Merge branch 'master' of https://github.com/ggerganov/llama.cpp into ntkv2

.github/workflows/build.yml

@@ -10,10 +10,10 @@ on:
   push:
     branches:
       - master
-    paths: ['.github/workflows/**', '**/CMakeLists.txt', '**/Makefile', '**/*.h', '**/*.hpp', '**/*.c', '**/*.cpp', '**/*.cu', '**/*.swift']
+    paths: ['.github/workflows/**', '**/CMakeLists.txt', '**/Makefile', '**/*.h', '**/*.hpp', '**/*.c', '**/*.cpp', '**/*.cu', '**/*.swift', '**/*.m']
   pull_request:
     types: [opened, synchronize, reopened]
-    paths: ['**/CMakeLists.txt', '**/Makefile', '**/*.h', '**/*.hpp', '**/*.c', '**/*.cpp', '**/*.cu', '**/*.swift']
+    paths: ['**/CMakeLists.txt', '**/Makefile', '**/*.h', '**/*.hpp', '**/*.c', '**/*.cpp', '**/*.cu', '**/*.swift', '**/*.m']
 
 env:
   BRANCH_NAME: ${{ github.head_ref || github.ref_name }}
@@ -276,6 +276,11 @@ jobs:
         run: |
           xcodebuild -scheme llama -destination "${{ matrix.destination }}"
 
+      - name: Build Swift Example
+        id: make_build_swift_example
+        run: |
+            make swift
+
   windows-latest-cmake:
     runs-on: windows-latest
 

.github/workflows/gguf-publish.yml

@@ -36,8 +36,9 @@ jobs:
         poetry install
 
     - name: Build package
-      run: poetry build
+      run: cd gguf-py && poetry build
     - name: Publish package
       uses: pypa/gh-action-pypi-publish@release/v1
       with:
         password: ${{ secrets.PYPI_API_TOKEN }}
+        packages-dir: gguf-py/dist

.github/workflows/zig-build.yml

@@ -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

.gitignore

@@ -44,6 +44,7 @@ models-mnt
 /infill
 /libllama.so
 /llama-bench
+/llava
 /main
 /metal
 /perplexity
@@ -55,6 +56,7 @@ models-mnt
 /server
 /simple
 /batched
+/batched-bench
 /export-lora
 /finetune
 /speculative

CMakeLists.txt

@@ -422,8 +422,7 @@ endif()
 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(c_flags -Wshadow -Wstrict-prototypes -Wpointer-arith -Wmissing-prototypes -Werror=implicit-int -Werror=implicit-function-declaration)
         set(cxx_flags -Wmissing-declarations -Wmissing-noreturn)
         set(host_cxx_flags "")
 
@@ -455,7 +454,8 @@ if (LLAMA_ALL_WARNINGS)
     set(c_flags   ${c_flags}   ${warning_flags})
     set(cxx_flags ${cxx_flags} ${warning_flags})
     add_compile_options("$<$<COMPILE_LANGUAGE:C>:${c_flags}>"
-                        "$<$<COMPILE_LANGUAGE:CXX>:${cxx_flags} ${host_cxx_flags}>")
+                        "$<$<COMPILE_LANGUAGE:CXX>:${cxx_flags}>"
+                        "$<$<COMPILE_LANGUAGE:CXX>:${host_cxx_flags}>")
 
 endif()
 
@@ -663,6 +663,8 @@ add_library(ggml OBJECT
             ggml.h
             ggml-alloc.c
             ggml-alloc.h
+            ggml-backend.c
+            ggml-backend.h
             ${GGML_SOURCES_CUDA} ${GGML_HEADERS_CUDA}
             ${GGML_SOURCES_OPENCL} ${GGML_HEADERS_OPENCL}
             ${GGML_SOURCES_METAL} ${GGML_HEADERS_METAL}

Makefile

@@ -1,8 +1,14 @@
 # Define the default target now so that it is always the first target
-BUILD_TARGETS = main quantize quantize-stats perplexity embedding vdot q8dot train-text-from-scratch convert-llama2c-to-ggml simple batched save-load-state server embd-input-test gguf llama-bench baby-llama beam-search speculative infill benchmark-matmult parallel finetune export-lora tests/test-c.o
+BUILD_TARGETS = \
+	main quantize quantize-stats perplexity embedding vdot q8dot train-text-from-scratch convert-llama2c-to-ggml \
+	simple batched batched-bench save-load-state server embd-input-test gguf llama-bench llava baby-llama beam-search  \
+	speculative infill benchmark-matmult parallel finetune export-lora tests/test-c.o
 
 # Binaries only useful for tests
-TEST_TARGETS = tests/test-llama-grammar tests/test-grammar-parser tests/test-double-float tests/test-grad0 tests/test-opt tests/test-quantize-fns tests/test-quantize-perf tests/test-sampling tests/test-tokenizer-0-llama tests/test-tokenizer-0-falcon tests/test-tokenizer-1-llama tests/test-tokenizer-1-bpe
+TEST_TARGETS = \
+	tests/test-llama-grammar tests/test-grammar-parser tests/test-double-float tests/test-grad0 tests/test-opt \
+	tests/test-quantize-fns tests/test-quantize-perf tests/test-sampling tests/test-tokenizer-0-llama          \
+	tests/test-tokenizer-0-falcon tests/test-tokenizer-1-llama tests/test-tokenizer-1-bpe
 
 # Code coverage output files
 COV_TARGETS = *.gcno tests/*.gcno *.gcda tests/*.gcda *.gcov tests/*.gcov lcov-report gcovr-report
@@ -172,6 +178,24 @@ else
 	MK_CPPFLAGS += -DNDEBUG
 endif
 
+ifdef LLAMA_SANITIZE_THREAD
+	MK_CFLAGS   += -fsanitize=thread -g
+	MK_CXXFLAGS += -fsanitize=thread -g
+	MK_LDFLAGS  += -fsanitize=thread -g
+endif
+
+ifdef LLAMA_SANITIZE_ADDRESS
+	MK_CFLAGS   += -fsanitize=address -fno-omit-frame-pointer -g
+	MK_CXXFLAGS += -fsanitize=address -fno-omit-frame-pointer -g
+	MK_LDFLAGS  += -fsanitize=address -fno-omit-frame-pointer -g
+endif
+
+ifdef LLAMA_SANITIZE_UNDEFINED
+	MK_CFLAGS   += -fsanitize=undefined -g
+	MK_CXXFLAGS += -fsanitize=undefined -g
+	MK_LDFLAGS  += -fsanitize=undefined -g
+endif
+
 ifdef LLAMA_SERVER_VERBOSE
 	MK_CPPFLAGS += -DSERVER_VERBOSE=$(LLAMA_SERVER_VERBOSE)
 endif
@@ -512,12 +536,21 @@ ggml.o: ggml.c ggml.h ggml-cuda.h
 ggml-alloc.o: ggml-alloc.c ggml.h ggml-alloc.h
 	$(CC)  $(CFLAGS)   -c $< -o $@
 
-OBJS += ggml-alloc.o
+ggml-backend.o: ggml-backend.c ggml.h ggml-backend.h
+	$(CC)  $(CFLAGS)   -c $< -o $@
 
-llama.o: llama.cpp ggml.h ggml-alloc.h ggml-cuda.h ggml-metal.h llama.h
+OBJS += ggml-alloc.o ggml-backend.o
+
+llama.o: llama.cpp ggml.h ggml-alloc.h ggml-backend.h ggml-cuda.h ggml-metal.h llama.h
 	$(CXX) $(CXXFLAGS) -c $< -o $@
 
-common.o: common/common.cpp common/common.h build-info.h common/log.h
+COMMON_H_DEPS = common/common.h common/sampling.h build-info.h common/log.h
+COMMON_DEPS   = $(COMMON_H_DEPS) common.o sampling.o
+
+common.o: common/common.cpp $(COMMON_H_DEPS)
+	$(CXX) $(CXXFLAGS) -c $< -o $@
+
+sampling.o: common/sampling.cpp $(COMMON_H_DEPS)
 	$(CXX) $(CXXFLAGS) -c $< -o $@
 
 console.o: common/console.cpp common/console.h
@@ -539,19 +572,22 @@ clean:
 # Examples
 #
 
-main: examples/main/main.cpp                                  build-info.h ggml.o llama.o common.o console.o grammar-parser.o $(OBJS)
+main: examples/main/main.cpp                                  build-info.h ggml.o llama.o $(COMMON_DEPS) console.o grammar-parser.o $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 	@echo
 	@echo '====  Run ./main -h for help.  ===='
 	@echo
 
-infill: examples/infill/infill.cpp                            build-info.h ggml.o llama.o common.o console.o grammar-parser.o $(OBJS)
+infill: examples/infill/infill.cpp                            build-info.h ggml.o llama.o $(COMMON_DEPS) console.o grammar-parser.o $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-simple: examples/simple/simple.cpp                            build-info.h ggml.o llama.o common.o $(OBJS)
+simple: examples/simple/simple.cpp                            build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-batched: examples/batched/batched.cpp                         build-info.h ggml.o llama.o common.o $(OBJS)
+batched: examples/batched/batched.cpp                         build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
+	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
+
+batched-bench: examples/batched-bench/batched-bench.cpp       build-info.h ggml.o llama.o common.o $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
 quantize: examples/quantize/quantize.cpp                      build-info.h ggml.o llama.o $(OBJS)
@@ -560,53 +596,56 @@ quantize: examples/quantize/quantize.cpp                      build-info.h ggml.
 quantize-stats: examples/quantize-stats/quantize-stats.cpp    build-info.h ggml.o llama.o $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-perplexity: examples/perplexity/perplexity.cpp                build-info.h ggml.o llama.o common.o $(OBJS)
+perplexity: examples/perplexity/perplexity.cpp                build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-embedding: examples/embedding/embedding.cpp                   build-info.h ggml.o llama.o common.o $(OBJS)
+embedding: examples/embedding/embedding.cpp                   build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-save-load-state: examples/save-load-state/save-load-state.cpp build-info.h ggml.o llama.o common.o $(OBJS)
+save-load-state: examples/save-load-state/save-load-state.cpp build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-server: examples/server/server.cpp examples/server/httplib.h examples/server/json.hpp examples/server/index.html.hpp examples/server/index.js.hpp examples/server/completion.js.hpp build-info.h ggml.o llama.o common.o grammar-parser.o $(OBJS)
+server: examples/server/server.cpp examples/server/httplib.h examples/server/json.hpp examples/server/index.html.hpp examples/server/index.js.hpp examples/server/completion.js.hpp build-info.h ggml.o llama.o $(COMMON_DEPS) grammar-parser.o $(OBJS)
 	$(CXX) $(CXXFLAGS) -Iexamples/server $(filter-out %.h,$(filter-out %.hpp,$^)) -o $@ $(LDFLAGS) $(LWINSOCK2)
 
-$(LIB_PRE)embdinput$(DSO_EXT): examples/embd-input/embd-input.h examples/embd-input/embd-input-lib.cpp build-info.h ggml.o llama.o common.o $(OBJS)
+$(LIB_PRE)embdinput$(DSO_EXT): examples/embd-input/embd-input.h examples/embd-input/embd-input-lib.cpp build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
 	$(CXX) --shared $(CXXFLAGS) $(filter-out %.h,$(filter-out %.hpp,$^)) -o $@ $(LDFLAGS)
 
 
-embd-input-test: $(LIB_PRE)embdinput$(DSO_EXT) examples/embd-input/embd-input-test.cpp build-info.h ggml.o llama.o common.o $(OBJS)
+embd-input-test: $(LIB_PRE)embdinput$(DSO_EXT) examples/embd-input/embd-input-test.cpp build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %$(DSO_EXT),$(filter-out %.h,$(filter-out %.hpp,$^))) -o $@ $(LDFLAGS) -L. -lembdinput
 
 gguf: examples/gguf/gguf.cpp ggml.o llama.o $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-train-text-from-scratch: examples/train-text-from-scratch/train-text-from-scratch.cpp ggml.o llama.o common.o train.o $(OBJS)
+train-text-from-scratch: examples/train-text-from-scratch/train-text-from-scratch.cpp ggml.o llama.o $(COMMON_DEPS) train.o $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
 convert-llama2c-to-ggml: examples/convert-llama2c-to-ggml/convert-llama2c-to-ggml.cpp ggml.o llama.o $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-llama-bench: examples/llama-bench/llama-bench.cpp build-info.h ggml.o llama.o common.o $(OBJS)
+llama-bench: examples/llama-bench/llama-bench.cpp build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-baby-llama: examples/baby-llama/baby-llama.cpp ggml.o llama.o common.o train.o $(OBJS)
+llava: examples/llava/llava.cpp examples/llava/llava-utils.h examples/llava/clip.cpp examples/llava/clip.h common/stb_image.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
+	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS) -Wno-cast-qual
+
+baby-llama: examples/baby-llama/baby-llama.cpp ggml.o llama.o $(COMMON_DEPS) train.o $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-beam-search: examples/beam-search/beam-search.cpp build-info.h ggml.o llama.o common.o $(OBJS)
+beam-search: examples/beam-search/beam-search.cpp build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-finetune: examples/finetune/finetune.cpp build-info.h ggml.o llama.o common.o train.o $(OBJS)
+finetune: examples/finetune/finetune.cpp build-info.h ggml.o llama.o $(COMMON_DEPS) train.o $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-export-lora: examples/export-lora/export-lora.cpp build-info.h ggml.o llama.o common.o $(OBJS)
+export-lora: examples/export-lora/export-lora.cpp build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-speculative: examples/speculative/speculative.cpp build-info.h ggml.o llama.o common.o grammar-parser.o $(OBJS)
+speculative: examples/speculative/speculative.cpp build-info.h ggml.o llama.o $(COMMON_DEPS) grammar-parser.o $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-parallel: examples/parallel/parallel.cpp build-info.h ggml.o llama.o common.o $(OBJS)
+parallel: examples/parallel/parallel.cpp build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
 ifdef LLAMA_METAL
@@ -614,6 +653,11 @@ metal: examples/metal/metal.cpp ggml.o $(OBJS)
 	$(CXX) $(CXXFLAGS) $^ -o $@ $(LDFLAGS)
 endif
 
+ifeq ($(UNAME_S),Darwin)
+swift: examples/batched.swift
+	(cd examples/batched.swift; make build)
+endif
+
 build-info.h: $(wildcard .git/index) scripts/build-info.sh
 	@sh scripts/build-info.sh $(CC) > $@.tmp
 	@if ! cmp -s $@.tmp $@; then \
@@ -634,7 +678,7 @@ benchmark-matmult: examples/benchmark/benchmark-matmult.cpp build-info.h ggml.o
 run-benchmark-matmult: benchmark-matmult
 	./$@
 
