;
|
||||
json tools;
|
||||
json tool_choice;
|
||||
json json_schema;
|
||||
bool parallel_tool_calls;
|
||||
bool stream;
|
||||
std::string grammar;
|
||||
bool add_generation_prompt = true;
|
||||
};
|
||||
|
||||
enum common_chat_format {
|
||||
COMMON_CHAT_FORMAT_CONTENT_ONLY,
|
||||
COMMON_CHAT_FORMAT_GENERIC,
|
||||
COMMON_CHAT_FORMAT_MISTRAL_NEMO,
|
||||
COMMON_CHAT_FORMAT_LLAMA_3_X,
|
||||
COMMON_CHAT_FORMAT_LLAMA_3_X_WITH_BUILTIN_TOOLS,
|
||||
COMMON_CHAT_FORMAT_DEEPSEEK_R1,
|
||||
COMMON_CHAT_FORMAT_FIREFUNCTION_V2,
|
||||
COMMON_CHAT_FORMAT_FUNCTIONARY_V3_2,
|
||||
COMMON_CHAT_FORMAT_FUNCTIONARY_V3_1_LLAMA_3_1,
|
||||
COMMON_CHAT_FORMAT_HERMES_2_PRO,
|
||||
COMMON_CHAT_FORMAT_COMMAND_R7B,
|
||||
|
||||
COMMON_CHAT_FORMAT_COUNT, // Not a format, just the # formats
|
||||
};
|
||||
|
||||
struct common_chat_params {
|
||||
common_chat_format format = COMMON_CHAT_FORMAT_CONTENT_ONLY;
|
||||
json prompt;
|
||||
std::string grammar;
|
||||
bool grammar_lazy = false;
|
||||
std::vector<common_grammar_trigger> grammar_triggers;
|
||||
std::vector<std::string> preserved_tokens;
|
||||
std::vector<std::string> additional_stops;
|
||||
};
|
||||
|
||||
struct common_chat_params common_chat_params_init(const common_chat_template & tmpl, const struct common_chat_inputs & params);
|
||||
std::string common_chat_format_name(common_chat_format format);
|
||||
common_chat_msg common_chat_parse( const std::string & input, common_chat_format format);
|
3228
common/common.cpp
3228
common/common.cpp
File diff suppressed because it is too large
Load diff
743
common/common.h
743
common/common.h
|
@ -2,12 +2,20 @@
|
|||
|
||||
#pragma once
|
||||
|
||||
#include "llama-cpp.h"
|
||||
#include "llama.h"
|
||||
|
||||
#include <set>
|
||||
#include "sampling.h"
|
||||
|
||||
#define LOG_NO_FILE_LINE_FUNCTION
|
||||
#include "log.h"
|
||||
|
||||
#include <cmath>
|
||||
#include <string>
|
||||
#include <vector>
|
||||
#include <sstream>
|
||||
#include <random>
|
||||
#include <thread>
|
||||
#include <unordered_map>
|
||||
#include <tuple>
|
||||
|
||||
#ifdef _WIN32
|
||||
#define DIRECTORY_SEPARATOR '\\'
|
||||
|
@ -19,699 +27,216 @@
|
|||
#define die_fmt(fmt, ...) do { fprintf(stderr, "error: " fmt "\n", __VA_ARGS__); exit(1); } while (0)
|
||||
|
||||
#define print_build_info() do { \
|
||||
fprintf(stderr, "%s: build = %d (%s)\n", __func__, LLAMA_BUILD_NUMBER, LLAMA_COMMIT); \
|
||||
fprintf(stderr, "%s: build = %d (%s)\n", __func__, LLAMA_BUILD_NUMBER, LLAMA_COMMIT); \
|
||||
fprintf(stderr, "%s: built with %s for %s\n", __func__, LLAMA_COMPILER, LLAMA_BUILD_TARGET); \
|
||||
} while(0)
|
||||
|
||||
#define DEFAULT_MODEL_PATH "models/7B/ggml-model-f16.gguf"
|
||||
|
||||
struct common_adapter_lora_info {
|
||||
std::string path;
|
||||
float scale;
|
||||
|
||||
struct llama_adapter_lora * ptr;
|
||||
};
|
||||
|
||||
using llama_tokens = std::vector<llama_token>;
|
||||
|
||||
// build info
|
||||
extern int LLAMA_BUILD_NUMBER;
|
||||
extern const char * LLAMA_COMMIT;
|
||||
extern const char * LLAMA_COMPILER;
|
||||
extern const char * LLAMA_BUILD_TARGET;
|
||||
|
||||
struct common_control_vector_load_info;
|
||||
extern char const *LLAMA_COMMIT;
|
||||
extern char const *LLAMA_COMPILER;
|
||||
extern char const *LLAMA_BUILD_TARGET;
|
||||
|
||||
//
|
||||
// CPU utils
|
||||
// CLI argument parsing
|
||||
//
|
||||
int32_t get_num_physical_cores();
|
||||
|
||||
struct cpu_params {
|
||||
int n_threads = -1;
|
||||
bool cpumask[GGML_MAX_N_THREADS] = {false}; // CPU affinity mask.
|
||||
bool mask_valid = false; // Default: any CPU
|
||||
enum ggml_sched_priority priority = GGML_SCHED_PRIO_NORMAL; // Scheduling prio : (0 - normal, 1 - medium, 2 - high, 3 - realtime)
|
||||
bool strict_cpu = false; // Use strict CPU placement
|
||||
uint32_t poll = 50; // Polling (busywait) level (0 - no polling, 100 - mostly polling)
|
||||
};
|
||||
struct gpt_params {
|
||||
uint32_t seed = -1; // RNG seed
|
||||
|
||||
int32_t cpu_get_num_physical_cores();
|
||||
int32_t cpu_get_num_math();
|
||||
int32_t n_threads = get_num_physical_cores();
|
||||
int32_t n_threads_batch = -1; // number of threads to use for batch processing (-1 = use n_threads)
|
||||
int32_t n_predict = -1; // new tokens to predict
|
||||
int32_t n_ctx = 512; // context size
|
||||
int32_t n_batch = 512; // batch size for prompt processing (must be >=32 to use BLAS)
|
||||
int32_t n_keep = 0; // number of tokens to keep from initial prompt
|
||||
int32_t n_draft = 16; // number of tokens to draft during speculative decoding
|
||||
int32_t n_chunks = -1; // max number of chunks to process (-1 = unlimited)
|
||||
int32_t n_parallel = 1; // number of parallel sequences to decode
|
||||
int32_t n_sequences = 1; // number of sequences to decode
|
||||
float p_accept = 0.5f; // speculative decoding accept probability
|
||||
float p_split = 0.1f; // speculative decoding split probability
|
||||
int32_t n_gpu_layers = -1; // number of layers to store in VRAM (-1 - use default)
|
||||
int32_t n_gpu_layers_draft = -1; // number of layers to store in VRAM for the draft model (-1 - use default)
|
||||
int32_t main_gpu = 0; // the GPU that is used for scratch and small tensors
|
||||
float tensor_split[LLAMA_MAX_DEVICES] = {0}; // how split tensors should be distributed across GPUs
|
||||
int32_t n_beams = 0; // if non-zero then use beam search of given width.