-.PHONY: run-benchmark-matmult
+.PHONY: run-benchmark-matmult swift
 
 vdot: pocs/vdot/vdot.cpp ggml.o $(OBJS)
 	$(CXX) $(CXXFLAGS) $^ -o $@ $(LDFLAGS)
@@ -642,40 +686,40 @@ vdot: pocs/vdot/vdot.cpp ggml.o $(OBJS)
 q8dot: pocs/vdot/q8dot.cpp ggml.o $(OBJS)
 	$(CXX) $(CXXFLAGS) $^ -o $@ $(LDFLAGS)
 
-tests/test-llama-grammar: tests/test-llama-grammar.cpp build-info.h ggml.o common.o grammar-parser.o $(OBJS)
+tests/test-llama-grammar: tests/test-llama-grammar.cpp build-info.h ggml.o $(COMMON_DEPS) grammar-parser.o $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-tests/test-grammar-parser: tests/test-grammar-parser.cpp build-info.h ggml.o llama.o common.o grammar-parser.o $(OBJS)
+tests/test-grammar-parser: tests/test-grammar-parser.cpp build-info.h ggml.o llama.o $(COMMON_DEPS) grammar-parser.o $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-tests/test-double-float: tests/test-double-float.cpp build-info.h ggml.o llama.o common.o $(OBJS)
+tests/test-double-float: tests/test-double-float.cpp build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-tests/test-grad0: tests/test-grad0.cpp build-info.h ggml.o llama.o common.o $(OBJS)
+tests/test-grad0: tests/test-grad0.cpp build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-tests/test-opt: tests/test-opt.cpp build-info.h ggml.o llama.o common.o $(OBJS)
+tests/test-opt: tests/test-opt.cpp build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-tests/test-quantize-fns: tests/test-quantize-fns.cpp build-info.h ggml.o llama.o common.o $(OBJS)
+tests/test-quantize-fns: tests/test-quantize-fns.cpp build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-tests/test-quantize-perf: tests/test-quantize-perf.cpp build-info.h ggml.o llama.o common.o $(OBJS)
+tests/test-quantize-perf: tests/test-quantize-perf.cpp build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-tests/test-sampling: tests/test-sampling.cpp build-info.h ggml.o llama.o common.o $(OBJS)
+tests/test-sampling: tests/test-sampling.cpp build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-tests/test-tokenizer-0-falcon: tests/test-tokenizer-0-falcon.cpp build-info.h ggml.o llama.o common.o $(OBJS)
+tests/test-tokenizer-0-falcon: tests/test-tokenizer-0-falcon.cpp build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-tests/test-tokenizer-0-llama: tests/test-tokenizer-0-llama.cpp build-info.h ggml.o llama.o common.o $(OBJS)
+tests/test-tokenizer-0-llama: tests/test-tokenizer-0-llama.cpp build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-tests/test-tokenizer-1-bpe: tests/test-tokenizer-1-bpe.cpp build-info.h ggml.o llama.o common.o $(OBJS)
+tests/test-tokenizer-1-bpe: tests/test-tokenizer-1-bpe.cpp build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
-tests/test-tokenizer-1-llama: tests/test-tokenizer-1-llama.cpp build-info.h ggml.o llama.o common.o $(OBJS)
+tests/test-tokenizer-1-llama: tests/test-tokenizer-1-llama.cpp build-info.h ggml.o llama.o $(COMMON_DEPS) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
 tests/test-c.o: tests/test-c.c llama.h

Package.swift

@@ -1,10 +1,10 @@
-// swift-tools-version:5.3
+// swift-tools-version:5.5
 
 import PackageDescription
 
 #if arch(arm) || arch(arm64)
 let platforms: [SupportedPlatform]? = [
-    .macOS(.v11),
+    .macOS(.v12),
     .iOS(.v14),
     .watchOS(.v4),
     .tvOS(.v14)
@@ -41,12 +41,13 @@ let package = Package(
                 "ggml.c",
                 "llama.cpp",
                 "ggml-alloc.c",
+                "ggml-backend.c",
                 "k_quants.c",
             ] + additionalSources,
             resources: resources,
             publicHeadersPath: "spm-headers",
             cSettings: [
-                .unsafeFlags(["-Wno-shorten-64-to-32"]),
+                .unsafeFlags(["-Wno-shorten-64-to-32", "-O3", "-DNDEBUG"]),
                 .define("GGML_USE_K_QUANTS"),
                 .define("GGML_USE_ACCELERATE")
                 // NOTE: NEW_LAPACK will required iOS version 16.4+

README.md

@@ -96,6 +96,8 @@ as the main playground for developing new features for the [ggml](https://github
 - [X] [Starcoder models](https://github.com/ggerganov/llama.cpp/pull/3187)
 - [X] [Mistral AI v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1)
 - [X] [Refact](https://huggingface.co/smallcloudai/Refact-1_6B-fim)
+- [X] [Bloom](https://github.com/ggerganov/llama.cpp/pull/3553)
+- [X] [MPT](https://github.com/ggerganov/llama.cpp/pull/3417)
 
 **Bindings:**
 
@@ -277,7 +279,7 @@ In order to build llama.cpp you have three different options.
 On MacOS, Metal is enabled by default. Using Metal makes the computation run on the GPU.
 To disable the Metal build at compile time use the `LLAMA_NO_METAL=1` flag or the `LLAMA_METAL=OFF` cmake option.
 
-When built with Metal support, you can explicitly disable GPU inference with the `--gpu-layers|-ngl 0` command-line
+When built with Metal support, you can explicitly disable GPU inference with the `--n-gpu-layers|-ngl 0` command-line
 argument.
 
 ### MPI Build

build.zig

@@ -36,14 +36,17 @@ const Maker = struct {
     }
 
     fn init(builder: *std.build.Builder) !Maker {
-        // const commit_hash = @embedFile(".git/refs/heads/master");
         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" } },
+        );
         const config_header = builder.addConfigHeader(
             .{ .style = .blank, .include_path = "build-info.h" },
             .{
                 .BUILD_NUMBER = 0,
-                .BUILD_COMMIT = "12345", // omit newline
-                .BUILD_COMPILER = "Zig 0.11.0",
+                .BUILD_COMMIT = commit_hash.stdout[0 .. commit_hash.stdout.len - 1], // omit newline
+                .BUILD_COMPILER = builder.fmt("Zig {s}", .{zig_version}),
                 .BUILD_TARGET = try target.allocDescription(builder.allocator),
             },
         );
@@ -67,13 +70,21 @@ const Maker = struct {
 
     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);
+        o.addConfigHeader(m.config_header);
         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();
             }
+        }
         o.addConfigHeader(m.config_header);
         for (m.include_dirs.items) |i| o.addIncludePath(.{ .path = i });
         o.want_lto = m.enable_lto;
@@ -86,8 +97,14 @@ const Maker = struct {
         for (deps) |d| e.addObject(d);
         for (m.objs.items) |o| e.addObject(o);
         for (m.include_dirs.items) |i| e.addIncludePath(.{ .path = i });
-        e.linkLibC();
+
+        // 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();
+        }
         e.addConfigHeader(m.config_header);
         m.builder.installArtifact(e);
         e.want_lto = m.enable_lto;
@@ -107,18 +124,22 @@ pub fn build(b: *std.build.Builder) !void {
 
     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 llama = make.obj("llama", "llama.cpp");
     const common = make.obj("common", "common/common.cpp");
-    const console = make.obj("common", "common/console.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");
 
-    _ = make.exe("main", "examples/main/main.cpp", &.{ ggml, ggml_alloc, llama, common, console, grammar_parser });
-    _ = make.exe("quantize", "examples/quantize/quantize.cpp", &.{ ggml, ggml_alloc, llama, common });
-    _ = make.exe("perplexity", "examples/perplexity/perplexity.cpp", &.{ ggml, ggml_alloc, llama, common });
-    _ = make.exe("embedding", "examples/embedding/embedding.cpp", &.{ ggml, ggml_alloc, llama, common });
-    _ = make.exe("train-text-from-scratch", "examples/train-text-from-scratch/train-text-from-scratch.cpp", &.{ ggml, ggml_alloc, llama, common });
+    _ = make.exe("main", "examples/main/main.cpp", &.{ ggml, ggml_alloc, ggml_backend, llama, common, sampling, console, grammar_parser });
+    _ = make.exe("quantize", "examples/quantize/quantize.cpp", &.{ ggml, ggml_alloc, ggml_backend, llama, common });
+    _ = make.exe("perplexity", "examples/perplexity/perplexity.cpp", &.{ ggml, ggml_alloc, ggml_backend, llama, common });
+    _ = make.exe("embedding", "examples/embedding/embedding.cpp", &.{ ggml, ggml_alloc, ggml_backend, llama, common });
+    _ = make.exe("finetune", "examples/finetune/finetune.cpp", &.{ ggml, ggml_alloc, ggml_backend, llama, common, train });
+    _ = make.exe("train-text-from-scratch", "examples/train-text-from-scratch/train-text-from-scratch.cpp", &.{ ggml, ggml_alloc, ggml_backend, llama, common, train });
 
-    const server = make.exe("server", "examples/server/server.cpp", &.{ ggml, ggml_alloc, llama, common, grammar_parser });
+    const server = make.exe("server", "examples/server/server.cpp", &.{ ggml, ggml_alloc, ggml_backend, llama, common, sampling, grammar_parser });
     if (server.target.isWindows()) {
         server.linkSystemLibrary("ws2_32");
     }

ci/run.sh

@@ -496,11 +496,13 @@ test $ret -eq 0 && gg_run ctest_debug
 test $ret -eq 0 && gg_run ctest_release
 
 if [ -z ${GG_BUILD_LOW_PERF} ]; then
+    if [ -z ${GG_BUILD_VRAM_GB} ] || [ ${GG_BUILD_VRAM_GB} -ge 8 ]; then
         if [ -z ${GG_BUILD_CUDA} ]; then
             test $ret -eq 0 && gg_run open_llama_3b_v2
         else
             test $ret -eq 0 && gg_run open_llama_7b_v2
         fi
     fi
+fi
 
 exit $ret

common/CMakeLists.txt

@@ -5,6 +5,8 @@ set(TARGET common)
 add_library(${TARGET} OBJECT
     common.h
     common.cpp
+    sampling.h
+    sampling.cpp
     console.h
     console.cpp
     grammar-parser.h

common/common.cpp

@@ -107,6 +107,7 @@ bool gpt_params_parse(int argc, char ** argv, gpt_params & params) {
     std::string arg;
     gpt_params default_params;
     const std::string arg_prefix = "--";
+    llama_sampling_params & sparams = params.sampling_params;
 
     for (int i = 1; i < argc; i++) {
         arg = argv[i];
@@ -170,7 +171,7 @@ bool gpt_params_parse(int argc, char ** argv, gpt_params & params) {
             // store the external file name in params
             params.prompt_file = argv[i];
             std::copy(std::istreambuf_iterator<char>(file), std::istreambuf_iterator<char>(), back_inserter(params.prompt));
-            if (params.prompt.back() == '\n') {
+            if (!params.prompt.empty() && params.prompt.back() == '\n') {
                 params.prompt.pop_back();
             }
         } else if (arg == "-n" || arg == "--n-predict") {
@@ -184,7 +185,7 @@ bool gpt_params_parse(int argc, char ** argv, gpt_params & params) {
                 invalid_param = true;
                 break;
             }
-            params.top_k = std::stoi(argv[i]);
+            sparams.top_k = std::stoi(argv[i]);
         } else if (arg == "-c" || arg == "--ctx-size") {
             if (++i >= argc) {
                 invalid_param = true;
@@ -250,73 +251,73 @@ bool gpt_params_parse(int argc, char ** argv, gpt_params & params) {
                 invalid_param = true;
                 break;
             }
-            params.top_p = std::stof(argv[i]);
+            sparams.top_p = std::stof(argv[i]);
         } else if (arg == "--temp") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            params.temp = std::stof(argv[i]);
+            sparams.temp = std::stof(argv[i]);
         } else if (arg == "--tfs") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            params.tfs_z = std::stof(argv[i]);
+            sparams.tfs_z = std::stof(argv[i]);
         } else if (arg == "--typical") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            params.typical_p = std::stof(argv[i]);
+            sparams.typical_p = std::stof(argv[i]);
         } else if (arg == "--repeat-last-n") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            params.repeat_last_n = std::stoi(argv[i]);
+            sparams.repeat_last_n = std::stoi(argv[i]);
         } else if (arg == "--repeat-penalty") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            params.repeat_penalty = std::stof(argv[i]);
+            sparams.repeat_penalty = std::stof(argv[i]);
         } else if (arg == "--frequency-penalty") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            params.frequency_penalty = std::stof(argv[i]);
+            sparams.frequency_penalty = std::stof(argv[i]);
         } else if (arg == "--presence-penalty") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            params.presence_penalty = std::stof(argv[i]);
+            sparams.presence_penalty = std::stof(argv[i]);
         } else if (arg == "--mirostat") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            params.mirostat = std::stoi(argv[i]);
+            sparams.mirostat = std::stoi(argv[i]);
         } else if (arg == "--mirostat-lr") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            params.mirostat_eta = std::stof(argv[i]);
+            sparams.mirostat_eta = std::stof(argv[i]);
         } else if (arg == "--mirostat-ent") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            params.mirostat_tau = std::stof(argv[i]);
+            sparams.mirostat_tau = std::stof(argv[i]);
         } else if (arg == "--cfg-negative-prompt") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            params.cfg_negative_prompt = argv[i];
+            sparams.cfg_negative_prompt = argv[i];
         } else if (arg == "--cfg-negative-prompt-file") {
             if (++i >= argc) {
                 invalid_param = true;
@@ -328,16 +329,16 @@ bool gpt_params_parse(int argc, char ** argv, gpt_params & params) {
                 invalid_param = true;
                 break;
             }
-            std::copy(std::istreambuf_iterator<char>(file), std::istreambuf_iterator<char>(), back_inserter(params.cfg_negative_prompt));
-            if (params.cfg_negative_prompt.back() == '\n') {
-                params.cfg_negative_prompt.pop_back();
+            std::copy(std::istreambuf_iterator<char>(file), std::istreambuf_iterator<char>(), back_inserter(sparams.cfg_negative_prompt));
+            if (!sparams.cfg_negative_prompt.empty() && sparams.cfg_negative_prompt.back() == '\n') {
+                sparams.cfg_negative_prompt.pop_back();
             }
         } else if (arg == "--cfg-scale") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            params.cfg_scale = std::stof(argv[i]);
+            sparams.cfg_scale = std::stof(argv[i]);
         } else if (arg == "-b" || arg == "--batch-size") {
             if (++i >= argc) {
                 invalid_param = true;
@@ -417,6 +418,18 @@ bool gpt_params_parse(int argc, char ** argv, gpt_params & params) {
                 break;
             }
             params.lora_base = argv[i];
+        } else if (arg == "--mmproj") {
+            if (++i >= argc) {
+                invalid_param = true;
+                break;
+            }
+            params.mmproj = argv[i];
+        } else if (arg == "--image") {
+            if (++i >= argc) {
+                invalid_param = true;
+                break;
+            }
+            params.image = argv[i];
         } else if (arg == "-i" || arg == "--interactive") {
             params.interactive = true;
         } else if (arg == "--embedding") {
@@ -546,7 +559,7 @@ bool gpt_params_parse(int argc, char ** argv, gpt_params & params) {
         } else if (arg == "--ignore-eos") {
             params.ignore_eos = true;
         } else if (arg == "--no-penalize-nl") {
-            params.penalize_nl = false;
+            sparams.penalize_nl = false;
         } else if (arg == "-l" || arg == "--logit-bias") {
             if (++i >= argc) {
                 invalid_param = true;
@@ -558,7 +571,7 @@ bool gpt_params_parse(int argc, char ** argv, gpt_params & params) {
             std::string value_str;
             try {
                 if (ss >> key && ss >> sign && std::getline(ss, value_str) && (sign == '+' || sign == '-')) {
-                    params.logit_bias[key] = std::stof(value_str) * ((sign == '-') ? -1.0f : 1.0f);
+                    sparams.logit_bias[key] = std::stof(value_str) * ((sign == '-') ? -1.0f : 1.0f);
                 } else {
                     throw std::exception();
                 }
@@ -661,6 +674,8 @@ bool gpt_params_parse(int argc, char ** argv, gpt_params & params) {
 }
 
 void gpt_print_usage(int /*argc*/, char ** argv, const gpt_params & params) {
+    const llama_sampling_params & sparams = params.sampling_params;
+
     printf("usage: %s [options]\n", argv[0]);
     printf("\n");
     printf("options:\n");
@@ -693,19 +708,19 @@ void gpt_print_usage(int /*argc*/, char ** argv, const gpt_params & params) {
     printf("  -n N, --n-predict N   number of tokens to predict (default: %d, -1 = infinity, -2 = until context filled)\n", params.n_predict);
     printf("  -c N, --ctx-size N    size of the prompt context (default: %d, 0 = loaded from model)\n", params.n_ctx);
     printf("  -b N, --batch-size N  batch size for prompt processing (default: %d)\n", params.n_batch);
-    printf("  --top-k N             top-k sampling (default: %d, 0 = disabled)\n", params.top_k);
-    printf("  --top-p N             top-p sampling (default: %.1f, 1.0 = disabled)\n", (double)params.top_p);
-    printf("  --tfs N               tail free sampling, parameter z (default: %.1f, 1.0 = disabled)\n", (double)params.tfs_z);
-    printf("  --typical N           locally typical sampling, parameter p (default: %.1f, 1.0 = disabled)\n", (double)params.typical_p);
-    printf("  --repeat-last-n N     last n tokens to consider for penalize (default: %d, 0 = disabled, -1 = ctx_size)\n", params.repeat_last_n);
-    printf("  --repeat-penalty N    penalize repeat sequence of tokens (default: %.1f, 1.0 = disabled)\n", (double)params.repeat_penalty);
-    printf("  --presence-penalty N  repeat alpha presence penalty (default: %.1f, 0.0 = disabled)\n", (double)params.presence_penalty);
-    printf("  --frequency-penalty N repeat alpha frequency penalty (default: %.1f, 0.0 = disabled)\n", (double)params.frequency_penalty);
+    printf("  --top-k N             top-k sampling (default: %d, 0 = disabled)\n", sparams.top_k);
+    printf("  --top-p N             top-p sampling (default: %.1f, 1.0 = disabled)\n", (double)sparams.top_p);
+    printf("  --tfs N               tail free sampling, parameter z (default: %.1f, 1.0 = disabled)\n", (double)sparams.tfs_z);
+    printf("  --typical N           locally typical sampling, parameter p (default: %.1f, 1.0 = disabled)\n", (double)sparams.typical_p);
+    printf("  --repeat-last-n N     last n tokens to consider for penalize (default: %d, 0 = disabled, -1 = ctx_size)\n", sparams.repeat_last_n);
+    printf("  --repeat-penalty N    penalize repeat sequence of tokens (default: %.1f, 1.0 = disabled)\n", (double)sparams.repeat_penalty);
+    printf("  --presence-penalty N  repeat alpha presence penalty (default: %.1f, 0.0 = disabled)\n", (double)sparams.presence_penalty);
+    printf("  --frequency-penalty N repeat alpha frequency penalty (default: %.1f, 0.0 = disabled)\n", (double)sparams.frequency_penalty);
     printf("  --mirostat N          use Mirostat sampling.\n");
     printf("                        Top K, Nucleus, Tail Free and Locally Typical samplers are ignored if used.\n");
-    printf("                        (default: %d, 0 = disabled, 1 = Mirostat, 2 = Mirostat 2.0)\n", params.mirostat);
-    printf("  --mirostat-lr N       Mirostat learning rate, parameter eta (default: %.1f)\n", (double)params.mirostat_eta);
-    printf("  --mirostat-ent N      Mirostat target entropy, parameter tau (default: %.1f)\n", (double)params.mirostat_tau);
+    printf("                        (default: %d, 0 = disabled, 1 = Mirostat, 2 = Mirostat 2.0)\n", sparams.mirostat);
+    printf("  --mirostat-lr N       Mirostat learning rate, parameter eta (default: %.1f)\n", (double)sparams.mirostat_eta);
+    printf("  --mirostat-ent N      Mirostat target entropy, parameter tau (default: %.1f)\n", (double)sparams.mirostat_tau);
     printf("  -l TOKEN_ID(+/-)BIAS, --logit-bias TOKEN_ID(+/-)BIAS\n");
     printf("                        modifies the likelihood of token appearing in the completion,\n");
     printf("                        i.e. `--logit-bias 15043+1` to increase likelihood of token ' Hello',\n");
@@ -716,7 +731,7 @@ void gpt_print_usage(int /*argc*/, char ** argv, const gpt_params & params) {
     printf("                        negative prompt to use for guidance. (default: empty)\n");
     printf("  --cfg-negative-prompt-file FNAME\n");
     printf("                        negative prompt file to use for guidance. (default: empty)\n");
-    printf("  --cfg-scale N         strength of guidance (default: %f, 1.0 = disable)\n", params.cfg_scale);
+    printf("  --cfg-scale N         strength of guidance (default: %f, 1.0 = disable)\n", sparams.cfg_scale);
     printf("  --rope-scaling {none,linear,yarn}\n");
     printf("                        RoPE frequency scaling method, defaults to linear unless specified by the model\n");
     printf("  --rope-scale N        RoPE context scaling factor, expands context by a factor of N\n");
@@ -730,7 +745,7 @@ void gpt_print_usage(int /*argc*/, char ** argv, const gpt_params & params) {
     printf("  --no-penalize-nl      do not penalize newline token\n");
     printf("  --memory-f32          use f32 instead of f16 for memory key+value (default: disabled)\n");
     printf("                        not recommended: doubles context memory required and no measurable increase in quality\n");
-    printf("  --temp N              temperature (default: %.1f)\n", (double)params.temp);
+    printf("  --temp N              temperature (default: %.1f)\n", (double)sparams.temp);
     printf("  --logits-all          return logits for all tokens in the batch (default: disabled)\n");
     printf("  --hellaswag           compute HellaSwag score over random tasks from datafile supplied with -f\n");
     printf("  --hellaswag-tasks N   number of tasks to use when computing the HellaSwag score (default: %zu)\n", params.hellaswag_tasks);
@@ -740,6 +755,8 @@ void gpt_print_usage(int /*argc*/, char ** argv, const gpt_params & params) {
     printf("  -np N, --parallel N   number of parallel sequences to decode (default: %d)\n", params.n_parallel);
     printf("  -ns N, --sequences N  number of sequences to decode (default: %d)\n", params.n_sequences);
     printf("  -cb, --cont-batching  enable continuous batching (a.k.a dynamic batching) (default: disabled)\n");
+    printf("  --mmproj MMPROJ_FILE  path to a multimodal projector file for LLaVA. see examples/llava/README.md\n");
+    printf("  --image IMAGE_FILE    path to an image file. use with multimodal models\n");
     if (llama_mlock_supported()) {
         printf("  --mlock               force system to keep model in RAM rather than swapping or compressing\n");
     }
@@ -885,7 +902,7 @@ std::tuple<struct llama_model *, struct llama_context *> llama_init_from_gpt_par
     }
 
     if (params.ignore_eos) {
-        params.logit_bias[llama_token_eos(lctx)] = -INFINITY;
+        params.sampling_params.logit_bias[llama_token_eos(lctx)] = -INFINITY;
     }
 
     {
@@ -977,127 +994,6 @@ std::string llama_detokenize_bpe(llama_context * ctx, const std::vector<llama_to
     return result;
 }
 
-//
-// Sampling utils
-//
-
-llama_token llama_sample_token(
-                  struct llama_context * ctx,
-                  struct llama_context * ctx_guidance,
-                  struct llama_grammar * grammar,
-               const struct gpt_params & params,
-        const std::vector<llama_token> & last_tokens,
-         std::vector<llama_token_data> & candidates,
-                                   int   idx) {
-    const int n_ctx   = llama_n_ctx(ctx);
-    const int n_vocab = llama_n_vocab(llama_get_model(ctx));
-
-    const float   temp            = params.temp;
-    const int32_t top_k           = params.top_k <= 0 ? n_vocab : params.top_k;
-    const float   top_p           = params.top_p;
-    const float   tfs_z           = params.tfs_z;
-    const float   typical_p       = params.typical_p;
-    const int32_t repeat_last_n   = params.repeat_last_n < 0 ? n_ctx : params.repeat_last_n;
-    const float   repeat_penalty  = params.repeat_penalty;
-    const float   alpha_presence  = params.presence_penalty;
-    const float   alpha_frequency = params.frequency_penalty;
-    const int     mirostat        = params.mirostat;
-    const float   mirostat_tau    = params.mirostat_tau;
-    const float   mirostat_eta    = params.mirostat_eta;
-    const bool    penalize_nl     = params.penalize_nl;
-
-    llama_token id = 0;
-
-    float * logits = llama_get_logits_ith(ctx, idx);
-
-    // Apply params.logit_bias map
-    for (auto it = params.logit_bias.begin(); it != params.logit_bias.end(); it++) {
-        logits[it->first] += it->second;
-    }
-
-    candidates.clear();
-    for (llama_token token_id = 0; token_id < n_vocab; token_id++) {
-        candidates.emplace_back(llama_token_data{token_id, logits[token_id], 0.0f});
-    }
-
-    llama_token_data_array cur_p = { candidates.data(), candidates.size(), false };
-
-    if (ctx_guidance) {
-        llama_sample_classifier_free_guidance(ctx, &cur_p, ctx_guidance, params.cfg_scale);
-    }
-
-    // apply penalties
-    if (!last_tokens.empty()) {
-        const float nl_logit = logits[llama_token_nl(ctx)];
-        const int last_n_repeat = std::min(std::min((int)last_tokens.size(), repeat_last_n), n_ctx);
-
-        llama_sample_repetition_penalty(ctx, &cur_p,
-                last_tokens.data() + last_tokens.size() - last_n_repeat,
-                last_n_repeat, repeat_penalty);
-        llama_sample_frequency_and_presence_penalties(ctx, &cur_p,
-                last_tokens.data() + last_tokens.size() - last_n_repeat,
-                last_n_repeat, alpha_frequency, alpha_presence);
-
-        if (!penalize_nl) {
-            for (size_t idx = 0; idx < cur_p.size; idx++) {
-                if (cur_p.data[idx].id == llama_token_nl(ctx)) {
-                    cur_p.data[idx].logit = nl_logit;
-                    break;
-                }
-            }
-        }
-    }
-
-    if (grammar != NULL) {
-        llama_sample_grammar(ctx, &cur_p, grammar);
-    }
-
-    if (temp <= 0) {
-        // Greedy sampling
-        id = llama_sample_token_greedy(ctx, &cur_p);
-    } else {
-        if (mirostat == 1) {
-            static float mirostat_mu = 2.0f * mirostat_tau;
-            const int mirostat_m = 100;
-            llama_sample_temp(ctx, &cur_p, temp);
-            id = llama_sample_token_mirostat(ctx, &cur_p, mirostat_tau, mirostat_eta, mirostat_m, &mirostat_mu);
-        } else if (mirostat == 2) {
-            static float mirostat_mu = 2.0f * mirostat_tau;
-            llama_sample_temp(ctx, &cur_p, temp);
-            id = llama_sample_token_mirostat_v2(ctx, &cur_p, mirostat_tau, mirostat_eta, &mirostat_mu);
-        } else {
-            // Temperature sampling
-            size_t min_keep = std::max(1, params.n_probs);
-            llama_sample_top_k      (ctx, &cur_p, top_k, min_keep);
-            llama_sample_tail_free  (ctx, &cur_p, tfs_z, min_keep);
-            llama_sample_typical    (ctx, &cur_p, typical_p, min_keep);
-            llama_sample_top_p      (ctx, &cur_p, top_p, min_keep);
-            llama_sample_temp(ctx, &cur_p, temp);
-
-            {
-                const int n_top = 10;
-                LOG("top %d candidates:\n", n_top);
-
-                for (int i = 0; i < n_top; i++) {
-                    const llama_token id = cur_p.data[i].id;
-                    LOG(" - %5d: '%12s' (%.3f)\n", id, llama_token_to_piece(ctx, id).c_str(), cur_p.data[i].p);
-                }
-            }
-
-            id = llama_sample_token(ctx, &cur_p);
-
-            LOG("sampled token: %5d: '%s'\n", id, llama_token_to_piece(ctx, id).c_str());
-        }
-    }
-    // printf("`%d`", candidates_p.size);
-
-    if (grammar != NULL) {
-        llama_grammar_accept_token(ctx, grammar, id);
-    }
-
-    return id;
-}
-
 //
 // YAML utils
 //
@@ -1249,6 +1145,8 @@ std::string get_sortable_timestamp() {
 