|
||||
float rope_freq_base = 0.0f; // RoPE base frequency
|
||||
float rope_freq_scale = 0.0f; // RoPE frequency scaling factor
|
||||
float yarn_ext_factor = -1.0f; // YaRN extrapolation mix factor
|
||||
float yarn_attn_factor = 1.0f; // YaRN magnitude scaling factor
|
||||
float yarn_beta_fast = 32.0f; // YaRN low correction dim
|
||||
float yarn_beta_slow = 1.0f; // YaRN high correction dim
|
||||
int32_t yarn_orig_ctx = 0; // YaRN original context length
|
||||
int8_t rope_scaling_type = LLAMA_ROPE_SCALING_UNSPECIFIED; // TODO: better to be int32_t for alignment
|
||||
// pinging @cebtenzzre
|
||||
|
||||
//
|
||||
// Common params
|
||||
//
|
||||
// // sampling parameters
|
||||
struct llama_sampling_params sparams;
|
||||
|
||||
enum llama_example {
|
||||
LLAMA_EXAMPLE_COMMON,
|
||||
LLAMA_EXAMPLE_SPECULATIVE,
|
||||
LLAMA_EXAMPLE_MAIN,
|
||||
LLAMA_EXAMPLE_INFILL,
|
||||
LLAMA_EXAMPLE_EMBEDDING,
|
||||
LLAMA_EXAMPLE_PERPLEXITY,
|
||||
LLAMA_EXAMPLE_RETRIEVAL,
|
||||
LLAMA_EXAMPLE_PASSKEY,
|
||||
LLAMA_EXAMPLE_IMATRIX,
|
||||
LLAMA_EXAMPLE_BENCH,
|
||||
LLAMA_EXAMPLE_SERVER,
|
||||
LLAMA_EXAMPLE_CVECTOR_GENERATOR,
|
||||
LLAMA_EXAMPLE_EXPORT_LORA,
|
||||
LLAMA_EXAMPLE_LLAVA,
|
||||
LLAMA_EXAMPLE_LOOKUP,
|
||||
LLAMA_EXAMPLE_PARALLEL,
|
||||
LLAMA_EXAMPLE_TTS,
|
||||
std::string model = "models/7B/ggml-model-f16.gguf"; // model path
|
||||
std::string model_draft = ""; // draft model for speculative decoding
|
||||
std::string model_alias = "unknown"; // model alias
|
||||
std::string prompt = "";
|
||||
std::string prompt_file = ""; // store the external prompt file name
|
||||
std::string path_prompt_cache = ""; // path to file for saving/loading prompt eval state
|
||||
std::string input_prefix = ""; // string to prefix user inputs with
|
||||
std::string input_suffix = ""; // string to suffix user inputs with
|
||||
std::vector<std::string> antiprompt; // string upon seeing which more user input is prompted
|
||||
std::string logdir = ""; // directory in which to save YAML log files
|
||||
|
||||
LLAMA_EXAMPLE_COUNT,
|
||||
};
|
||||
|
||||
enum common_sampler_type {
|
||||
COMMON_SAMPLER_TYPE_NONE = 0,
|
||||
COMMON_SAMPLER_TYPE_DRY = 1,
|
||||
COMMON_SAMPLER_TYPE_TOP_K = 2,
|
||||
COMMON_SAMPLER_TYPE_TOP_P = 3,
|
||||
COMMON_SAMPLER_TYPE_MIN_P = 4,
|
||||
//COMMON_SAMPLER_TYPE_TFS_Z = 5,
|
||||
COMMON_SAMPLER_TYPE_TYPICAL_P = 6,
|
||||
COMMON_SAMPLER_TYPE_TEMPERATURE = 7,
|
||||
COMMON_SAMPLER_TYPE_XTC = 8,
|
||||
COMMON_SAMPLER_TYPE_INFILL = 9,
|
||||
COMMON_SAMPLER_TYPE_PENALTIES = 10,
|
||||
};
|
||||
|
||||
// dimensionality reduction methods, used by cvector-generator
|
||||
enum dimre_method {
|
||||
DIMRE_METHOD_PCA,
|
||||
DIMRE_METHOD_MEAN,
|
||||
};
|
||||
|
||||
enum common_conversation_mode {
|
||||
COMMON_CONVERSATION_MODE_DISABLED = 0,
|
||||
COMMON_CONVERSATION_MODE_ENABLED = 1,
|
||||
COMMON_CONVERSATION_MODE_AUTO = 2,
|
||||
};
|
||||
|
||||
struct common_grammar_trigger {
|
||||
std::string word;
|
||||
bool at_start;
|
||||
};
|
||||
|
||||
// sampling parameters
|
||||
struct common_params_sampling {
|
||||
uint32_t seed = LLAMA_DEFAULT_SEED; // the seed used to initialize llama_sampler
|
||||
|
||||
int32_t n_prev = 64; // number of previous tokens to remember
|
||||
int32_t n_probs = 0; // if greater than 0, output the probabilities of top n_probs tokens.
|
||||
int32_t min_keep = 0; // 0 = disabled, otherwise samplers should return at least min_keep tokens
|
||||
int32_t top_k = 40; // <= 0 to use vocab size
|
||||
float top_p = 0.95f; // 1.0 = disabled
|
||||
float min_p = 0.05f; // 0.0 = disabled
|
||||
float xtc_probability = 0.00f; // 0.0 = disabled
|
||||
float xtc_threshold = 0.10f; // > 0.5 disables XTC
|
||||
float typ_p = 1.00f; // typical_p, 1.0 = disabled
|
||||
float temp = 0.80f; // <= 0.0 to sample greedily, 0.0 to not output probabilities
|
||||
float dynatemp_range = 0.00f; // 0.0 = disabled
|
||||
float dynatemp_exponent = 1.00f; // controls how entropy maps to temperature in dynamic temperature sampler
|
||||
int32_t penalty_last_n = 64; // last n tokens to penalize (0 = disable penalty, -1 = context size)
|
||||
float penalty_repeat = 1.00f; // 1.0 = disabled
|
||||
float penalty_freq = 0.00f; // 0.0 = disabled
|
||||
float penalty_present = 0.00f; // 0.0 = disabled
|
||||
float dry_multiplier = 0.0f; // 0.0 = disabled; DRY repetition penalty for tokens extending repetition:
|
||||
float dry_base = 1.75f; // 0.0 = disabled; multiplier * base ^ (length of sequence before token - allowed length)
|
||||
int32_t dry_allowed_length = 2; // tokens extending repetitions beyond this receive penalty
|
||||
int32_t dry_penalty_last_n = -1; // how many tokens to scan for repetitions (0 = disable penalty, -1 = context size)
|
||||
int32_t mirostat = 0; // 0 = disabled, 1 = mirostat, 2 = mirostat 2.0
|
||||
float mirostat_tau = 5.00f; // target entropy
|
||||
float mirostat_eta = 0.10f; // learning rate
|
||||
bool ignore_eos = false;
|
||||
bool no_perf = false; // disable performance metrics
|
||||
bool timing_per_token = false;
|
||||
|
||||
std::vector<std::string> dry_sequence_breakers = {"\n", ":", "\"", "*"}; // default sequence breakers for DRY
|
||||
|
||||
|
||||
std::vector<enum common_sampler_type> samplers = {
|
||||
COMMON_SAMPLER_TYPE_PENALTIES,
|
||||
COMMON_SAMPLER_TYPE_DRY,
|
||||
COMMON_SAMPLER_TYPE_TOP_K,
|
||||
COMMON_SAMPLER_TYPE_TYPICAL_P,
|
||||
COMMON_SAMPLER_TYPE_TOP_P,
|
||||
COMMON_SAMPLER_TYPE_MIN_P,
|
||||
COMMON_SAMPLER_TYPE_XTC,
|
||||
COMMON_SAMPLER_TYPE_TEMPERATURE,
|
||||
};
|
||||
|
||||
std::string grammar; // optional BNF-like grammar to constrain sampling
|
||||
bool grammar_lazy = false;
|
||||
std::vector<common_grammar_trigger> grammar_trigger_words; // optional trigger words to trigger lazy grammar
|
||||
std::vector<llama_token> grammar_trigger_tokens; // optional trigger tokens to trigger lazy grammar and print trigger special tokens.
|
||||
std::set<llama_token> preserved_tokens;
|
||||
|
||||
std::vector<llama_logit_bias> logit_bias; // logit biases to apply
|
||||
|
||||
// print the parameters into a string
|
||||
std::string print() const;
|
||||
};
|
||||
|
||||
struct common_params_speculative {
|
||||
std::vector<ggml_backend_dev_t> devices; // devices to use for offloading
|
||||
|
||||
int32_t n_ctx = 0; // draft context size
|
||||
int32_t n_max = 16; // maximum number of tokens to draft during speculative decoding
|
||||
int32_t n_min = 5; // minimum number of draft tokens to use for speculative decoding
|
||||
int32_t n_gpu_layers = -1; // number of layers to store in VRAM for the draft model (-1 - use default)
|
||||
float p_split = 0.1f; // speculative decoding split probability
|
||||
float p_min = 0.9f; // minimum speculative decoding probability (greedy)
|
||||
|
||||
struct cpu_params cpuparams;
|
||||
struct cpu_params cpuparams_batch;
|
||||
|
||||
std::string hf_repo = ""; // HF repo // NOLINT
|
||||
std::string hf_file = ""; // HF file // NOLINT
|
||||
|
||||
std::string model = ""; // draft model for speculative decoding // NOLINT
|
||||
std::string model_url = ""; // model url to download // NOLINT
|
||||
};
|
||||
|
||||
struct common_params_vocoder {
|
||||
std::string hf_repo = ""; // HF repo // NOLINT
|
||||
std::string hf_file = ""; // HF file // NOLINT
|
||||
|
||||
std::string model = ""; // model path // NOLINT
|
||||
std::string model_url = ""; // model url to download // NOLINT
|
||||
|
||||
bool use_guide_tokens = false; // enable guide tokens to improve TTS accuracy // NOLINT
|
||||
};
|
||||
|
||||
struct common_params {
|
||||
int32_t n_predict = -1; // new tokens to predict
|
||||
int32_t n_ctx = 4096; // context size
|
||||
int32_t n_batch = 2048; // logical batch size for prompt processing (must be >=32 to use BLAS)
|
||||
int32_t n_ubatch = 512; // physical batch size for prompt processing (must be >=32 to use BLAS)
|
||||
int32_t n_keep = 0; // number of tokens to keep from initial prompt
|
||||
int32_t n_chunks = -1; // max number of chunks to process (-1 = unlimited)
|
||||
int32_t n_parallel = 1; // number of parallel sequences to decode
|
||||
int32_t n_sequences = 1; // number of sequences to decode
|
||||
int32_t grp_attn_n = 1; // group-attention factor
|
||||
int32_t grp_attn_w = 512; // group-attention width
|
||||
int32_t n_print = -1; // print token count every n tokens (-1 = disabled)
|
||||
float rope_freq_base = 0.0f; // RoPE base frequency
|
||||
float rope_freq_scale = 0.0f; // RoPE frequency scaling factor
|
||||
float yarn_ext_factor = -1.0f; // YaRN extrapolation mix factor
|
||||
float yarn_attn_factor = 1.0f; // YaRN magnitude scaling factor
|
||||
float yarn_beta_fast = 32.0f; // YaRN low correction dim
|
||||
float yarn_beta_slow = 1.0f; // YaRN high correction dim
|
||||
int32_t yarn_orig_ctx = 0; // YaRN original context length
|
||||
float defrag_thold = 0.1f; // KV cache defragmentation threshold
|
||||
|
||||
// offload params
|
||||
std::vector<ggml_backend_dev_t> devices; // devices to use for offloading
|
||||
|
||||
int32_t n_gpu_layers = -1; // number of layers to store in VRAM (-1 - use default)
|
||||
int32_t main_gpu = 0; // the GPU that is used for scratch and small tensors
|
||||
float tensor_split[128] = {0}; // how split tensors should be distributed across GPUs
|
||||
|
||||
enum llama_split_mode split_mode = LLAMA_SPLIT_MODE_LAYER; // how to split the model across GPUs
|
||||
|
||||
struct cpu_params cpuparams;
|
||||
struct cpu_params cpuparams_batch;
|
||||
|
||||
ggml_backend_sched_eval_callback cb_eval = nullptr;
|
||||
void * cb_eval_user_data = nullptr;
|
||||
|
||||
ggml_numa_strategy numa = GGML_NUMA_STRATEGY_DISABLED;
|
||||
|
||||
enum llama_rope_scaling_type rope_scaling_type = LLAMA_ROPE_SCALING_TYPE_UNSPECIFIED;
|
||||
enum llama_pooling_type pooling_type = LLAMA_POOLING_TYPE_UNSPECIFIED; // pooling type for embeddings
|
||||
enum llama_attention_type attention_type = LLAMA_ATTENTION_TYPE_UNSPECIFIED; // attention type for embeddings
|
||||
|
||||
struct common_params_sampling sampling;
|
||||
struct common_params_speculative speculative;
|
||||
struct common_params_vocoder vocoder;
|
||||
|
||||
std::string model = ""; // model path // NOLINT
|
||||
std::string model_alias = ""; // model alias // NOLINT
|
||||
std::string model_url = ""; // model url to download // NOLINT
|
||||
std::string hf_token = ""; // HF token // NOLINT
|
||||
std::string hf_repo = ""; // HF repo // NOLINT
|
||||
std::string hf_file = ""; // HF file // NOLINT
|
||||
std::string prompt = ""; // NOLINT
|
||||
std::string prompt_file = ""; // store the external prompt file name // NOLINT
|
||||
std::string path_prompt_cache = ""; // path to file for saving/loading prompt eval state // NOLINT
|
||||
std::string input_prefix = ""; // string to prefix user inputs with // NOLINT
|
||||
std::string input_suffix = ""; // string to suffix user inputs with // NOLINT
|
||||
std::string lookup_cache_static = ""; // path of static ngram cache file for lookup decoding // NOLINT
|
||||
std::string lookup_cache_dynamic = ""; // path of dynamic ngram cache file for lookup decoding // NOLINT
|
||||
std::string logits_file = ""; // file for saving *all* logits // NOLINT
|
||||
|
||||
std::vector<std::string> in_files; // all input files
|
||||
std::vector<std::string> antiprompt; // strings upon which more user input is prompted (a.k.a. reverse prompts)
|
||||
std::vector<llama_model_kv_override> kv_overrides;
|
||||
|
||||
bool lora_init_without_apply = false; // only load lora to memory, but do not apply it to ctx (user can manually apply lora later using llama_adapter_lora_apply)
|
||||
std::vector<common_adapter_lora_info> lora_adapters; // lora adapter path with user defined scale
|
||||
// TODO: avoid tuple, use struct
|
||||
std::vector<std::tuple<std::string, float>> lora_adapter; // lora adapter path with user defined scale
|
||||
std::string lora_base = ""; // base model path for the lora adapter
|
||||
|
||||
std::vector<common_control_vector_load_info> control_vectors; // control vector with user defined scale
|
||||
int ppl_stride = 0; // stride for perplexity calculations. If left at 0, the pre-existing approach will be used.
|
||||
int ppl_output_type = 0; // = 0 -> ppl output is as usual, = 1 -> ppl output is num_tokens, ppl, one per line
|
||||
// (which is more convenient to use for plotting)
|
||||
//
|
||||
bool hellaswag = false; // compute HellaSwag score over random tasks from datafile supplied in prompt
|
||||
size_t hellaswag_tasks = 400; // number of tasks to use when computing the HellaSwag score
|
||||
|
||||
int32_t verbosity = 0;
|
||||
int32_t control_vector_layer_start = -1; // layer range for control vector
|
||||
int32_t control_vector_layer_end = -1; // layer range for control vector
|
||||
|
||||
int32_t ppl_stride = 0; // stride for perplexity calculations. If left at 0, the pre-existing approach will be used.
|
||||
int32_t ppl_output_type = 0; // = 0 -> ppl output is as usual, = 1 -> ppl output is num_tokens, ppl, one per line
|
||||
// (which is more convenient to use for plotting)
|
||||
//
|
||||
bool hellaswag = false; // compute HellaSwag score over random tasks from datafile supplied in prompt
|
||||
size_t hellaswag_tasks = 400; // number of tasks to use when computing the HellaSwag score
|
||||
|
||||
bool winogrande = false; // compute Winogrande score over random tasks from datafile supplied in prompt
|
||||
size_t winogrande_tasks = 0; // number of tasks to use when computing the Winogrande score. If 0, all tasks will be computed
|
||||
|
||||
bool multiple_choice = false; // compute TruthfulQA score over random tasks from datafile supplied in prompt
|
||||
size_t multiple_choice_tasks = 0; // number of tasks to use when computing the TruthfulQA score. If 0, all tasks will be computed
|
||||
|
||||
bool kl_divergence = false; // compute KL divergence
|
||||
|
||||
bool usage = false; // print usage
|
||||
bool mul_mat_q = true; // if true, use mul_mat_q kernels instead of cuBLAS
|
||||
bool random_prompt = false; // do not randomize prompt if none provided
|
||||
bool use_color = false; // use color to distinguish generations and inputs
|
||||
bool special = false; // enable special token output
|
||||
bool interactive = false; // interactive mode
|
||||
bool interactive_first = false; // wait for user input immediately
|
||||
bool chatml = false; // chatml mode (used for models trained on chatml syntax)
|
||||
bool prompt_cache_all = false; // save user input and generations to prompt cache
|
||||
bool prompt_cache_ro = false; // open the prompt cache read-only and do not update it
|
||||
|
||||
bool escape = true; // escape "\n", "\r", "\t", "\'", "\"", and "\\"
|
||||
bool embedding = false; // get only sentence embedding
|
||||
bool escape = false; // escape "\n", "\r", "\t", "\'", "\"", and "\\"
|
||||
bool interactive_first = false; // wait for user input immediately
|
||||
bool multiline_input = false; // reverse the usage of `\`