 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) {
+    const llama_sampling_params & sparams = params.sampling_params;
+
     fprintf(stream, "build_commit: %s\n", BUILD_COMMIT);
     fprintf(stream, "build_number: %d\n", BUILD_NUMBER);
     fprintf(stream, "cpu_has_arm_fma: %s\n", ggml_cpu_has_arm_fma() ? "true" : "false");
@@ -1295,21 +1193,21 @@ void dump_non_result_info_yaml(FILE * stream, const gpt_params & params, const l
 
     fprintf(stream, "alias: %s # default: unknown\n", params.model_alias.c_str());
     fprintf(stream, "batch_size: %d # default: 512\n", params.n_batch);
-    dump_string_yaml_multiline(stream, "cfg_negative_prompt", params.cfg_negative_prompt.c_str());
-    fprintf(stream, "cfg_scale: %f # default: 1.0\n", params.cfg_scale);
+    dump_string_yaml_multiline(stream, "cfg_negative_prompt", sparams.cfg_negative_prompt.c_str());
+    fprintf(stream, "cfg_scale: %f # default: 1.0\n", sparams.cfg_scale);
     fprintf(stream, "chunks: %d # default: -1 (unlimited)\n", params.n_chunks);
     fprintf(stream, "color: %s # default: false\n", params.use_color ? "true" : "false");
     fprintf(stream, "ctx_size: %d # default: 512\n", params.n_ctx);
     fprintf(stream, "escape: %s # default: false\n", params.escape ? "true" : "false");
     fprintf(stream, "file: # never logged, see prompt instead. Can still be specified for input.\n");
-    fprintf(stream, "frequency_penalty: %f # default: 0.0 \n", params.frequency_penalty);
+    fprintf(stream, "frequency_penalty: %f # default: 0.0 \n", sparams.frequency_penalty);
     dump_string_yaml_multiline(stream, "grammar", params.grammar.c_str());
     fprintf(stream, "grammar-file: # never logged, see grammar instead. Can still be specified for input.\n");
     fprintf(stream, "hellaswag: %s # default: false\n", params.hellaswag ? "true" : "false");
     fprintf(stream, "hellaswag_tasks: %zu # default: 400\n", params.hellaswag_tasks);
 
-    const auto logit_bias_eos = params.logit_bias.find(llama_token_eos(lctx));
-    const bool ignore_eos = logit_bias_eos != params.logit_bias.end() && logit_bias_eos->second == -INFINITY;
+    const auto logit_bias_eos = sparams.logit_bias.find(llama_token_eos(lctx));
+    const bool ignore_eos = logit_bias_eos != sparams.logit_bias.end() && logit_bias_eos->second == -INFINITY;
     fprintf(stream, "ignore_eos: %s # default: false\n", ignore_eos ? "true" : "false");
 
     dump_string_yaml_multiline(stream, "in_prefix", params.input_prefix.c_str());
@@ -1322,7 +1220,7 @@ void dump_non_result_info_yaml(FILE * stream, const gpt_params & params, const l
     fprintf(stream, "logdir: %s # default: unset (no logging)\n", params.logdir.c_str());
 
     fprintf(stream, "logit_bias:\n");
-    for (std::pair<llama_token, float> lb : params.logit_bias) {
+    for (std::pair<llama_token, float> lb : sparams.logit_bias) {
         if (ignore_eos && lb.first == logit_bias_eos->first) {
             continue;
         }
@@ -1346,30 +1244,30 @@ void dump_non_result_info_yaml(FILE * stream, const gpt_params & params, const l
     fprintf(stream, "lora_base: %s\n", params.lora_base.c_str());
     fprintf(stream, "main_gpu: %d # default: 0\n", params.main_gpu);
     fprintf(stream, "memory_f32: %s # default: false\n", !params.memory_f16 ? "true" : "false");
-    fprintf(stream, "mirostat: %d # default: 0 (disabled)\n", params.mirostat);
-    fprintf(stream, "mirostat_ent: %f # default: 5.0\n", params.mirostat_tau);
-    fprintf(stream, "mirostat_lr: %f # default: 0.1\n", params.mirostat_eta);
+    fprintf(stream, "mirostat: %d # default: 0 (disabled)\n", sparams.mirostat);
+    fprintf(stream, "mirostat_ent: %f # default: 5.0\n", sparams.mirostat_tau);
+    fprintf(stream, "mirostat_lr: %f # default: 0.1\n", sparams.mirostat_eta);
     fprintf(stream, "mlock: %s # default: false\n", params.use_mlock ? "true" : "false");
     fprintf(stream, "model: %s # default: models/7B/ggml-model.bin\n", params.model.c_str());
     fprintf(stream, "model_draft: %s # default:\n", params.model_draft.c_str());
     fprintf(stream, "multiline_input: %s # default: false\n", params.multiline_input ? "true" : "false");
     fprintf(stream, "n_gpu_layers: %d # default: -1\n", params.n_gpu_layers);
     fprintf(stream, "n_predict: %d # default: -1 (unlimited)\n", params.n_predict);
-    fprintf(stream, "n_probs: %d # only used by server binary, default: 0\n", params.n_probs);
+    fprintf(stream, "n_probs: %d # only used by server binary, default: 0\n", sparams.n_probs);
     fprintf(stream, "no_mmap: %s # default: false\n", !params.use_mmap ? "true" : "false");
     fprintf(stream, "no_mul_mat_q: %s # default: false\n", !params.mul_mat_q ? "true" : "false");
-    fprintf(stream, "no_penalize_nl: %s # default: false\n", !params.penalize_nl ? "true" : "false");
+    fprintf(stream, "no_penalize_nl: %s # default: false\n", !sparams.penalize_nl ? "true" : "false");
     fprintf(stream, "numa: %s # default: false\n", params.numa ? "true" : "false");
     fprintf(stream, "ppl_output_type: %d # default: 0\n", params.ppl_output_type);
     fprintf(stream, "ppl_stride: %d # default: 0\n", params.ppl_stride);
-    fprintf(stream, "presence_penalty: %f # default: 0.0\n", params.presence_penalty);
+    fprintf(stream, "presence_penalty: %f # default: 0.0\n", sparams.presence_penalty);
     dump_string_yaml_multiline(stream, "prompt", params.prompt.c_str());
     fprintf(stream, "prompt_cache: %s\n", params.path_prompt_cache.c_str());
     fprintf(stream, "prompt_cache_all: %s # default: false\n", params.prompt_cache_all ? "true" : "false");
     fprintf(stream, "prompt_cache_ro: %s # default: false\n", params.prompt_cache_ro ? "true" : "false");
     dump_vector_int_yaml(stream, "prompt_tokens", prompt_tokens);
     fprintf(stream, "random_prompt: %s # default: false\n", params.random_prompt ? "true" : "false");
-    fprintf(stream, "repeat_penalty: %f # default: 1.1\n", params.repeat_penalty);
+    fprintf(stream, "repeat_penalty: %f # default: 1.1\n", sparams.repeat_penalty);
 
     fprintf(stream, "reverse_prompt:\n");
     for (std::string ap : params.antiprompt) {
@@ -1387,15 +1285,15 @@ void dump_non_result_info_yaml(FILE * stream, const gpt_params & params, const l
     fprintf(stream, "seed: %d # default: -1 (random seed)\n", params.seed);
     fprintf(stream, "simple_io: %s # default: false\n", params.simple_io ? "true" : "false");
     fprintf(stream, "cont_batching: %s # default: false\n", params.cont_batching ? "true" : "false");
-    fprintf(stream, "temp: %f # default: 0.8\n", params.temp);
+    fprintf(stream, "temp: %f # default: 0.8\n", sparams.temp);
 
     const std::vector<float> tensor_split_vector(params.tensor_split, params.tensor_split + LLAMA_MAX_DEVICES);
     dump_vector_float_yaml(stream, "tensor_split", tensor_split_vector);
 
-    fprintf(stream, "tfs: %f # default: 1.0\n", params.tfs_z);
+    fprintf(stream, "tfs: %f # default: 1.0\n", sparams.tfs_z);
     fprintf(stream, "threads: %d # default: %d\n", params.n_threads, std::thread::hardware_concurrency());
-    fprintf(stream, "top_k: %d # default: 40\n", params.top_k);
-    fprintf(stream, "top_p: %f # default: 0.95\n", params.top_p);
-    fprintf(stream, "typical_p: %f # default: 1.0\n", params.typical_p);
+    fprintf(stream, "top_k: %d # default: 40\n", sparams.top_k);
+    fprintf(stream, "top_p: %f # default: 0.95\n", sparams.top_p);
+    fprintf(stream, "typical_p: %f # default: 1.0\n", sparams.typical_p);
     fprintf(stream, "verbose_prompt: %s # default: false\n", params.verbose_prompt ? "true" : "false");
 }

common/common.h

@@ -4,6 +4,8 @@
 
 #include "llama.h"
 
+#include "sampling.h"
+
 #define LOG_NO_FILE_LINE_FUNCTION
 #include "log.h"
 
@@ -50,7 +52,6 @@ struct gpt_params {
     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_probs                         = 0;    // if greater than 0, output the probabilities of top n_probs tokens.
     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
@@ -60,26 +61,8 @@ struct gpt_params {
     float   yarn_beta_slow                  = 1.0f;  // YaRN high correction dim
     int8_t  rope_scaling_type               = LLAMA_ROPE_SCALING_UNSPECIFIED;
 
-    // sampling parameters
-    int32_t top_k             = 40;    // <= 0 to use vocab size
-    float   top_p             = 0.95f; // 1.0 = disabled
-    float   tfs_z             = 1.00f; // 1.0 = disabled
-    float   typical_p         = 1.00f; // 1.0 = disabled
-    float   temp              = 0.80f; // 1.0 = disabled
-    float   repeat_penalty    = 1.10f; // 1.0 = disabled
-    int32_t repeat_last_n     = 64;    // last n tokens to penalize (0 = disable penalty, -1 = context size)
-    float   frequency_penalty = 0.00f; // 0.0 = disabled
-    float   presence_penalty  = 0.00f; // 0.0 = disabled
-    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
-
-    std::unordered_map<llama_token, float> logit_bias; // logit bias for specific tokens
-
-    // Classifier-Free Guidance
-    // https://arxiv.org/abs/2306.17806
-    std::string cfg_negative_prompt;       // string to help guidance
-    float       cfg_scale         = 1.f;   // How strong is guidance
+    // // sampling parameters
+    struct llama_sampling_params sampling_params;
 
     std::string model             = "models/7B/ggml-model-f16.gguf"; // model path
     std::string model_draft       = "";                              // draft model for speculative decoding
@@ -121,13 +104,16 @@ struct gpt_params {
     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 penalize_nl       = true;  // consider newlines as a repeatable token
     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 infill            = false; // use infill mode
+
+    // multimodal models (see examples/llava)
+    std::string mmproj = ""; // path to multimodal projector
+    std::string image = ""; // path to an image file
 };
 
 bool gpt_params_parse(int argc, char ** argv, gpt_params & params);
@@ -186,36 +172,6 @@ std::string llama_detokenize_bpe(
                          llama_context * ctx,
         const std::vector<llama_token> & tokens);
 
-//
-// Sampling utils
-//
-
-// this is a common sampling function used across the examples for convenience
-// it can serve as a starting point for implementing your own sampling function
-//
-// required:
-//  - ctx:    context to use for sampling
-//  - params: sampling parameters
-//
-// optional:
-//  - ctx_guidance:  context to use for classifier-free guidance, ignore if NULL
-//  - grammar:       grammar to use for sampling, ignore if NULL
-//  - last_tokens:   needed for repetition penalty, ignore if empty
-//  - idx:           sample from llama_get_logits_ith(ctx, idx)
-//
-// returns:
-//  - token:      sampled token
-//  - candidates: vector of candidate tokens
-//
-llama_token llama_sample_token(
-                  struct llama_context * ctx,
-                  struct llama_context * ctx_guidance,
-                  struct llama_grammar * grammar,
-               const struct gpt_params & params,
-        const std::vector<llama_token> & last_tokens,
-         std::vector<llama_token_data> & candidates,
-                                   int   idx = 0);
-
 //
 // YAML utils
 //