|
||||
bool simple_io = false; // improves compatibility with subprocesses and limited consoles
|
||||
bool cont_batching = true; // insert new sequences for decoding on-the-fly
|
||||
bool flash_attn = false; // flash attention
|
||||
bool no_perf = false; // disable performance metrics
|
||||
bool ctx_shift = true; // context shift on inifinite text generation
|
||||
bool cont_batching = false; // insert new sequences for decoding on-the-fly
|
||||
|
||||
bool input_prefix_bos = false; // prefix BOS to user inputs, preceding input_prefix
|
||||
bool ignore_eos = false; // ignore generated EOS tokens
|
||||
bool instruct = false; // instruction mode (used for Alpaca models)
|
||||
bool logits_all = false; // return logits for all tokens in the batch
|
||||
bool use_mmap = true; // use mmap for faster loads
|
||||
bool use_mlock = false; // use mlock to keep model in memory
|
||||
bool numa = false; // attempt optimizations that help on some NUMA systems
|
||||
bool verbose_prompt = false; // print prompt tokens before generation
|
||||
bool display_prompt = true; // print prompt before generation
|
||||
bool infill = false; // use infill mode
|
||||
bool dump_kv_cache = false; // dump the KV cache contents for debugging purposes
|
||||
bool no_kv_offload = false; // disable KV offloading
|
||||
bool warmup = true; // warmup run
|
||||
bool check_tensors = false; // validate tensor data
|
||||
|
||||
ggml_type cache_type_k = GGML_TYPE_F16; // KV cache data type for the K
|
||||
ggml_type cache_type_v = GGML_TYPE_F16; // KV cache data type for the V
|
||||
|
||||
common_conversation_mode conversation_mode = COMMON_CONVERSATION_MODE_AUTO;
|
||||
std::string cache_type_k = "f16"; // KV cache data type for the K
|
||||
std::string cache_type_v = "f16"; // KV cache data type for the V
|
||||
|
||||
// multimodal models (see examples/llava)
|
||||
std::string mmproj = ""; // path to multimodal projector // NOLINT
|
||||
std::vector<std::string> image; // path to image file(s)
|
||||
|
||||
// embedding
|
||||
bool embedding = false; // get only sentence embedding
|
||||
int32_t embd_normalize = 2; // normalisation for embeddings (-1=none, 0=max absolute int16, 1=taxicab, 2=euclidean, >2=p-norm)
|
||||
std::string embd_out = ""; // empty = default, "array" = [[],[]...], "json" = openai style, "json+" = same "json" + cosine similarity matrix
|
||||
std::string embd_sep = "\n"; // separator of embeddings
|
||||
bool reranking = false; // enable reranking support on server
|
||||
|
||||
// server params
|
||||
int32_t port = 8080; // server listens on this network port
|
||||
int32_t timeout_read = 600; // http read timeout in seconds
|
||||
int32_t timeout_write = timeout_read; // http write timeout in seconds
|
||||
int32_t n_threads_http = -1; // number of threads to process HTTP requests (TODO: support threadpool)
|
||||
int32_t n_cache_reuse = 0; // min chunk size to reuse from the cache via KV shifting
|
||||
|
||||
std::string hostname = "127.0.0.1";
|
||||
std::string public_path = ""; // NOLINT
|
||||
std::string chat_template = ""; // NOLINT
|
||||
bool use_jinja = false; // NOLINT
|
||||
bool enable_chat_template = true;
|
||||
|
||||
std::vector<std::string> api_keys;
|
||||
|
||||
std::string ssl_file_key = ""; // NOLINT
|
||||
std::string ssl_file_cert = ""; // NOLINT
|
||||
|
||||
// "advanced" endpoints are disabled by default for better security
|
||||
bool webui = true;
|
||||
bool endpoint_slots = false;
|
||||
bool endpoint_props = false; // only control POST requests, not GET
|
||||
bool endpoint_metrics = false;
|
||||
|
||||
bool log_json = false;
|
||||
|
||||
std::string slot_save_path;
|
||||
|
||||
float slot_prompt_similarity = 0.5f;
|
||||
|
||||
// batched-bench params
|
||||
bool is_pp_shared = false;
|
||||
|
||||
std::vector<int32_t> n_pp;
|
||||
std::vector<int32_t> n_tg;
|
||||
std::vector<int32_t> n_pl;
|
||||
|
||||
// retrieval params
|
||||
std::vector<std::string> context_files; // context files to embed
|
||||
|
||||
int32_t chunk_size = 64; // chunk size for context embedding
|
||||
|
||||
std::string chunk_separator = "\n"; // chunk separator for context embedding
|
||||
|
||||
// passkey params
|
||||
int32_t n_junk = 250; // number of times to repeat the junk text
|
||||
int32_t i_pos = -1; // position of the passkey in the junk text
|
||||
|
||||
// imatrix params
|
||||
std::string out_file = "imatrix.dat"; // save the resulting imatrix to this file
|
||||
|
||||
int32_t n_out_freq = 10; // output the imatrix every n_out_freq iterations
|
||||
int32_t n_save_freq = 0; // save the imatrix every n_save_freq iterations
|
||||
int32_t i_chunk = 0; // start processing from this chunk
|
||||
|
||||
bool process_output = false; // collect data for the output tensor
|
||||
bool compute_ppl = true; // whether to compute perplexity
|
||||
|
||||
// cvector-generator params
|
||||
int n_pca_batch = 100;
|
||||
int n_pca_iterations = 1000;
|
||||
dimre_method cvector_dimre_method = DIMRE_METHOD_PCA;
|
||||
std::string cvector_outfile = "control_vector.gguf";
|
||||
std::string cvector_positive_file = "examples/cvector-generator/positive.txt";
|
||||
std::string cvector_negative_file = "examples/cvector-generator/negative.txt";
|
||||
|
||||
bool spm_infill = false; // suffix/prefix/middle pattern for infill
|
||||
|
||||
std::string lora_outfile = "ggml-lora-merged-f16.gguf";
|
||||
|
||||
// batched-bench params
|
||||
bool batched_bench_output_jsonl = false;
|
||||
std::string mmproj = ""; // path to multimodal projector
|
||||
std::string image = ""; // path to an image file
|
||||
};
|
||||
|
||||
// call once at the start of a program if it uses libcommon
|
||||
// initializes the logging system and prints info about the build
|
||||
void common_init();
|
||||
bool gpt_params_parse_ex(int argc, char ** argv, gpt_params & params);
|
||||
|
||||
std::string common_params_get_system_info(const common_params & params);
|
||||
bool gpt_params_parse(int argc, char ** argv, gpt_params & params);
|
||||
|
||||
bool parse_cpu_range(const std::string & range, bool(&boolmask)[GGML_MAX_N_THREADS]);
|
||||
bool parse_cpu_mask(const std::string & mask, bool(&boolmask)[GGML_MAX_N_THREADS]);
|
||||
void postprocess_cpu_params(cpu_params & cpuparams, const cpu_params * role_model = nullptr);
|
||||
bool set_process_priority(enum ggml_sched_priority prio);
|
||||
void gpt_print_usage(int argc, char ** argv, const gpt_params & params);
|
||||
|
||||
std::string get_system_info(const gpt_params & params);
|
||||
|
||||
std::string gpt_random_prompt(std::mt19937 & rng);
|
||||
|
||||
void process_escapes(std::string& input);
|
||||
|
||||
//
|
||||
// String utils
|
||||
// String parsing
|
||||
//
|
||||
|
||||
#ifdef __GNUC__
|
||||
#ifdef __MINGW32__
|
||||
#define LLAMA_COMMON_ATTRIBUTE_FORMAT(...) __attribute__((format(gnu_printf, __VA_ARGS__)))
|
||||
#else
|
||||
#define LLAMA_COMMON_ATTRIBUTE_FORMAT(...) __attribute__((format(printf, __VA_ARGS__)))
|
||||
#endif
|
||||
#else
|
||||
#define LLAMA_COMMON_ATTRIBUTE_FORMAT(...)
|
||||
#endif
|
||||
|
||||
LLAMA_COMMON_ATTRIBUTE_FORMAT(1, 2)
|
||||
std::string string_format(const char * fmt, ...);
|
||||
|
||||
std::string string_strip(const std::string & str);
|
||||
std::string string_get_sortable_timestamp();
|
||||
|
||||
std::string string_join(const std::vector<std::string> & values, const std::string & separator);
|
||||
std::vector<std::string> string_split(const std::string & str, const std::string & delimiter);
|
||||
std::string string_repeat(const std::string & str, size_t n);
|
||||
|
||||
void string_replace_all(std::string & s, const std::string & search, const std::string & replace);
|
||||
|
||||
template<class T>
|
||||
static std::vector<T> string_split(const std::string & str, char delim) {
|
||||
static_assert(!std::is_same<T, std::string>::value, "Please use the specialized version for std::string");
|
||||
std::vector<T> values;
|
||||
std::istringstream str_stream(str);
|
||||
std::string token;
|
||||
while (std::getline(str_stream, token, delim)) {
|
||||
T value;
|
||||
std::istringstream token_stream(token);
|
||||
token_stream >> value;
|
||||
values.push_back(value);
|
||||
}
|
||||
return values;
|
||||
}
|
||||
|
||||
template<>
|
||||
std::vector<std::string> string_split<std::string>(const std::string & input, char separator)
|
||||
{
|
||||
std::vector<std::string> parts;
|
||||
size_t begin_pos = 0;
|
||||
size_t separator_pos = input.find(separator);
|
||||
while (separator_pos != std::string::npos) {
|
||||
std::string part = input.substr(begin_pos, separator_pos - begin_pos);
|
||||
parts.emplace_back(part);
|
||||
begin_pos = separator_pos + 1;
|
||||
separator_pos = input.find(separator, begin_pos);
|
||||
}
|
||||
parts.emplace_back(input.substr(begin_pos, separator_pos - begin_pos));
|
||||
return parts;
|
||||
}
|
||||
|
||||
static bool string_starts_with(const std::string & str,
|
||||
const std::string & prefix) { // While we wait for C++20's std::string::starts_with...
|
||||
return str.rfind(prefix, 0) == 0;
|
||||
}
|
||||
|
||||
static bool string_ends_with(const std::string & str,
|
||||
const std::string & suffix) { // While we wait for C++20's std::string::ends_with...
|
||||
return str.size() >= suffix.size() && str.compare(str.size()-suffix.size(), suffix.size(), suffix) == 0;
|
||||
}
|
||||
|
||||
bool string_parse_kv_override(const char * data, std::vector<llama_model_kv_override> & overrides);
|
||||
void string_process_escapes(std::string & input);
|
||||
|
||||
std::string string_from(bool value);
|
||||
std::string string_from(const std::vector<int> & values);
|
||||
std::string string_from(const struct llama_context * ctx, const std::vector<llama_token> & tokens);
|
||||
std::string string_from(const struct llama_context * ctx, const struct llama_batch & batch);
|
||||
|
||||
//
|
||||
// Filesystem utils
|
||||
//
|
||||
|
||||
bool fs_validate_filename(const std::string & filename);
|
||||
bool fs_create_directory_with_parents(const std::string & path);
|
||||
|
||||
std::string fs_get_cache_directory();
|
||||
std::string fs_get_cache_file(const std::string & filename);
|
||||
std::string parse_samplers_input(std::string input);
|
||||
|
||||
//
|
||||
// Model utils
|
||||
//
|
||||
|
||||
// note: defines object's lifetime
|
||||
struct common_init_result {
|
||||
llama_model_ptr model;
|
||||
llama_context_ptr context;
|
||||
// TODO: avoid tuplue, use struct
|
||||
std::tuple<struct llama_model *, struct llama_context *> llama_init_from_gpt_params(gpt_params & params);
|
||||
|
||||
std::vector<llama_adapter_lora_ptr> lora;
|
||||
};
|
||||
struct llama_model_params llama_model_params_from_gpt_params (const gpt_params & params);
|
||||
struct llama_context_params llama_context_params_from_gpt_params(const gpt_params & params);
|
||||
|
||||
struct common_init_result common_init_from_params(common_params & params);
|
||||
|
||||
struct llama_model_params common_model_params_to_llama ( common_params & params);
|
||||
struct llama_context_params common_context_params_to_llama(const common_params & params);
|
||||
struct ggml_threadpool_params ggml_threadpool_params_from_cpu_params(const cpu_params & params);
|
||||
|
||||
struct llama_model * common_load_model_from_url(
|
||||
const std::string & model_url,
|
||||
const std::string & local_path,
|
||||
const std::string & hf_token,
|
||||
const struct llama_model_params & params);
|
||||
|
||||
struct llama_model * common_load_model_from_hf(
|
||||
const std::string & repo,
|
||||
const std::string & remote_path,
|
||||
const std::string & local_path,
|
||||
const std::string & hf_token,
|
||||
const struct llama_model_params & params);
|
||||
|
||||
std::pair<std::string, std::string> common_get_hf_file(
|
||||
const std::string & hf_repo_with_tag,
|
||||
const std::string & hf_token);
|
||||
|
||||
// clear LoRA adapters from context, then apply new list of adapters
|
||||
void common_set_adapter_lora(struct llama_context * ctx, std::vector<common_adapter_lora_info> & lora);
|
||||
|
||||
//
|
||||
// Batch utils
|
||||
//
|
||||
|
||||
void common_batch_clear(struct llama_batch & batch);
|
||||
void llama_batch_clear(struct llama_batch & batch);
|
||||
|
||||
void common_batch_add(
|
||||
void llama_batch_add(
|
||||
struct llama_batch & batch,
|
||||
llama_token id,
|
||||
llama_pos pos,
|
||||
const std::vector<llama_seq_id> & seq_ids,
|
||||
bool logits);
|
||||
|
||||
//
|
||||
// Token utils
|
||||
//
|
||||
|
||||
// longest common prefix
|
||||
size_t common_lcp(const llama_tokens & a, const llama_tokens & b);
|
||||
|
||||
// longet common subsequence
|
||||
size_t common_lcs(const llama_tokens & a, const llama_tokens & b);
|
||||
|
||||
//
|
||||
// Vocab utils
|
||||
//
|
||||
|
||||
// tokenizes a string into a vector of tokens
|
||||
// should work similar to Python's `tokenizer.encode`
|
||||
std::vector<llama_token> common_tokenize(
|
||||
std::vector<llama_token> llama_tokenize(
|
||||
const struct llama_context * ctx,
|
||||
const std::string & text,
|
||||
bool add_special,
|
||||
bool parse_special = false);
|
||||
bool add_bos,
|
||||
bool special = false);
|
||||
|
||||
std::vector<llama_token> common_tokenize(
|
||||
const struct llama_vocab * vocab,
|
||||
std::vector<llama_token> llama_tokenize(
|
||||
const struct llama_model * model,
|
||||
const std::string & text,
|
||||
bool add_special,
|
||||
bool parse_special = false);
|
||||
bool add_bos,
|
||||
bool special = false);
|
||||
|
||||
// tokenizes a token into a piece, optionally renders special/control tokens
|
||||
// tokenizes a token into a piece
|
||||
// should work similar to Python's `tokenizer.id_to_piece`
|
||||
std::string common_token_to_piece(
|
||||
std::string llama_token_to_piece(
|
||||
const struct llama_context * ctx,
|
||||
llama_token token,
|
||||
bool special = true);
|
||||
llama_token token);
|
||||
|
||||
std::string common_token_to_piece(
|
||||
const struct llama_vocab * vocab,
|
||||
llama_token token,
|
||||
bool special = true);
|
||||
// TODO: these should be moved in llama.h C-style API under single `llama_detokenize` function
|
||||
// that takes into account the tokenizer type and decides how to handle the leading space
|
||||
//
|
||||
// detokenizes a vector of tokens into a string
|
||||
// should work similar to Python's `tokenizer.decode`
|
||||
// removes the leading space from the first non-BOS token
|
||||
std::string llama_detokenize_spm(
|
||||
llama_context * ctx,
|
||||
const std::vector<llama_token> & tokens);
|
||||
|
||||
// detokenizes a vector of tokens into a string
|
||||
// should work similar to Python's `tokenizer.decode`
|
||||
// optionally renders special/control tokens
|
||||
std::string common_detokenize(
|
||||
const struct llama_context * ctx,
|
||||
const std::vector<llama_token> & tokens,
|
||||
bool special = true);
|
||||
std::string llama_detokenize_bpe(
|
||||
llama_context * ctx,
|
||||
const std::vector<llama_token> & tokens);
|
||||
|
||||
std::string common_detokenize(
|
||||
const struct llama_vocab * vocab,
|
||||
const std::vector<llama_token> & tokens,
|
||||
bool special = true);
|
||||
// Uses the value from the model metadata if possible, otherwise
|
||||
// defaults to true when model type is SPM, otherwise false.
|
||||
bool llama_should_add_bos_token(const llama_model * model);
|
||||
|
||||
//
|
||||
// Chat template utils
|
||||
// YAML utils
|
||||
//
|
||||
|
||||
struct common_tool_call {
|
||||
std::string name;
|
||||
std::string arguments;
|
||||
std::string id;
|
||||
};
|
||||
bool create_directory_with_parents(const std::string & path);
|
||||
void dump_vector_float_yaml(FILE * stream, const char * prop_name, const std::vector<float> & data);
|
||||
void dump_vector_int_yaml(FILE * stream, const char * prop_name, const std::vector<int> & data);
|
||||
void dump_string_yaml_multiline(FILE * stream, const char * prop_name, const char * data);
|
||||
std::string get_sortable_timestamp();
|
||||
|
||||
// same with llama_chat_message, but uses std::string
|
||||
struct common_chat_msg {
|
||||
std::string role;
|
||||
std::string content;
|
||||
std::vector<common_tool_call> tool_calls;
|
||||
std::string tool_plan = "";
|
||||
};
|
||||
|
||||
// Check if the template supplied via "--chat-template" is supported or not. Returns true if it's valid
|
||||
bool common_chat_verify_template(const std::string & tmpl, bool use_jinja);
|
||||
|
||||
namespace minja {
|
||||
class chat_template;
|
||||
}
|
||||
|
||||
typedef minja::chat_template common_chat_template;
|
||||
|
||||
struct common_chat_templates {
|
||||
bool has_explicit_template; // Model had builtin template or template overridde was specified.