common/sampling.cpp

@@ -0,0 +1,166 @@
+#include "sampling.h"
+
+llama_sampling_context::~llama_sampling_context() {
+    for (auto & it : sequence_contexts) {
+        if (it.second.grammar != NULL) {
+            llama_grammar_free(it.second.grammar);
+            it.second.grammar = NULL;
+        }
+    }
+}
+
+llama_sampling_context llama_sampling_context_init(
+        const struct gpt_params & params,
+                  llama_grammar * grammar) {
+  llama_sampling_context result;
+
+  result.params = params.sampling_params;
+  result.grammar = grammar;
+  return result;
+}
+
+// Note: Creates the context if it doesn't exist, so this always return something.
+llama_sampler_sequence_context & llama_sampling_get_sequence_context(
+              llama_sampling_context & ctx_sampling,
+        const llama_seq_id             seq) {
+    const auto it = ctx_sampling.sequence_contexts.find(seq);
+    if (it != ctx_sampling.sequence_contexts.end()) {
+        return it->second;
+    }
+    llama_sampler_sequence_context new_ctx = {
+        2.0f * ctx_sampling.params.mirostat_tau,
+        ctx_sampling.grammar != NULL ? llama_grammar_copy(ctx_sampling.grammar) : NULL,
+    };
+    return ctx_sampling.sequence_contexts.insert({seq, new_ctx}).first->second;
+}
+
+bool llama_sampling_context_reset(
+              llama_sampling_context & ctx_sampling,
+        const llama_seq_id             seq) {
+    const auto it = ctx_sampling.sequence_contexts.find(seq);
+    if (it == ctx_sampling.sequence_contexts.end()) return false;
+    if (it->second.grammar != NULL) {
+        llama_grammar_free(it->second.grammar);
+        it->second.grammar = NULL;
+    }
+    ctx_sampling.sequence_contexts.erase(it);
+    return true;
+}
+
+llama_token llama_sampling_sample(
+                  struct llama_context * ctx,
+                  struct llama_context * ctx_guidance,
+                  struct llama_sampling_context & ctx_sampling,
+        const std::vector<llama_token> & last_tokens,
+         std::vector<llama_token_data> & candidates,
+        const                      int   idx,
+                          llama_seq_id   seq) {
+    const int n_ctx   = llama_n_ctx(ctx);
+    const int n_vocab = llama_n_vocab(llama_get_model(ctx));
+
+    const llama_sampling_params & params = ctx_sampling.params;
+    const float   temp            = params.temp;
+    const int32_t top_k           = params.top_k <= 0 ? n_vocab : params.top_k;
+    const float   top_p           = params.top_p;
+    const float   tfs_z           = params.tfs_z;
+    const float   typical_p       = params.typical_p;
+    const int32_t repeat_last_n   = params.repeat_last_n < 0 ? n_ctx : params.repeat_last_n;
+    const float   repeat_penalty  = params.repeat_penalty;
+    const float   alpha_presence  = params.presence_penalty;
+    const float   alpha_frequency = params.frequency_penalty;
+    const int     mirostat        = params.mirostat;
+    const float   mirostat_tau    = params.mirostat_tau;
+    const float   mirostat_eta    = params.mirostat_eta;
+    const bool    penalize_nl     = params.penalize_nl;
+
+    llama_token id = 0;
+
+    float * logits = llama_get_logits_ith(ctx, idx);
+
+    // Apply params.logit_bias map
+    for (auto it = params.logit_bias.begin(); it != params.logit_bias.end(); it++) {
+        logits[it->first] += it->second;
+    }
+
+    candidates.clear();
+    for (llama_token token_id = 0; token_id < n_vocab; token_id++) {
+        candidates.emplace_back(llama_token_data{token_id, logits[token_id], 0.0f});
+    }
+
+    llama_token_data_array cur_p = { candidates.data(), candidates.size(), false };
+
+    if (ctx_guidance) {
+        llama_sample_classifier_free_guidance(ctx, &cur_p, ctx_guidance, params.cfg_scale);
+    }
+
+    // apply penalties
+    if (!last_tokens.empty()) {
+        const float nl_logit = logits[llama_token_nl(ctx)];
+        const int last_n_repeat = std::min(std::min((int)last_tokens.size(), repeat_last_n), n_ctx);
+
+        llama_sample_repetition_penalty(ctx, &cur_p,
+                last_tokens.data() + last_tokens.size() - last_n_repeat,
+                last_n_repeat, repeat_penalty);
+        llama_sample_frequency_and_presence_penalties(ctx, &cur_p,
+                last_tokens.data() + last_tokens.size() - last_n_repeat,
+                last_n_repeat, alpha_frequency, alpha_presence);
+
+        if (!penalize_nl) {
+            for (size_t idx = 0; idx < cur_p.size; idx++) {
+                if (cur_p.data[idx].id == llama_token_nl(ctx)) {
+                    cur_p.data[idx].logit = nl_logit;
+                    break;
+                }
+            }
+        }
+    }
+
+    llama_sampler_sequence_context & ctx_seq = llama_sampling_get_sequence_context(ctx_sampling, seq);
+
+    if (ctx_seq.grammar != NULL) {
+        llama_sample_grammar(ctx, &cur_p, ctx_seq.grammar);
+    }
+
+    if (temp <= 0) {
+        // Greedy sampling
+        id = llama_sample_token_greedy(ctx, &cur_p);
+    } else {
+        if (mirostat == 1) {
+            const int mirostat_m = 100;
+            llama_sample_temp(ctx, &cur_p, temp);
+            id = llama_sample_token_mirostat(ctx, &cur_p, mirostat_tau, mirostat_eta, mirostat_m, &ctx_seq.mirostat_mu);
+        } else if (mirostat == 2) {
+            llama_sample_temp(ctx, &cur_p, temp);
+            id = llama_sample_token_mirostat_v2(ctx, &cur_p, mirostat_tau, mirostat_eta, &ctx_seq.mirostat_mu);
+        } else {
+            // Temperature sampling
+            size_t min_keep = std::max(1, params.n_probs);
+            llama_sample_top_k      (ctx, &cur_p, top_k, min_keep);
+            llama_sample_tail_free  (ctx, &cur_p, tfs_z, min_keep);
+            llama_sample_typical    (ctx, &cur_p, typical_p, min_keep);
+            llama_sample_top_p      (ctx, &cur_p, top_p, min_keep);
+            llama_sample_temp(ctx, &cur_p, temp);
+
+            {
+                const int n_top = 10;
+                LOG("top %d candidates:\n", n_top);
+
+                for (int i = 0; i < n_top; i++) {
+                    const llama_token id = cur_p.data[i].id;
+                    (void)id; // To avoid a warning that id is unused when logging is disabled.
+                    LOG(" - %5d: '%12s' (%.3f)\n", id, llama_token_to_piece(ctx, id).c_str(), cur_p.data[i].p);
+                }
+            }
+
+            id = llama_sample_token(ctx, &cur_p);
+
+            LOG("sampled token: %5d: '%s'\n", id, llama_token_to_piece(ctx, id).c_str());
+        }
+    }
+
+    if (ctx_seq.grammar != NULL) {
+        llama_grammar_accept_token(ctx, ctx_seq.grammar, id);
+    }
+
+    return id;
+}

common/sampling.h

@@ -0,0 +1,108 @@
+#pragma once
+
+#include "llama.h"
+
+#include <string>
+#include <vector>
+#include <unordered_map>
+
+// sampling parameters
+typedef struct llama_sampling_params {
+    int32_t top_k             = 40;    // <= 0 to use vocab size
+    float   top_p             = 0.95f; // 1.0 = disabled
+    float   tfs_z             = 1.00f; // 1.0 = disabled
+    float   typical_p         = 1.00f; // 1.0 = disabled
+    float   temp              = 0.80f; // 1.0 = disabled
+    float   repeat_penalty    = 1.10f; // 1.0 = disabled
+    int32_t repeat_last_n     = 64;    // last n tokens to penalize (0 = disable penalty, -1 = context size)
+    float   frequency_penalty = 0.00f; // 0.0 = disabled
+    float   presence_penalty  = 0.00f; // 0.0 = disabled
+    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    penalize_nl       = true;  // consider newlines as a repeatable token
+
+    int32_t n_probs           = 0;     // if greater than 0, output the probabilities of top n_probs tokens.
+
+    // Classifier-Free Guidance
+    // https://arxiv.org/abs/2306.17806
+    std::string cfg_negative_prompt;   // string to help guidance
+    float       cfg_scale     = 1.f;   // How strong is guidance
+
+    std::unordered_map<llama_token, float> logit_bias; // logit bias for specific tokens
+
+} llama_sampling_params;
+
+// per-sequence sampler context
+typedef struct llama_sampler_sequence_context {
+    float mirostat_mu; // mirostat sampler state
+    llama_grammar * grammar;
+} llama_sampler_sequence_context;
+
+// general sampler context
+typedef struct llama_sampling_context {
+    ~llama_sampling_context();
+
+    // parameters that will be used for sampling and when creating
+    // new llama_sampler_sequence_context instances
+    llama_sampling_params params;
+
+    // map of sequence ids to sampler contexts
+    std::unordered_map<llama_seq_id, llama_sampler_sequence_context> sequence_contexts;
+
+    // when non-NULL, new instances of llama_sampler_sequence_context
+    // will get a copy of the grammar here
+    // note: only the pointer is stored here, it is not a copy of
+    //       the grammar and shouldn't be freed
+    llama_grammar * grammar;
+} llama_sampling_context;
+
+#include "common.h"
+
+// Create a new sampling context instance.
+llama_sampling_context llama_sampling_context_init(
+        const struct gpt_params & params,
+                  llama_grammar * grammar = NULL);
+
+// Fetches the sampler context for the specified sequence id (defaults to 0).
+// If the context for that sequence id doesn't already exist, it will be created with
+// default values based on the parameters in the ctx_sampling argument.
+llama_sampler_sequence_context & llama_sampling_get_sequence_context(
+              llama_sampling_context & ctx_sampling,
+        const llama_seq_id             seq = 0);
+
+// Reset the sampler context for the supplied sequence id (defaults to 0).
+// This is necessary to reuse a sequence id or free memory used by sequences
+// that are no longer required.
+bool llama_sampling_context_reset(
+              llama_sampling_context & ctx_sampling,
+        const llama_seq_id             seq = 0);
+
+// this is a common sampling function used across the examples for convenience
+// it can serve as a starting point for implementing your own sampling function
+// Note: When using multiple sequences, it is the caller's responsibility to call
+//       llama_sampling_context_reset when a sequence ends
+//
+// required:
+//  - ctx:          context to use for sampling
+//  - ctx_sampling: sampling-specific context
+//
+// optional:
+//  - ctx_guidance:  context to use for classifier-free guidance, ignore if NULL
+//  - last_tokens:   needed for repetition penalty, ignore if empty
+//  - idx:           sample from llama_get_logits_ith(ctx, idx)
+//  - seq:           sequence id to associate sampler state with
+//
+// returns:
+//  - token:      sampled token
+//  - candidates: vector of candidate tokens
+//
+llama_token llama_sampling_sample(
+                  struct llama_context * ctx,
+                  struct llama_context * ctx_guidance,
+                  struct llama_sampling_context & ctx_sampling,
+        const std::vector<llama_token> & last_tokens,
+         std::vector<llama_token_data> & candidates,
+        const                      int   idx = 0,
+                          llama_seq_id   seq = 0);

convert-bloom-hf-to-gguf.py

@@ -0,0 +1,238 @@
+#!/usr/bin/env python3
+# HF bloom --> gguf conversion
+
+from __future__ import annotations
+
+import argparse
+import json
+import os
+import re
+import struct
+import sys
+from pathlib import Path
+from typing import Any
+
+import numpy as np
+import torch
+from transformers import AutoTokenizer  # type: ignore[import]
+
+if 'NO_LOCAL_GGUF' not in os.environ:
+    sys.path.insert(1, str(Path(__file__).parent / 'gguf-py' / 'gguf'))
+import gguf
+
+
+def count_model_parts(dir_model: Path) -> int:
+    num_parts = 0
+    for filename in os.listdir(dir_model):
+        if filename.startswith("pytorch_model-"):
+            num_parts += 1
+
+    if num_parts > 0:
+        print("gguf: found " + str(num_parts) + " model parts")
+    return num_parts
+
+
+# Supported Models:
+#   https://huggingface.co/bigscience/bloom-1b7
+#   https://huggingface.co/bigscience/bloom-3b
+#   https://huggingface.co/bigscience/bloom-7b1
+#   https://huggingface.co/Langboat/bloom-1b4-zh
+def parse_args() -> argparse.Namespace:
+    parser = argparse.ArgumentParser(description="Convert a Bloom model to a GGML compatible file")
+    parser.add_argument("--vocab-only", action="store_true", help="extract only the vocab")
+    parser.add_argument("--outfile",    type=Path,           help="path to write to; default: based on input")
+    parser.add_argument("model",        type=Path,           help="directory containing model file, or model file itself (*.bin)")
+    parser.add_argument("ftype",        type=int,            help="output format - use 0 for float32, 1 for float16", choices=[0, 1], default = 1)
+    return parser.parse_args()
+
+args = parse_args()
+
+dir_model = args.model
+ftype = args.ftype
+if not dir_model.is_dir():
+    print(f'Error: {args.model} is not a directory', file = sys.stderr)
+    sys.exit(1)
+
+# possible tensor data types
+#   ftype == 0 -> float32
+#   ftype == 1 -> float16
+
+# map from ftype to string
+ftype_str = ["f32", "f16"]
+
+if args.outfile is not None:
+    fname_out = args.outfile
+else:
+    # output in the same directory as the model by default
+    fname_out = dir_model / f'ggml-model-{ftype_str[ftype]}.gguf'
+
+print("gguf: loading model "+dir_model.name)
+
+with open(dir_model / "config.json", "r", encoding="utf-8") as f:
+    hparams = json.load(f)
+
+if hparams["architectures"][0] != "BloomForCausalLM":
+    print("Model architecture not supported: " + hparams["architectures"][0])
+    sys.exit(1)
+
+# get number of model parts
+num_parts = count_model_parts(dir_model)
+
+ARCH=gguf.MODEL_ARCH.BLOOM
+gguf_writer = gguf.GGUFWriter(fname_out, gguf.MODEL_ARCH_NAMES[ARCH])
+
+print("gguf: get model metadata")
+
+block_count = hparams["n_layer"]
+
+gguf_writer.add_name("Bloom")
+n_embed = hparams.get("hidden_size", hparams.get("n_embed"))
+n_head = hparams.get("n_head", hparams.get("num_attention_heads"))
+gguf_writer.add_context_length(hparams.get("seq_length", n_embed))
+gguf_writer.add_embedding_length(n_embed)
+gguf_writer.add_feed_forward_length(4 * n_embed)
+gguf_writer.add_block_count(block_count)
+gguf_writer.add_head_count(n_head)
+gguf_writer.add_head_count_kv(n_head)
+gguf_writer.add_layer_norm_eps(hparams["layer_norm_epsilon"])
+gguf_writer.add_file_type(ftype)
+
+# TOKENIZATION
+
+print("gguf: get tokenizer metadata")
+
+tokens: list[bytearray] = []
+scores: list[float] = []
+toktypes: list[int] = []
+
+# gpt2 tokenizer
+gguf_writer.add_tokenizer_model("gpt2")
+
+print("gguf: get gpt2 tokenizer vocab")
+
+# ref: https://github.com/cmp-nct/ggllm.cpp/blob/master/falcon_convert.py
+tokenizer = AutoTokenizer.from_pretrained(dir_model)
+
+# The number of tokens in tokenizer.json can differ from the expected vocab size.
+# This causes downstream issues with mismatched tensor sizes when running the inference
+vocab_size = hparams.get("vocab_size", len(tokenizer.vocab))
+assert max(tokenizer.vocab.values()) < vocab_size
+
+reverse_vocab = {id: encoded_tok for encoded_tok, id in tokenizer.vocab.items()}
+
+for i in range(vocab_size):
+    tokens.append(reverse_vocab[i] if i in reverse_vocab else f"[PAD{i}]")
+    scores.append(0.0)  # dummy
+    toktypes.append(gguf.TokenType.NORMAL)
+
+gguf_writer.add_token_list(tokens)
+gguf_writer.add_token_scores(scores)
+gguf_writer.add_token_types(toktypes)
+
+special_vocab = gguf.SpecialVocab(dir_model, load_merges=True)
+special_vocab.add_to_gguf(gguf_writer)
+
+# TENSORS
+
+tensor_map = gguf.get_tensor_name_map(ARCH, block_count)
+
+# params for qkv transform
+n_head_kv = hparams.get("n_head_kv", n_head)
+head_dim = n_embed // n_head
+
+# tensor info
+print("gguf: get tensor metadata")
+
+if num_parts == 0:
+    part_names = iter(("pytorch_model.bin",))
+else:
+    part_names = (
+        f"pytorch_model-{n:05}-of-{num_parts:05}.bin" for n in range(1, num_parts + 1)
+    )
+
+for part_name in part_names:
+    if args.vocab_only:
+        break
+    print("gguf: loading model part '" + part_name + "'")
+    model_part = torch.load(dir_model / part_name, map_location="cpu")
+
+    has_lm_head = True
+    if "lm_head.weight" not in model_part.keys() and "output.weight" not in model_part.keys():
+        has_lm_head = False
+
+    for original_name in model_part.keys():
+        data = model_part[original_name]
+        name = re.sub(r'transformer\.', '', original_name)
+
+        old_dtype = data.dtype
+
+        # convert any unsupported data types to float32
+        if data.dtype != torch.float16 and data.dtype != torch.float32:
+            data = data.to(torch.float32)
+
+        data = data.squeeze().numpy()
+
+        if re.match(r"h\.\d+\.self_attention\.query_key_value\.weight", name):
+            # Map bloom-style qkv_linear to gpt-style qkv_linear
+            # bloom: https://github.com/huggingface/transformers/blob/main/src/transformers/models/bloom/modeling_bloom.py#L238-L252  # noqa
+            # gpt-2: https://github.com/huggingface/transformers/blob/main/src/transformers/models/gpt2/modeling_gpt2.py#L312  # noqa
+            qkv_weights = data.reshape((n_head, 3, n_embed // n_head, n_embed))
+            data = np.concatenate(
+                (qkv_weights[:, 0, :, :].reshape((-1, n_embed)),
+                 qkv_weights[:, 1, :, :].reshape((-1, n_embed)),
+                 qkv_weights[:, 2, :, :].reshape((-1, n_embed))),
+                axis=0
+            )
+            print("re-format attention.linear_qkv.weight")
+        elif re.match(r"h\.\d+\.self_attention\.query_key_value\.bias", name):
+            qkv_bias = data.reshape((n_head, 3, n_embed // n_head))
+            data = np.concatenate(
+                (qkv_bias[:, 0, :].reshape((n_embed,)),
+                 qkv_bias[:, 1, :].reshape((n_embed,)),
+                 qkv_bias[:, 2, :].reshape((n_embed,))),
+                axis=0
+            )
+            print("re-format attention.linear_qkv.bias")
+
+        # map tensor names
+        new_name = tensor_map.get_name(name, try_suffixes=(".weight", ".bias"))
+        if new_name is None:
+            print("Can not map tensor '" + name + "'")
+            sys.exit()
+
+        n_dims = len(data.shape)
+        data_dtype = data.dtype
+
+        # if f32 desired, convert any float16 to float32
+        if ftype == 0 and data_dtype == np.float16:
+            data = data.astype(np.float32)
+
+        # TODO: Why cant we use these float16 as-is? There should be not reason to store float16 as float32
+        if ftype == 1 and data_dtype == np.float16 and n_dims == 1:
+            data = data.astype(np.float32)
+
+        # if f16 desired, convert any float32 2-dim weight tensors to float16
+        if ftype == 1 and data_dtype == np.float32 and name.endswith(".weight") and n_dims == 2:
+            data = data.astype(np.float16)
+
+        print(name, "=>", new_name + ", shape = " + str(data.shape) + ", " + str(old_dtype) + " --> " + str(data.dtype))
+
+        gguf_writer.add_tensor(new_name, data)
+
+        if not has_lm_head and name == "word_embeddings.weight":
+            gguf_writer.add_tensor("output.weight", data)
+            print(name, "=>", "output.weight" + ", shape = " + str(data.shape) + ", " + str(old_dtype) + " --> " + str(data.dtype))  # noqa
+
+
+print("gguf: write header")
+gguf_writer.write_header_to_file()
+print("gguf: write metadata")
+gguf_writer.write_kv_data_to_file()
+if not args.vocab_only:
+    print("gguf: write tensors")
+    gguf_writer.write_tensors_to_file()
+
+gguf_writer.close()
+
+print(f"gguf: model successfully exported to '{fname_out}'")
+print("")

convert-mpt-hf-to-gguf.py

@@ -0,0 +1,216 @@
+#!/usr/bin/env python3
+# HF mpt--> gguf conversion
+
+from __future__ import annotations
+
+import argparse
+import json
+import os
+import struct
+import sys
+from pathlib import Path
+from typing import Any
+
+import numpy as np
+import torch
+from transformers import AutoTokenizer  # type: ignore[import]
+
+if 'NO_LOCAL_GGUF' not in os.environ:
+    sys.path.insert(1, str(Path(__file__).parent / 'gguf-py' / 'gguf'))
+import gguf
+
+
+def count_model_parts(dir_model: Path) -> int:
+    num_parts = 0
+    for filename in os.listdir(dir_model):
+        if filename.startswith("pytorch_model-"):
+            num_parts += 1
+
+    if num_parts > 0:
+        print("gguf: found " + str(num_parts) + " model parts")
+    return num_parts
+
+
+def parse_args() -> argparse.Namespace:
+    parser = argparse.ArgumentParser(description="Convert an MPT model to a GGML compatible file")
+    parser.add_argument(
+        "--vocab-only", action="store_true",
+        help="extract only the vocab",
+    )
+    parser.add_argument(
+        "--outfile", type=Path,
+        help="path to write to; default: based on input",
+    )
+    parser.add_argument(
+        "model", type=Path,
+        help="directory containing model file, or model file itself (*.bin)",
+    )
+    parser.add_argument(
+        "ftype", type=int, choices=[0, 1], default=1, nargs='?',
+        help="output format - use 0 for float32, 1 for float16",
+    )
+    return parser.parse_args()
+
+args = parse_args()
+
+dir_model = args.model
+ftype = args.ftype
+if not dir_model.is_dir():
+    print(f'Error: {args.model} is not a directory', file = sys.stderr)
+    sys.exit(1)
+
+# possible tensor data types
+#   ftype == 0 -> float32
+#   ftype == 1 -> float16
+
+# map from ftype to string
+ftype_str = ["f32", "f16"]
+
+if args.outfile is not None:
+    fname_out = args.outfile
+else:
+    # output in the same directory as the model by default
+    fname_out = dir_model / f'ggml-model-{ftype_str[ftype]}.gguf'
+
+print("gguf: loading model "+dir_model.name)
+
+with open(dir_model / "config.json", "r", encoding="utf-8") as f:
+    hparams = json.load(f)
+
+if hparams["architectures"][0] != "MPTForCausalLM":
+    print("Model architecture not supported: " + hparams["architectures"][0])
+
+    sys.exit()
+
+# get number of model parts
+num_parts = count_model_parts(dir_model)
+
+ARCH=gguf.MODEL_ARCH.MPT
+gguf_writer = gguf.GGUFWriter(fname_out, gguf.MODEL_ARCH_NAMES[ARCH])
+
+print("gguf: get model metadata")
+
+block_count = hparams["n_layers"]
+
+gguf_writer.add_name(dir_model.name)
+gguf_writer.add_context_length(hparams["max_seq_len"])
+gguf_writer.add_embedding_length(hparams["d_model"])
+gguf_writer.add_block_count(block_count)
+gguf_writer.add_feed_forward_length(4 * hparams["d_model"])
+gguf_writer.add_head_count(hparams["n_heads"])
+gguf_writer.add_layer_norm_eps(1e-05)
+if hparams["attn_config"]["clip_qkv"] is not None:
+    gguf_writer.add_clamp_kqv(hparams["attn_config"]["clip_qkv"])
+gguf_writer.add_max_alibi_bias(hparams["attn_config"]["alibi_bias_max"])
+
+# TOKENIZATION
+
+print("gguf: get tokenizer metadata")
+
+tokens: list[bytearray] = []
+scores: list[float] = []
+toktypes: list[int] = []
+
+# gpt2 tokenizer
+gguf_writer.add_tokenizer_model("gpt2")
+
+print("gguf: get gpt2 tokenizer vocab")
+
+# MPT token embedding tensors have dimension 50432 (hparams["vocab_size"]), but
+# there are only 50254 (len(tokenizer.vocab)) tokens in the vocab, presumably to
+# accomodate some "reserved" tokens; this is causing problems down the line in
+# llama.cpp, so we pad the vocab with dummy tokens:
+
+vocab_size = hparams["vocab_size"]
+
+# ref: https://github.com/cmp-nct/ggllm.cpp/blob/master/falcon_convert.py
+tokenizer = AutoTokenizer.from_pretrained(dir_model)
+
+reverse_vocab = {id: encoded_tok for encoded_tok, id in tokenizer.vocab.items()}
+
+for i in range(vocab_size):
+    tokens.append(reverse_vocab[i] if i in reverse_vocab else f"[PAD{i}]")
+    scores.append(0.0) # dummy
+    toktypes.append(gguf.TokenType.NORMAL)
+
+gguf_writer.add_token_list(tokens)
+gguf_writer.add_token_scores(scores)
+gguf_writer.add_token_types(toktypes)
+
+special_vocab = gguf.SpecialVocab(dir_model, load_merges = True)
+special_vocab.add_to_gguf(gguf_writer)
+
+# TENSORS
+
+tensor_map = gguf.get_tensor_name_map(ARCH,block_count)
+
+# tensor info
+print("gguf: get tensor metadata")
+
+if num_parts == 0:
+    part_names = iter(("pytorch_model.bin",))
+else:
+    part_names = (
+        f"pytorch_model-{n:05}-of-{num_parts:05}.bin" for n in range(1, num_parts + 1)
+    )
+
+for part_name in part_names:
+    if args.vocab_only:
+        break
+    print("gguf: loading model part '" + part_name + "'")
+    model_part = torch.load(f"{dir_model}/{part_name}", map_location="cpu")
+
+    for name in model_part.keys():
+        data = model_part[name]
+
+        old_dtype = data.dtype
+
+        # convert any unsupported data types to float32
+        if data.dtype != torch.float16 and data.dtype != torch.float32:
+            data = data.to(torch.float32)
+
+        data = data.squeeze().numpy()
+
+        # map tensor names
+        new_name = tensor_map.get_name(name, try_suffixes = (".weight", ".bias"))
+        if new_name is None:
+            print("Cannot map tensor '" + name + "'")
+            continue # for the sake of compatibility with some old published models, don't quit
+            sys.exit()
+
+        n_dims = len(data.shape)
+        data_dtype = data.dtype
+
+        # if f32 desired, convert any float16 to float32
+        if ftype == 0 and data_dtype == np.float16:
+            data = data.astype(np.float32)
+
+        # TODO: Why cant we use these float16 as-is? There should be not reason to store float16 as float32
+        if ftype == 1 and data_dtype == np.float16 and n_dims == 1:
+            data = data.astype(np.float32)
+
+        # if f16 desired, convert any float32 2-dim weight tensors to float16
+        if ftype == 1 and data_dtype == np.float32 and name.endswith(".weight") and n_dims == 2:
+            data = data.astype(np.float16)
+
+        print(new_name + ", n_dims = " + str(n_dims) + ", " + str(old_dtype) + " --> " + str(data.dtype))
+
+        gguf_writer.add_tensor(new_name, data)
+
+        # note: MPT output is tied to (same as) wte in original model;
+        # for easier implementation in llama.cpp it's duplicated in GGUF, though :/
+        if new_name == "token_embd.weight":
+            gguf_writer.add_tensor("output.weight", data)
+
+print("gguf: write header")
+gguf_writer.write_header_to_file()
+print("gguf: write metadata")
+gguf_writer.write_kv_data_to_file()
+if not args.vocab_only:
+    print("gguf: write tensors")
+    gguf_writer.write_tensors_to_file()
+
+gguf_writer.close()
+
+print(f"gguf: model successfully exported to '{fname_out}'")
+print("")

convert-refact-hf-to-gguf.py

@@ -17,33 +17,6 @@ if "NO_LOCAL_GGUF" not in os.environ:
     sys.path.insert(1, str(Path(__file__).parent / "gguf-py" / "gguf"))
 import gguf
 
-
-def bytes_to_unicode():
-    # ref: https://github.com/openai/gpt-2/blob/master/src/encoder.py
-    """
-    Returns list of utf-8 byte and a corresponding list of unicode strings.
-    The reversible bpe codes work on unicode strings.
-    This means you need a large # of unicode characters in your vocab if you want to avoid UNKs.
-    When you're at something like a 10B token dataset you end up needing around 5K for decent coverage.
-    This is a significant percentage of your normal, say, 32K bpe vocab.
-    To avoid that, we want lookup tables between utf-8 bytes and unicode strings.
-    And avoids mapping to whitespace/control characters the bpe code barfs on.
-    """
-    bs = (
-        list(range(ord("!"), ord("~") + 1))
-        + list(range(ord("¡"), ord("¬") + 1))
-        + list(range(ord("®"), ord("ÿ") + 1))
-    )
-    cs = bs[:]
-    n = 0
-    for b in range(2**8):
-        if b not in bs:
-            bs.append(b)
-            cs.append(2**8 + n)
-            n += 1
-    return dict(zip(bs, (chr(n) for n in cs)))
-
-
 def count_model_parts(dir_model: Path) -> int:
     num_parts = 0
     for filename in os.listdir(dir_model):
@@ -153,53 +126,25 @@ tokens: list[bytearray] = []
 scores: list[float] = []
 toktypes: list[int] = []
 
-tokenizer_json_file = dir_model / "tokenizer.json"
-if not tokenizer_json_file.is_file():
-    print(f"Error: Missing {tokenizer_json_file}", file=sys.stderr)
-    sys.exit(1)
-
 # gpt2 tokenizer
 gguf_writer.add_tokenizer_model("gpt2")
 
-with open(tokenizer_json_file, "r", encoding="utf-8") as f:
-    tokenizer_json = json.load(f)
-
 print("gguf: get gpt2 tokenizer vocab")
 
+# ref: https://github.com/cmp-nct/ggllm.cpp/blob/master/falcon_convert.py
+tokenizer = AutoTokenizer.from_pretrained(dir_model)
+
 # The number of tokens in tokenizer.json can differ from the expected vocab size.
 # This causes downstream issues with mismatched tensor sizes when running the inference
-vocab_size = (
-    hparams["vocab_size"]
-    if "vocab_size" in hparams
-    else len(tokenizer_json["model"]["vocab"])
-)
-
-tokenizer = AutoTokenizer.from_pretrained(dir_model, trust_remote_code=True)
+vocab_size = hparams.get("vocab_size", len(tokenizer.vocab))
+assert max(tokenizer.vocab.values()) < vocab_size
 
 reverse_vocab = {id: encoded_tok for encoded_tok, id in tokenizer.vocab.items()}
-byte_encoder = bytes_to_unicode()
-byte_decoder = {v: k for k, v in byte_encoder.items()}
 
 for i in range(vocab_size):
-    if i in reverse_vocab:
-        text = reverse_vocab[i]
-        try:
-            text = bytearray([byte_decoder[c] for c in reverse_vocab[i]])
-        except KeyError:
-            text = bytearray()
-            for c in reverse_vocab[i]:
-                if ord(c) < 256:  # single byte character
-                    text.append(byte_decoder[ord(c)])
-                else:  # multibyte special token character
-                    text.extend(c.encode("utf-8"))
-    else:
-        print(f"Key {i} not in tokenizer vocabulary. Padding with an arbitrary token.")
-        pad_token = f"[PAD{i}]".encode("utf8")
-        text = bytearray(pad_token)
-
-    tokens.append(text)
-    scores.append(0.0)  # dymmy
-    toktypes.append(gguf.TokenType.NORMAL)  # dummy
+    tokens.append(reverse_vocab[i] if i in reverse_vocab else f"[PAD{i}]")
+    scores.append(0.0) # dummy
+    toktypes.append(gguf.TokenType.NORMAL)
 
 gguf_writer.add_token_list(tokens)
 gguf_writer.add_token_scores(scores)

docs/BLIS.md

@@ -49,7 +49,7 @@ According to the BLIS documentation, we could set the following
 environment variables to modify the behavior of openmp:
 
 ```bash
-export GOMP_GPU_AFFINITY="0-19"
+export GOMP_CPU_AFFINITY="0-19"
 export BLIS_NUM_THREADS=14
 ```
 