|
||||
std::unique_ptr<common_chat_template> template_default; // always set (defaults to chatml)
|
||||
std::unique_ptr<common_chat_template> template_tool_use;
|
||||
};
|
||||
|
||||
// CPP wrapper for llama_chat_apply_template
|
||||
// If the built-in template is not supported, we default to chatml
|
||||
// If the custom "tmpl" is not supported, we throw an error
|
||||
std::string common_chat_apply_template(
|
||||
const common_chat_template & tmpl,
|
||||
const std::vector<common_chat_msg> & chat,
|
||||
bool add_ass,
|
||||
bool use_jinja);
|
||||
|
||||
// Format single message, while taking into account the position of that message in chat history
|
||||
std::string common_chat_format_single(
|
||||
const common_chat_template & tmpl,
|
||||
const std::vector<common_chat_msg> & past_msg,
|
||||
const common_chat_msg & new_msg,
|
||||
bool add_ass,
|
||||
bool use_jinja);
|
||||
|
||||
// Returns an example of formatted chat
|
||||
std::string common_chat_format_example(
|
||||
const common_chat_template & tmpl, bool use_jinja);
|
||||
|
||||
common_chat_templates common_chat_templates_from_model(const struct llama_model * model, const std::string & chat_template_override);
|
||||
void dump_non_result_info_yaml(
|
||||
FILE * stream, const gpt_params & params, const llama_context * lctx,
|
||||
const std::string & timestamp, const std::vector<int> & prompt_tokens, const char * model_desc);
|
||||
|
||||
//
|
||||
// KV cache utils
|
||||
//
|
||||
|
||||
// Dump the KV cache view with the number of sequences per cell.
|
||||
void common_kv_cache_dump_view(const llama_kv_cache_view & view, int row_size = 80);
|
||||
void dump_kv_cache_view(const llama_kv_cache_view & view, int row_size = 80);
|
||||
|
||||
// Dump the KV cache view showing individual sequences in each cell (long output).
|
||||
void common_kv_cache_dump_view_seqs(const llama_kv_cache_view & view, int row_size = 40);
|
||||
|
||||
//
|
||||
// Embedding utils
|
||||
//
|
||||
|
||||
// TODO: repace embd_norm with an enum
|
||||
void common_embd_normalize(const float * inp, float * out, int n, int embd_norm);
|
||||
|
||||
float common_embd_similarity_cos(const float * embd1, const float * embd2, int n);
|
||||
|
||||
//
|
||||
// Control vector utils
|
||||
//
|
||||
|
||||
struct common_control_vector_data {
|
||||
int n_embd;
|
||||
|
||||
// stores data for layers [1, n_layer] where n_layer = data.size() / n_embd
|
||||
std::vector<float> data;
|
||||
};
|
||||
|
||||
struct common_control_vector_load_info {
|
||||
float strength;
|
||||
|
||||
std::string fname;
|
||||
};
|
||||
|
||||
// Load control vectors, scale each by strength, and add them together.
|
||||
// On error, returns {-1, empty}
|
||||
common_control_vector_data common_control_vector_load(const std::vector<common_control_vector_load_info> & load_infos);
|
||||
|
||||
//
|
||||
// Split utils
|
||||
//
|
||||
|
||||
namespace {
|
||||
|
||||
const char * const LLM_KV_SPLIT_NO = "split.no";
|
||||
const char * const LLM_KV_SPLIT_COUNT = "split.count";
|
||||
const char * const LLM_KV_SPLIT_TENSORS_COUNT = "split.tensors.count";
|
||||
|
||||
}
|
||||
void dump_kv_cache_view_seqs(const llama_kv_cache_view & view, int row_size = 40);
|
||||
|
|
|
@ -94,9 +94,6 @@ namespace console {
|
|||
simple_io = true;
|
||||
}
|
||||
}
|
||||
if (simple_io) {
|
||||
_setmode(_fileno(stdin), _O_U8TEXT);
|
||||
}
|
||||
#else
|
||||
// POSIX-specific console initialization
|
||||
if (!simple_io) {
|
||||
|
|
424
common/grammar-parser.cpp
Normal file
424
common/grammar-parser.cpp
Normal file
|
@ -0,0 +1,424 @@
|
|||
#include "grammar-parser.h"
|
||||
#include <cstdint>
|
||||
#include <cwchar>
|
||||
#include <string>
|
||||
#include <utility>
|
||||
#include <stdexcept>
|
||||
#include <exception>
|
||||
|
||||
namespace grammar_parser {
|
||||
// NOTE: assumes valid utf8 (but checks for overrun)
|
||||
// copied from llama.cpp
|
||||
static std::pair<uint32_t, const char *> decode_utf8(const char * src) {
|
||||
static const int lookup[] = { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 3, 4 };
|
||||
uint8_t first_byte = static_cast<uint8_t>(*src);
|
||||
uint8_t highbits = first_byte >> 4;
|
||||
int len = lookup[highbits];
|
||||
uint8_t mask = (1 << (8 - len)) - 1;
|
||||
uint32_t value = first_byte & mask;
|
||||
const char * end = src + len; // may overrun!
|
||||
const char * pos = src + 1;
|
||||
for ( ; pos < end && *pos; pos++) {
|
||||
value = (value << 6) + (static_cast<uint8_t>(*pos) & 0x3F);
|
||||
}
|
||||
return std::make_pair(value, pos);
|
||||
}
|
||||
|
||||
static uint32_t get_symbol_id(parse_state & state, const char * src, size_t len) {
|
||||
uint32_t next_id = static_cast<uint32_t>(state.symbol_ids.size());
|
||||
auto result = state.symbol_ids.insert(std::make_pair(std::string(src, len), next_id));
|
||||
return result.first->second;
|
||||
}
|
||||
|
||||
static uint32_t generate_symbol_id(parse_state & state, const std::string & base_name) {
|
||||
uint32_t next_id = static_cast<uint32_t>(state.symbol_ids.size());
|
||||
state.symbol_ids[base_name + '_' + std::to_string(next_id)] = next_id;
|
||||
return next_id;
|
||||
}
|
||||
|
||||
static void add_rule(
|
||||
parse_state & state,
|
||||
uint32_t rule_id,
|
||||
const std::vector<llama_grammar_element> & rule) {
|
||||
if (state.rules.size() <= rule_id) {
|
||||
state.rules.resize(rule_id + 1);
|
||||
}
|
||||
state.rules[rule_id] = rule;
|
||||
}
|
||||
|
||||
static bool is_word_char(char c) {
|
||||
return ('a' <= c && c <= 'z') || ('A' <= c && c <= 'Z') || c == '-' || ('0' <= c && c <= '9');
|
||||
}
|
||||
|
||||
static std::pair<uint32_t, const char *> parse_hex(const char * src, int size) {
|
||||
const char * pos = src;
|
||||
const char * end = src + size;
|
||||
uint32_t value = 0;
|
||||
for ( ; pos < end && *pos; pos++) {
|
||||
value <<= 4;
|
||||
char c = *pos;
|
||||
if ('a' <= c && c <= 'f') {
|
||||
value += c - 'a' + 10;
|
||||
} else if ('A' <= c && c <= 'F') {
|
||||
value += c - 'A' + 10;
|
||||
} else if ('0' <= c && c <= '9') {
|
||||
value += c - '0';
|
||||
} else {
|
||||
break;
|
||||
}
|
||||
}
|
||||
if (pos != end) {
|
||||
throw std::runtime_error("expecting " + std::to_string(size) + " hex chars at " + src);
|
||||
}
|
||||
return std::make_pair(value, pos);
|
||||
}
|
||||
|
||||
static const char * parse_space(const char * src, bool newline_ok) {
|
||||
const char * pos = src;
|
||||
while (*pos == ' ' || *pos == '\t' || *pos == '#' ||
|
||||
(newline_ok && (*pos == '\r' || *pos == '\n'))) {
|
||||
if (*pos == '#') {
|
||||
while (*pos && *pos != '\r' && *pos != '\n') {
|
||||
pos++;
|
||||
}
|
||||
} else {
|
||||
pos++;
|
||||
}
|
||||
}
|
||||
return pos;
|
||||
}
|
||||
|
||||
static const char * parse_name(const char * src) {
|
||||
const char * pos = src;
|
||||
while (is_word_char(*pos)) {
|
||||
pos++;
|
||||
}
|
||||
if (pos == src) {
|
||||
throw std::runtime_error(std::string("expecting name at ") + src);
|
||||
}
|
||||
return pos;
|
||||
}
|
||||
|
||||
static std::pair<uint32_t, const char *> parse_char(const char * src) {
|
||||
if (*src == '\\') {
|
||||
switch (src[1]) {
|
||||
case 'x': return parse_hex(src + 2, 2);
|
||||
case 'u': return parse_hex(src + 2, 4);
|
||||
case 'U': return parse_hex(src + 2, 8);
|
||||
case 't': return std::make_pair('\t', src + 2);
|
||||
case 'r': return std::make_pair('\r', src + 2);
|
||||
case 'n': return std::make_pair('\n', src + 2);
|
||||
case '\\':
|
||||
case '"':
|
||||
case '[':
|
||||
case ']':
|
||||
return std::make_pair(src[1], src + 2);
|
||||
default:
|
||||
throw std::runtime_error(std::string("unknown escape at ") + src);
|
||||
}
|
||||
} else if (*src) {
|
||||
return decode_utf8(src);
|
||||
}
|
||||
throw std::runtime_error("unexpected end of input");
|
||||
}
|
||||
|
||||
const char * parse_alternates(
|
||||
parse_state & state,
|
||||
const char * src,
|
||||
const std::string & rule_name,
|
||||
uint32_t rule_id,
|
||||
bool is_nested);
|
||||
|
||||
static const char * parse_sequence(
|
||||
parse_state & state,
|
||||
const char * src,
|
||||
const std::string & rule_name,
|
||||
std::vector<llama_grammar_element> & out_elements,
|
||||
bool is_nested) {
|
||||
size_t last_sym_start = out_elements.size();
|
||||
const char * pos = src;
|
||||
while (*pos) {
|
||||
if (*pos == '"') { // literal string
|
||||
pos++;
|
||||
last_sym_start = out_elements.size();
|
||||
while (*pos != '"') {
|
||||
auto char_pair = parse_char(pos);
|
||||
pos = char_pair.second;
|
||||
out_elements.push_back({LLAMA_GRETYPE_CHAR, char_pair.first});
|
||||
}
|
||||
pos = parse_space(pos + 1, is_nested);
|
||||
} else if (*pos == '[') { // char range(s)
|
||||
pos++;
|
||||
enum llama_gretype start_type = LLAMA_GRETYPE_CHAR;
|
||||
if (*pos == '^') {
|
||||
pos++;
|
||||
start_type = LLAMA_GRETYPE_CHAR_NOT;
|
||||
}
|
||||
last_sym_start = out_elements.size();
|
||||
while (*pos != ']') {
|
||||
auto char_pair = parse_char(pos);
|
||||
pos = char_pair.second;
|
||||
enum llama_gretype type = last_sym_start < out_elements.size()
|
||||
? LLAMA_GRETYPE_CHAR_ALT
|
||||
: start_type;
|
||||
|
||||
out_elements.push_back({type, char_pair.first});
|
||||
if (pos[0] == '-' && pos[1] != ']') {
|
||||
auto endchar_pair = parse_char(pos + 1);
|
||||
pos = endchar_pair.second;
|
||||
out_elements.push_back({LLAMA_GRETYPE_CHAR_RNG_UPPER, endchar_pair.first});
|
||||
}
|
||||
}
|
||||
pos = parse_space(pos + 1, is_nested);
|
||||
} else if (is_word_char(*pos)) { // rule reference
|
||||
const char * name_end = parse_name(pos);
|
||||
uint32_t ref_rule_id = get_symbol_id(state, pos, name_end - pos);
|
||||
pos = parse_space(name_end, is_nested);
|
||||
last_sym_start = out_elements.size();
|
||||
out_elements.push_back({LLAMA_GRETYPE_RULE_REF, ref_rule_id});
|
||||
} else if (*pos == '(') { // grouping
|
||||
// parse nested alternates into synthesized rule
|
||||
pos = parse_space(pos + 1, true);
|
||||
uint32_t sub_rule_id = generate_symbol_id(state, rule_name);
|
||||
pos = parse_alternates(state, pos, rule_name, sub_rule_id, true);
|
||||
last_sym_start = out_elements.size();
|
||||
// output reference to synthesized rule
|
||||
out_elements.push_back({LLAMA_GRETYPE_RULE_REF, sub_rule_id});
|
||||
if (*pos != ')') {
|
||||
throw std::runtime_error(std::string("expecting ')' at ") + pos);
|
||||
}
|
||||
pos = parse_space(pos + 1, is_nested);
|
||||
} else if (*pos == '*' || *pos == '+' || *pos == '?') { // repetition operator
|
||||
if (last_sym_start == out_elements.size()) {
|
||||
throw std::runtime_error(std::string("expecting preceding item to */+/? at ") + pos);
|
||||
}
|
||||
|
||||
// apply transformation to previous symbol (last_sym_start to end) according to
|
||||
// rewrite rules:
|
||||
// S* --> S' ::= S S' |
|
||||
// S+ --> S' ::= S S' | S
|
||||
// S? --> S' ::= S |
|
||||
uint32_t sub_rule_id = generate_symbol_id(state, rule_name);
|
||||
std::vector<llama_grammar_element> sub_rule;
|
||||
// add preceding symbol to generated rule
|
||||
sub_rule.insert(
|
||||
sub_rule.end(), out_elements.begin() + last_sym_start, out_elements.end());
|
||||
if (*pos == '*' || *pos == '+') {
|
||||
// cause generated rule to recurse
|
||||
sub_rule.push_back({LLAMA_GRETYPE_RULE_REF, sub_rule_id});
|
||||
}
|
||||
// mark start of alternate def
|
||||
sub_rule.push_back({LLAMA_GRETYPE_ALT, 0});
|
||||
if (*pos == '+') {
|
||||
// add preceding symbol as alternate only for '+' (otherwise empty)
|
||||
sub_rule.insert(
|
||||
sub_rule.end(), out_elements.begin() + last_sym_start, out_elements.end());
|
||||
}
|
||||
sub_rule.push_back({LLAMA_GRETYPE_END, 0});
|
||||
add_rule(state, sub_rule_id, sub_rule);
|
||||
|
||||
// in original rule, replace previous symbol with reference to generated rule
|
||||
out_elements.resize(last_sym_start);
|
||||
out_elements.push_back({LLAMA_GRETYPE_RULE_REF, sub_rule_id});
|
||||
|
||||
pos = parse_space(pos + 1, is_nested);
|
||||
} else {
|
||||
break;
|
||||
}
|
||||
}
|
||||
return pos;
|
||||
}
|
||||
|
||||
const char * parse_alternates(
|
||||
parse_state & state,
|
||||
const char * src,
|
||||
const std::string & rule_name,
|
||||
uint32_t rule_id,
|
||||
bool is_nested) {
|
||||
std::vector<llama_grammar_element> rule;
|
||||
const char * pos = parse_sequence(state, src, rule_name, rule, is_nested);
|
||||
while (*pos == '|') {
|
||||
rule.push_back({LLAMA_GRETYPE_ALT, 0});
|
||||
pos = parse_space(pos + 1, true);
|
||||
pos = parse_sequence(state, pos, rule_name, rule, is_nested);
|
||||
}
|
||||
rule.push_back({LLAMA_GRETYPE_END, 0});
|
||||
add_rule(state, rule_id, rule);
|
||||
return pos;
|
||||
}
|
||||
|
||||
static const char * parse_rule(parse_state & state, const char * src) {
|
||||
const char * name_end = parse_name(src);
|
||||
const char * pos = parse_space(name_end, false);
|
||||
size_t name_len = name_end - src;
|
||||
uint32_t rule_id = get_symbol_id(state, src, name_len);
|
||||
const std::string name(src, name_len);
|
||||
|
||||
if (!(pos[0] == ':' && pos[1] == ':' && pos[2] == '=')) {
|
||||
throw std::runtime_error(std::string("expecting ::= at ") + pos);
|
||||
}
|
||||
pos = parse_space(pos + 3, true);
|
||||
|
||||
pos = parse_alternates(state, pos, name, rule_id, false);
|
||||
|
||||
if (*pos == '\r') {
|
||||
pos += pos[1] == '\n' ? 2 : 1;
|
||||
} else if (*pos == '\n') {