examples/CMakeLists.txt

@@ -25,9 +25,11 @@ else()
     add_subdirectory(convert-llama2c-to-ggml)
     add_subdirectory(simple)
     add_subdirectory(batched)
+    add_subdirectory(batched-bench)
     add_subdirectory(speculative)
     add_subdirectory(parallel)
     add_subdirectory(embd-input)
+    add_subdirectory(llava)
     add_subdirectory(llama-bench)
     add_subdirectory(beam-search)
     if (LLAMA_METAL)

examples/batched-bench/CMakeLists.txt

@@ -0,0 +1,5 @@
+set(TARGET batched-bench)
+add_executable(${TARGET} batched-bench.cpp)
+install(TARGETS ${TARGET} RUNTIME)
+target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT})
+target_compile_features(${TARGET} PRIVATE cxx_std_11)

examples/batched-bench/README.md

@@ -0,0 +1,51 @@
+# llama.cpp/example/batched-bench
+
+Benchmark the batched decoding performance of `llama.cpp`
+
+## Usage
+
+There are 2 modes of operation:
+
+- `prompt not shared` - each batch has a separate prompt of size `PP` (i.e. `N_KV = B*(PP + TG)`)
+- `prompt is shared` - there is a common prompt of size `PP` used by all batches (i.e. `N_KV = PP + B*TG`)
+
+```bash
+./batched-bench MODEL_PATH [N_KV_MAX] [IS_PP_SHARED] [NGL] [MMQ] <PP> <TG> <PL>
+
+# LLaMA 7B, F16, N_KV_MAX = 16384 (8GB), prompt not shared
+./batched-bench ./models/llama-7b/ggml-model-f16.gguf 16384 0 99
+
+# LLaMA 7B, Q8_0, N_KV_MAX = 16384 (8GB), prompt is shared
+./batched-bench ./models/llama-7b/ggml-model-q8_0.gguf 16384 1 99
+
+# custom set of batches
+./batched-bench ./models/llama-7b/ggml-model-q8_0.gguf 2048 0 999 0 128,256,512 128,256 1,2,4,8,16,32
+```
+
+## Sample results
+
+- `PP` - prompt tokens per batch
+- `TG` - generated tokens per batch
+- `B` - number of batches
+- `N_KV` - required KV cache size
+- `T_PP` - prompt processing time (i.e. time to first token)
+- `S_PP` - prompt processing speed (`(B*PP)/T_PP` or `PP/T_PP`)
+- `T_TG` - time to generate all batches
+- `S_TG` - text generation speed (`(B*TG)/T_TG`)
+- `T` - total time
+- `S` - total speed (i.e. all tokens / total time)
+
+|    PP |     TG |    B |   N_KV |   T_PP s | S_PP t/s |   T_TG s | S_TG t/s |      T s |    S t/s |
+|-------|--------|------|--------|----------|----------|----------|----------|----------|----------|
+|   128 |    128 |    1 |    256 |    0.108 |  1186.64 |    3.079 |    41.57 |    3.187 |    80.32 |
+|   128 |    128 |    2 |    512 |    0.198 |  1295.19 |    5.029 |    50.90 |    5.227 |    97.95 |
+|   128 |    128 |    4 |   1024 |    0.373 |  1373.96 |    6.878 |    74.44 |    7.251 |   141.23 |
+|   128 |    128 |    8 |   2048 |    0.751 |  1363.27 |    7.344 |   139.43 |    8.095 |   252.99 |
+|   128 |    128 |   16 |   4096 |    1.570 |  1304.68 |    8.455 |   242.23 |   10.024 |   408.60 |
+|   128 |    128 |   32 |   8192 |    3.408 |  1201.73 |    8.801 |   465.40 |   12.209 |   670.96 |
+|   128 |    256 |    1 |    384 |    0.107 |  1196.70 |    6.329 |    40.45 |    6.436 |    59.67 |
+|   128 |    256 |    2 |    768 |    0.194 |  1317.45 |   10.239 |    50.00 |   10.433 |    73.61 |
+|   128 |    256 |    4 |   1536 |    0.366 |  1399.03 |   13.960 |    73.35 |   14.326 |   107.22 |
+|   128 |    256 |    8 |   3072 |    0.751 |  1363.92 |   15.110 |   135.54 |   15.861 |   193.69 |
+|   128 |    256 |   16 |   6144 |    1.569 |  1304.93 |   18.073 |   226.64 |   19.642 |   312.80 |
+|   128 |    256 |   32 |  12288 |    3.409 |  1201.35 |   19.223 |   426.15 |   22.633 |   542.93 |

examples/batched-bench/batched-bench.cpp

@@ -0,0 +1,251 @@
+#include "common.h"
+#include "llama.h"
+
+#include <algorithm>
+#include <cmath>
+#include <cstdio>
+#include <string>
+#include <vector>
+
+// mutates the input string
+static std::vector<int> parse_list(char * p) {
+    std::vector<int> ret;
+
+    char * q = p;
+
+    while (*p) {
+        if (*p == ',') {
+            *p = '\0';
+            ret.push_back(std::atoi(q));
+            q = p + 1;
+        }
+
+        ++p;
+    }
+
+    ret.push_back(std::atoi(q));
+
+    return ret;
+}
+
+int main(int argc, char ** argv) {
+    gpt_params params;
+
+    if (argc == 1 || argv[1][0] == '-') {
+        printf("usage: %s MODEL_PATH [N_KV_MAX] [IS_PP_SHARED] [NGL] [MMQ] <PP> <TG> <PL>\n" , argv[0]);
+        printf("  <PP>, <TG> and PL are comma-separated lists of numbers without spaces\n\n");
+        printf("  example: %s ggml-model-f16.gguf 2048 0 999 0 128,256,512 128,256 1,2,4,8,16,32\n\n", argv[0]);
+        return 1 ;
+    }
+
+    int n_kv_max     = 2048;
+    int is_pp_shared = 0;
+    int n_gpu_layers = 0;
+    int mmq          = 0;
+
+    std::vector<int> n_pp = { 128, 256, 512, 1024, 2048, 3584, 7680, };
+    std::vector<int> n_tg = { 128, 256, };
+    std::vector<int> n_pl = { 1, 2, 4, 8, 16, 32, };
+    //std::vector<int> n_pl = { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 32, };
+
+    if (argc >= 2) {
+        params.model = argv[1];
+    }
+
+    if (argc >= 3) {
+        n_kv_max = std::atoi(argv[2]);
+    }
+
+    if (argc >= 4) {
+        is_pp_shared = std::atoi(argv[3]);
+    }
+
+    if (argc >= 5) {
+        n_gpu_layers = std::atoi(argv[4]);
+    }
+
+    if (argc >= 6) {
+        mmq = std::atoi(argv[5]);
+    }
+
+    if (argc >= 7) {
+        n_pp = parse_list(argv[6]);
+    }
+
+    if (argc >= 8) {
+        n_tg = parse_list(argv[7]);
+    }
+
+    if (argc >= 9) {
+        n_pl = parse_list(argv[8]);
+    }
+
+    // init LLM
+
+    llama_backend_init(params.numa);
+
+    // initialize the model
+
+    llama_model_params model_params = llama_model_default_params();
+
+    model_params.n_gpu_layers = n_gpu_layers;
+
+    llama_model * model = llama_load_model_from_file(params.model.c_str(), model_params);
+
+    if (model == NULL) {
+        fprintf(stderr , "%s: error: unable to load model\n" , __func__);
+        return 1;
+    }
+
+    llama_context_params ctx_params = llama_context_default_params();
+
+    ctx_params.seed      = 1234;
+    ctx_params.n_ctx     = n_kv_max;
+    ctx_params.n_batch   = 512;
+    ctx_params.mul_mat_q = mmq;
+
+    ctx_params.n_threads       = params.n_threads;
+    ctx_params.n_threads_batch = params.n_threads_batch == -1 ? params.n_threads : params.n_threads_batch;
+
+    llama_context * ctx = llama_new_context_with_model(model, ctx_params);
+
+    if (ctx == NULL) {
+        fprintf(stderr , "%s: error: failed to create the llama_context\n" , __func__);
+        return 1;
+    }
+
+    llama_batch batch = llama_batch_init(n_kv_max, 0);
+
+    // decode in batches of ctx_params.n_batch tokens
+    auto decode_helper = [](llama_context * ctx, llama_batch & batch, int32_t n_batch) {
+        for (int32_t i = 0; i < (int32_t) batch.n_tokens; i += n_batch) {
+            const int32_t n_tokens = std::min(n_batch, (int32_t) (batch.n_tokens - i));
+
+            llama_batch batch_view = {
+                n_tokens,
+                batch.token  + i,
+                nullptr,
+                batch.pos    + i,
+                batch.seq_id + i,
+                batch.logits + i,
+                0, 0, 0, // unused
+            };
+
+            const int ret = llama_decode(ctx, batch_view);
+            if (ret != 0) {
+                LOG_TEE("failed to decode the batch, n_batch = %d, ret = %d\n", n_batch, ret);
+                return false;
+            }
+        }
+
+        return true;
+    };
+
+    // warm up
+    {
+        batch.n_tokens = 16;
+
+        for (int i = 0; i < batch.n_tokens; ++i) {
+            batch.token[i]  = 0;
+            batch.pos[i]    = i;
+            batch.seq_id[i] = 0;
+            batch.logits[i] = false;
+        }
+
+        if (!decode_helper(ctx, batch, ctx_params.n_batch)) {
+            LOG_TEE("%s: llama_decode() failed\n", __func__);
+            return 1;
+        }
+    }
+
+    LOG_TEE("|%6s | %6s | %4s | %6s | %8s | %8s | %8s | %8s | %8s | %8s |\n", "PP",     "TG",     "B",    "N_KV",     "T_PP s",   "S_PP t/s", "T_TG s",   "S_TG t/s", "T s",      "S t/s");
+    LOG_TEE("|%6s-|-%6s-|-%4s-|-%6s-|-%8s-|-%8s-|-%8s-|-%8s-|-%8s-|-%8s-|\n", "------", "------", "----", "------", "--------", "--------", "--------", "--------", "--------", "--------");
+
+    for (        int i_pp = 0; i_pp < (int) n_pp.size(); ++i_pp) {
+        for (    int i_tg = 0; i_tg < (int) n_tg.size(); ++i_tg) {
+            for (int i_pl = 0; i_pl < (int) n_pl.size(); ++i_pl) {
+                const int pp = n_pp[i_pp];
+                const int tg = n_tg[i_tg];
+                const int pl = n_pl[i_pl];
+
+                const int n_ctx_req = is_pp_shared ? pp + pl*tg : pl*(pp + tg);
+
+                if (n_ctx_req > n_kv_max) {
+                    continue;
+                }
+
+                batch.n_tokens = is_pp_shared ? pp : pl*pp;
+
+                for (int i = 0; i < batch.n_tokens; ++i) {
+                    batch.token[i]  = 0;
+                    batch.pos[i]    = i;
+                    batch.seq_id[i] = 0;
+                    batch.logits[i] = false;
+                }
+                batch.logits[batch.n_tokens - 1] = true;
+
+                const auto t_pp_start = ggml_time_us();
+
+                llama_kv_cache_tokens_rm(ctx, -1, -1);
+
+                if (!decode_helper(ctx, batch, ctx_params.n_batch)) {
+                    LOG_TEE("%s: llama_decode() failed\n", __func__);
+                    return 1;
+                }
+
+                if (is_pp_shared) {
+                    for (int32_t i = 1; i < pl; ++i) {
+                        llama_kv_cache_seq_cp(ctx, 0, i, 0, pp);
+                    }
+                }
+
+                const auto t_pp_end = ggml_time_us();
+
+                const auto t_tg_start = ggml_time_us();
+
+                for (int i = 0; i < tg; ++i) {
+                    batch.n_tokens = pl;
+
+                    for (int j = 0; j < pl; ++j) {
+                        batch.token[j]  = 0;
+                        batch.pos[j]    = pp + i;
+                        batch.seq_id[j] = j;
+                        batch.logits[j] = true;
+                    }
+
+                    if (!decode_helper(ctx, batch, ctx_params.n_batch)) {
+                        LOG_TEE("%s: llama_decode() failed\n", __func__);
+                        return 1;
+                    }
+                }
+
+                const auto t_tg_end = ggml_time_us();
+
+                const int32_t n_kv = n_ctx_req;
+
+                const float t_pp = (t_pp_end - t_pp_start) / 1000000.0f;
+                const float t_tg = (t_tg_end - t_tg_start) / 1000000.0f;
+                const float t    = t_pp + t_tg;
+
+                const float speed_pp = is_pp_shared ? pp / t_pp : pl*pp / t_pp;
+                const float speed_tg = pl*tg / t_tg;
+                const float speed    = n_kv / t;
+
+                LOG_TEE("|%6d | %6d | %4d | %6d | %8.3f | %8.2f | %8.3f | %8.2f | %8.3f | %8.2f |\n", pp, tg, pl, n_kv, t_pp, speed_pp, t_tg, speed_tg, t, speed);
+            }
+        }
+    }
+
+    llama_print_timings(ctx);
+
+    llama_batch_free(batch);
+
+    llama_free(ctx);
+    llama_free_model(model);
+
+    llama_backend_free();
+
+    fprintf(stderr, "\n\n");
+
+    return 0;
+}

examples/batched.swift/.gitignore

@@ -0,0 +1,9 @@
+.DS_Store
+/.build
+/Packages
+xcuserdata/
+DerivedData/
+.swiftpm/configuration/registries.json
+.swiftpm/xcode/package.xcworkspace/contents.xcworkspacedata
+.netrc
+batched_swift

examples/batched.swift/Makefile

@@ -0,0 +1,6 @@
+.PHONY: build
+
+build:
+	xcodebuild -scheme batched_swift -destination "generic/platform=macOS" -derivedDataPath build
+	rm -f ./batched_swift
+	ln -s ./build/Build/Products/Debug/batched_swift ./batched_swift

examples/batched.swift/Package.swift

@@ -0,0 +1,22 @@
+// swift-tools-version: 5.5
+// The swift-tools-version declares the minimum version of Swift required to build this package.
+
+import PackageDescription
+
+let package = Package(
+    name: "batched_swift",
+    platforms: [.macOS(.v12)],
+    dependencies: [
+        .package(name: "llama", path: "../../"),
+    ],
+    targets: [
+        // Targets are the basic building blocks of a package, defining a module or a test suite.
+        // Targets can depend on other targets in this package and products from dependencies.
+        .executableTarget(
+            name: "batched_swift",
+            dependencies: ["llama"],
+            path: "Sources",
+            linkerSettings: [.linkedFramework("Foundation"), .linkedFramework("AppKit")]
+        ),
+    ]
+)