|
||||
pos++;
|
||||
} else if (*pos) {
|
||||
throw std::runtime_error(std::string("expecting newline or end at ") + pos);
|
||||
}
|
||||
return parse_space(pos, true);
|
||||
}
|
||||
|
||||
parse_state parse(const char * src) {
|
||||
try {
|
||||
parse_state state;
|
||||
const char * pos = parse_space(src, true);
|
||||
while (*pos) {
|
||||
pos = parse_rule(state, pos);
|
||||
}
|
||||
return state;
|
||||
} catch (const std::exception & err) {
|
||||
fprintf(stderr, "%s: error parsing grammar: %s\n", __func__, err.what());
|
||||
return parse_state();
|
||||
}
|
||||
}
|
||||
|
||||
static void print_grammar_char(FILE * file, uint32_t c) {
|
||||
if (0x20 <= c && c <= 0x7f) {
|
||||
fprintf(file, "%c", static_cast<char>(c));
|
||||
} else {
|
||||
// cop out of encoding UTF-8
|
||||
fprintf(file, "<U+%04X>", c);
|
||||
}
|
||||
}
|
||||
|
||||
static bool is_char_element(llama_grammar_element elem) {
|
||||
switch (elem.type) {
|
||||
case LLAMA_GRETYPE_CHAR: return true;
|
||||
case LLAMA_GRETYPE_CHAR_NOT: return true;
|
||||
case LLAMA_GRETYPE_CHAR_ALT: return true;
|
||||
case LLAMA_GRETYPE_CHAR_RNG_UPPER: return true;
|
||||
default: return false;
|
||||
}
|
||||
}
|
||||
|
||||
static void print_rule_binary(FILE * file, const std::vector<llama_grammar_element> & rule) {
|
||||
for (auto elem : rule) {
|
||||
switch (elem.type) {
|
||||
case LLAMA_GRETYPE_END: fprintf(file, "END"); break;
|
||||
case LLAMA_GRETYPE_ALT: fprintf(file, "ALT"); break;
|
||||
case LLAMA_GRETYPE_RULE_REF: fprintf(file, "RULE_REF"); break;
|
||||
case LLAMA_GRETYPE_CHAR: fprintf(file, "CHAR"); break;
|
||||
case LLAMA_GRETYPE_CHAR_NOT: fprintf(file, "CHAR_NOT"); break;
|
||||
case LLAMA_GRETYPE_CHAR_RNG_UPPER: fprintf(file, "CHAR_RNG_UPPER"); break;
|
||||
case LLAMA_GRETYPE_CHAR_ALT: fprintf(file, "CHAR_ALT"); break;
|
||||
}
|
||||
switch (elem.type) {
|
||||
case LLAMA_GRETYPE_END:
|
||||
case LLAMA_GRETYPE_ALT:
|
||||
case LLAMA_GRETYPE_RULE_REF:
|
||||
fprintf(file, "(%u) ", elem.value);
|
||||
break;
|
||||
case LLAMA_GRETYPE_CHAR:
|
||||
case LLAMA_GRETYPE_CHAR_NOT:
|
||||
case LLAMA_GRETYPE_CHAR_RNG_UPPER:
|
||||
case LLAMA_GRETYPE_CHAR_ALT:
|
||||
fprintf(file, "(\"");
|
||||
print_grammar_char(file, elem.value);
|
||||
fprintf(file, "\") ");
|
||||
break;
|
||||
}
|
||||
}
|
||||
fprintf(file, "\n");
|
||||
}
|
||||
|
||||
static void print_rule(
|
||||
FILE * file,
|
||||
uint32_t rule_id,
|
||||
const std::vector<llama_grammar_element> & rule,
|
||||
const std::map<uint32_t, std::string> & symbol_id_names) {
|
||||
if (rule.empty() || rule.back().type != LLAMA_GRETYPE_END) {
|
||||
throw std::runtime_error(
|
||||
"malformed rule, does not end with LLAMA_GRETYPE_END: " + std::to_string(rule_id));
|
||||
}
|
||||
fprintf(file, "%s ::= ", symbol_id_names.at(rule_id).c_str());
|
||||
for (size_t i = 0, end = rule.size() - 1; i < end; i++) {
|
||||
llama_grammar_element elem = rule[i];
|
||||
switch (elem.type) {
|
||||
case LLAMA_GRETYPE_END:
|
||||
throw std::runtime_error(
|
||||
"unexpected end of rule: " + std::to_string(rule_id) + "," +
|
||||
std::to_string(i));
|
||||
case LLAMA_GRETYPE_ALT:
|
||||
fprintf(file, "| ");
|
||||
break;
|
||||
case LLAMA_GRETYPE_RULE_REF:
|
||||
fprintf(file, "%s ", symbol_id_names.at(elem.value).c_str());
|
||||
break;
|
||||
case LLAMA_GRETYPE_CHAR:
|
||||
fprintf(file, "[");
|
||||
print_grammar_char(file, elem.value);
|
||||
break;
|
||||
case LLAMA_GRETYPE_CHAR_NOT:
|
||||
fprintf(file, "[^");
|
||||
print_grammar_char(file, elem.value);
|
||||
break;
|
||||
case LLAMA_GRETYPE_CHAR_RNG_UPPER:
|
||||
if (i == 0 || !is_char_element(rule[i - 1])) {
|
||||
throw std::runtime_error(
|
||||
"LLAMA_GRETYPE_CHAR_RNG_UPPER without preceding char: " +
|
||||
std::to_string(rule_id) + "," + std::to_string(i));
|
||||
}
|
||||
fprintf(file, "-");
|
||||
print_grammar_char(file, elem.value);
|
||||
break;
|
||||
case LLAMA_GRETYPE_CHAR_ALT:
|
||||
if (i == 0 || !is_char_element(rule[i - 1])) {
|
||||
throw std::runtime_error(
|
||||
"LLAMA_GRETYPE_CHAR_ALT without preceding char: " +
|
||||
std::to_string(rule_id) + "," + std::to_string(i));
|
||||
}
|
||||
print_grammar_char(file, elem.value);
|
||||
break;
|
||||
}
|
||||
if (is_char_element(elem)) {
|
||||
switch (rule[i + 1].type) {
|
||||
case LLAMA_GRETYPE_CHAR_ALT:
|
||||
case LLAMA_GRETYPE_CHAR_RNG_UPPER:
|
||||
break;
|
||||
default:
|
||||
fprintf(file, "] ");
|
||||
}
|
||||
}
|
||||
}
|
||||
fprintf(file, "\n");
|
||||
}
|
||||
|
||||
void print_grammar(FILE * file, const parse_state & state) {
|
||||
try {
|
||||
std::map<uint32_t, std::string> symbol_id_names;
|
||||
for (const auto & kv : state.symbol_ids) {
|
||||
symbol_id_names[kv.second] = kv.first;
|
||||
}
|
||||
for (size_t i = 0, end = state.rules.size(); i < end; i++) {
|
||||
// fprintf(file, "%zu: ", i);
|
||||
// print_rule_binary(file, state.rules[i]);
|
||||
print_rule(file, uint32_t(i), state.rules[i], symbol_id_names);
|
||||
// fprintf(file, "\n");
|
||||
}
|
||||
} catch (const std::exception & err) {
|
||||
fprintf(stderr, "\n%s: error printing grammar: %s\n", __func__, err.what());
|
||||
}
|
||||
}
|
||||
|
||||
std::vector<const llama_grammar_element *> parse_state::c_rules() {
|
||||
std::vector<const llama_grammar_element *> ret;
|
||||
ret.reserve(rules.size());
|
||||
for (const auto & rule : rules) {
|
||||
ret.push_back(rule.data());
|
||||
}
|
||||
return ret;
|
||||
}
|
||||
}
|
29
common/grammar-parser.h
Normal file
29
common/grammar-parser.h
Normal file
|
@ -0,0 +1,29 @@
|
|||
// Implements a parser for an extended Backus-Naur form (BNF), producing the
|
||||
// binary context-free grammar format specified by llama.h. Supports character
|
||||
// ranges, grouping, and repetition operators. As an example, a grammar for
|
||||
// arithmetic might look like:
|
||||
//
|
||||
// root ::= expr
|
||||
// expr ::= term ([-+*/] term)*
|
||||
// term ::= num | "(" space expr ")" space
|
||||
// num ::= [0-9]+ space
|
||||
// space ::= [ \t\n]*
|
||||
|
||||
#pragma once
|
||||
#include "llama.h"
|
||||
#include <vector>
|
||||
#include <map>
|
||||
#include <cstdint>
|
||||
#include <string>
|
||||
|
||||
namespace grammar_parser {
|
||||
struct parse_state {
|
||||
std::map<std::string, uint32_t> symbol_ids;
|
||||
std::vector<std::vector<llama_grammar_element>> rules;
|
||||
|
||||
std::vector<const llama_grammar_element *> c_rules();
|
||||
};
|
||||
|
||||
parse_state parse(const char * src);
|
||||
void print_grammar(FILE * file, const parse_state & state);
|
||||
}
|
Some files were not shown because too many files have changed in this diff Show more
Loading…
Add table
Add a link
Reference in a new issue