examples/batched.swift/README.md

@@ -0,0 +1,4 @@
+This is a swift clone of `examples/batched`.
+
+$ `make`
+$ `./swift MODEL_PATH [PROMPT] [PARALLEL]`

examples/batched.swift/Sources/main.swift

@@ -0,0 +1,255 @@
+import Foundation
+import llama
+
+let arguments = CommandLine.arguments
+
+// Check that we have at least one argument (the model path)
+guard arguments.count > 1 else {
+    print("Usage: swift MODEL_PATH [PROMPT] [PARALLEL]")
+    exit(1)
+}
+
+let modelPath: String = arguments[1]
+let prompt: String = arguments.count > 2 ? arguments[2] : "Hello my name is"
+let n_parallel: Int = arguments.count > 3 && Int(arguments[3]) != nil ? Int(arguments[3])! : 1
+
+// total length of the sequences including the prompt
+let n_len: Int = 32
+
+// init LLM
+llama_backend_init(false)
+defer {
+    llama_backend_free()
+}
+
+let model_params = llama_model_default_params()
+guard let model = llama_load_model_from_file(modelPath.cString(using: .utf8), model_params) else {
+    print("Failed to load model")
+    exit(1)
+}
+
+defer {
+    llama_free_model(model)
+}
+
+var tokens = tokenize(text: prompt, add_bos: true)
+
+let n_kv_req = UInt32(tokens.count) + UInt32((n_len - Int(tokens.count)) * n_parallel)
+
+var context_params = llama_context_default_params()
+context_params.seed = 1234
+context_params.n_ctx = n_kv_req
+context_params.n_batch = UInt32(max(n_len, n_parallel))
+context_params.n_threads = 8
+context_params.n_threads_batch = 8
+
+let context = llama_new_context_with_model(model, context_params)
+guard context != nil else {
+    print("Failed to initialize context")
+    exit(1)
+}
+
+defer {
+    llama_free(context)
+}
+
+let n_ctx = llama_n_ctx(context)
+
+print("\nn_len = \(n_len), n_ctx = \(n_ctx), n_batch = \(context_params.n_batch), n_parallel = \(n_parallel), n_kv_req = \(n_kv_req)\n")
+
+if n_kv_req > n_ctx {
+    print("error: n_kv_req (%d) > n_ctx, the required KV cache size is not big enough\n", n_kv_req)
+    exit(1)
+}
+
+var buffer: [CChar] = []
+for id: llama_token in tokens {
+    print(token_to_piece(token: id, buffer: &buffer) ?? "", terminator: "")
+}
+
+print("\n")
+
+var batch = llama_batch_init(max(Int32(tokens.count), Int32(n_parallel)), 0)
+defer {
+    llama_batch_free(batch)
+}
+
+// evaluate the initial prompt
+batch.n_tokens = Int32(tokens.count)
+
+for (i, token) in tokens.enumerated() {
+    batch.token[i] = token
+    batch.pos[i] = Int32(i)
+    batch.seq_id[i] = 0
+    batch.logits[i] = 0
+}
+
+// llama_decode will output logits only for the last token of the prompt
+batch.logits[Int(batch.n_tokens) - 1] = 1
+
+if llama_decode(context, batch) != 0 {
+    print("llama_decode() failed")
+    exit(1)
+}
+
+for i in 1 ..< n_parallel {
+    llama_kv_cache_seq_cp(context, 0, Int32(i), 0, batch.n_tokens)
+}
+
+if n_parallel > 1 {
+    print("generating \(n_parallel) sequences ...\n")
+}
+
+var streams: [String] = .init(repeating: "", count: n_parallel)
+var streamBuffers: [[CChar]] = .init(repeating: [], count: n_parallel)
+var i_batch = [Int32](repeating: batch.n_tokens - 1, count: n_parallel)
+
+var n_cur = batch.n_tokens
+var n_decode = 0
+
+let t_main_start = ggml_time_us()
+
+while n_cur <= n_len {
+    // prepare the next batch
+    batch.n_tokens = 0
+
+    // sample the next token for each parallel sequence / stream
+    for i in 0 ..< n_parallel {
+        if i_batch[i] < 0 {
+            // the stream has already finished
+            continue
+        }
+
+        var n_vocab = llama_n_vocab(model)
+        var logits = llama_get_logits_ith(context, i_batch[i])
+
+        var candidates: [llama_token_data] = .init(repeating: llama_token_data(), count: Int(n_vocab))
+
+        for token_id in 0 ..< n_vocab {
+            candidates.append(llama_token_data(id: token_id, logit: logits![Int(token_id)], p: 0.0))
+        }
+
+        var candidates_p: llama_token_data_array = .init(
+            data: &candidates,
+            size: candidates.count,
+            sorted: false
+        )
+
+        let top_k: Int32 = 40
+        let top_p: Float = 0.9
+        let temp: Float = 0.4
+
+        llama_sample_top_k(context, &candidates_p, top_k, 1)
+        llama_sample_top_p(context, &candidates_p, top_p, 1)
+        llama_sample_temp(context, &candidates_p, temp)
+
+        let new_token_id = llama_sample_token(context, &candidates_p)
+
+        // const llama_token new_token_id = llama_sample_token_greedy(ctx, &candidates_p);
+
+        // is it an end of stream? -> mark the stream as finished
+        if new_token_id == llama_token_eos(context) || n_cur == n_len {
+            i_batch[i] = -1
+            // print("")
+            if n_parallel > 1 {
+                print("stream \(i) finished at n_cur = \(n_cur)")
+            }
+
+            continue
+        }
+
+        let nextStringPiece = token_to_piece(token: new_token_id, buffer: &streamBuffers[i]) ?? ""
+
+        // if there is only one stream, we print immediately to stdout
+        if n_parallel == 1 {
+            print(nextStringPiece, terminator: "")
+        }
+        streams[i] += nextStringPiece
+
+        // push this new token for next evaluation
+        batch.token[Int(batch.n_tokens)] = new_token_id
+        batch.pos[Int(batch.n_tokens)] = n_cur
+        batch.seq_id[Int(batch.n_tokens)] = Int32(i)
+        batch.logits[Int(batch.n_tokens)] = 1
+
+        i_batch[i] = batch.n_tokens
+
+        batch.n_tokens += 1
+
+        n_decode += 1
+    }
+
+    // all streams are finished
+    if batch.n_tokens == 0 {
+        break
+    }
+
+    n_cur += 1
+
+    // evaluate the current batch with the transformer model
+    if llama_decode(context, batch) != 0 {
+        print("llama_decode() failed")
+        exit(1)
+    }
+}
+
+if n_parallel > 1 {
+    print("\n")
+    for (i, stream) in streams.enumerated() {
+        print("sequence \(i):\n\n\(prompt)\(stream)\n")
+    }
+}
+
+let t_main_end = ggml_time_us()
+
+print("decoded \(n_decode) tokens in \(String(format: "%.2f", Double(t_main_end - t_main_start) / 1_000_000.0)) s, speed: \(String(format: "%.2f", Double(n_decode) / (Double(t_main_end - t_main_start) / 1_000_000.0))) t/s\n")
+
+llama_print_timings(context)
+
+private func tokenize(text: String, add_bos: Bool) -> [llama_token] {
+    let n_tokens = text.count + (add_bos ? 1 : 0)
+    let tokens = UnsafeMutablePointer<llama_token>.allocate(capacity: n_tokens)
+    let tokenCount = llama_tokenize(model, text, Int32(text.count), tokens, Int32(n_tokens), add_bos)
+    var swiftTokens: [llama_token] = []
+    for i in 0 ..< tokenCount {
+        swiftTokens.append(tokens[Int(i)])
+    }
+    tokens.deallocate()
+    return swiftTokens
+}
+
+private func token_to_piece(token: llama_token, buffer: inout [CChar]) -> String? {
+    var result = [CChar](repeating: 0, count: 8)
+    let nTokens = llama_token_to_piece(model, token, &result, Int32(result.count))
+    if nTokens < 0 {
+        if result.count >= -Int(nTokens) {
+            result.removeLast(-Int(nTokens))
+        } else {
+            result.removeAll()
+        }
+        let check = llama_token_to_piece(
+            model,
+            token,
+            &result,
+            Int32(result.count)
+        )
+        assert(check == nTokens)
+    } else {
+        result.removeLast(result.count - Int(nTokens))
+    }
+    if buffer.isEmpty, let utfString = String(cString: result + [0], encoding: .utf8) {
+        return utfString
+    } else {
+        buffer.append(contentsOf: result)
+        let data = Data(buffer.map { UInt8(bitPattern: $0) })
+        if buffer.count >= 4 { // 4 bytes is the max length of a utf8 character so if we're here we need to reset the buffer
+            buffer = []
+        }
+        guard let bufferString = String(data: data, encoding: .utf8) else {
+            return nil
+        }
+        buffer = []
+        return bufferString
+    }
+    return nil
+}

examples/embd-input/embd-input-lib.cpp

@@ -128,21 +128,22 @@ bool eval_string(struct MyModel * mymodel,const char* str){
 llama_token sampling_id(struct MyModel* mymodel) {
     llama_context* ctx = mymodel->ctx;
     gpt_params params = mymodel->params;
+    llama_sampling_params & sparams = params.sampling_params;
     // int n_ctx = llama_n_ctx(ctx);
 
     // out of user input, sample next token
-    const float   temp            = params.temp;
-    const int32_t top_k           = params.top_k <= 0 ? llama_n_vocab(llama_get_model(ctx)) : params.top_k;
-    const float   top_p           = params.top_p;
-    const float   tfs_z           = params.tfs_z;
-    const float   typical_p       = params.typical_p;
+    const float   temp            = sparams.temp;
+    const int32_t top_k           = sparams.top_k <= 0 ? llama_n_vocab(llama_get_model(ctx)) : sparams.top_k;
+    const float   top_p           = sparams.top_p;
+    const float   tfs_z           = sparams.tfs_z;
+    const float   typical_p       = sparams.typical_p;
     // const int32_t repeat_last_n   = params.repeat_last_n < 0 ? n_ctx : params.repeat_last_n;
     // const float   repeat_penalty  = params.repeat_penalty;
     // const float   alpha_presence  = params.presence_penalty;
     // const float   alpha_frequency = params.frequency_penalty;
-    const int     mirostat        = params.mirostat;
-    const float   mirostat_tau    = params.mirostat_tau;
-    const float   mirostat_eta    = params.mirostat_eta;
+    const int     mirostat        = sparams.mirostat;
+    const float   mirostat_tau    = sparams.mirostat_tau;
+    const float   mirostat_eta    = sparams.mirostat_eta;
     // const bool    penalize_nl     = params.penalize_nl;
 
     llama_token id = 0;
@@ -151,7 +152,7 @@ llama_token sampling_id(struct MyModel* mymodel) {
         auto n_vocab = llama_n_vocab(llama_get_model(ctx));
 
         // Apply params.logit_bias map
-        for (auto it = params.logit_bias.begin(); it != params.logit_bias.end(); it++) {
+        for (auto it = sparams.logit_bias.begin(); it != sparams.logit_bias.end(); it++) {
             logits[it->first] += it->second;
         }
 

examples/finetune/finetune.cpp

@@ -529,13 +529,14 @@ static void init_lora(const struct my_llama_model * model, struct my_llama_lora
     set_param_lora(lora);
 
     // measure data size
-    struct ggml_allocr * alloc = NULL;
-    alloc = ggml_allocr_new_measure(tensor_alignment);
-    alloc_lora(alloc, lora);
+    size_t size = 0;
+    for (struct ggml_tensor * t = ggml_get_first_tensor(ctx); t != NULL; t = ggml_get_next_tensor(ctx, t)) {
+        size += GGML_PAD(ggml_nbytes(t), tensor_alignment);
+    }
 
     // allocate data
-    lora->data.resize(ggml_allocr_max_size(alloc) + tensor_alignment);
-    ggml_allocr_free(alloc);
+    struct ggml_allocr * alloc = NULL;
+    lora->data.resize(size + tensor_alignment);
     alloc = ggml_allocr_new(lora->data.data(), lora->data.size(), tensor_alignment);
     alloc_lora(alloc, lora);
     ggml_allocr_free(alloc);
@@ -1715,11 +1716,9 @@ int main(int argc, char ** argv) {
     struct ggml_tensor * target_probs  = ggml_new_tensor_3d(ctx_input, GGML_TYPE_F32, n_vocab,  n_tokens, n_batch);
 
     // measure required memory for input tensors
-    alloc = ggml_allocr_new_measure(tensor_alignment);
-    ggml_allocr_alloc(alloc, tokens_input);
-    ggml_allocr_alloc(alloc, target_probs);
-    size_t max_input_size = ggml_allocr_max_size(alloc) + tensor_alignment;
-    ggml_allocr_free(alloc);
+    size_t max_input_size = GGML_PAD(ggml_nbytes(tokens_input), tensor_alignment) +
+                            GGML_PAD(ggml_nbytes(target_probs), tensor_alignment) +
+                            tensor_alignment;
     printf("%s: input_size = %zu bytes (%.1f MB)\n", __func__, max_input_size, (float) max_input_size / (1024.0f*1024.0f));
 
     // allocate input tensors

examples/infill/infill.cpp

@@ -104,6 +104,7 @@ static void sigint_handler(int signo) {
 
 int main(int argc, char ** argv) {
     gpt_params params;
+    llama_sampling_params & sparams = params.sampling_params;
     g_params = &params;
 
     if (!gpt_params_parse(argc, argv, params)) {
@@ -206,7 +207,7 @@ int main(int argc, char ** argv) {
     // load the model and apply lora adapter, if any
     LOG("%s: load the model and apply lora adapter, if any\n", __func__);
     std::tie(model, ctx) = llama_init_from_gpt_params(params);
-    if (params.cfg_scale > 1.f) {
+    if (sparams.cfg_scale > 1.f) {
         struct llama_context_params lparams = llama_context_params_from_gpt_params(params);
         ctx_guidance = llama_new_context_with_model(model, lparams);
     }
@@ -233,10 +234,22 @@ int main(int argc, char ** argv) {
     const bool add_bos = llama_vocab_type(model) == LLAMA_VOCAB_TYPE_SPM;
     LOG("add_bos: %d\n", add_bos);
 
+    bool suff_rm_leading_spc = params.escape;
+    if (suff_rm_leading_spc && params.input_suffix.find_first_of(" ") == 0 && params.input_suffix.size() > 1) {
+        params.input_suffix.erase(0, 1);
+        suff_rm_leading_spc = false;
+    }
     std::vector<llama_token> embd_inp;
-    std::vector<llama_token> inp_pfx = ::llama_tokenize(ctx, params.input_prefix, add_bos);
-    std::vector<llama_token> inp_sfx = ::llama_tokenize(ctx, params.input_suffix, add_bos);
+    std::vector<llama_token> inp_pfx = ::llama_tokenize(ctx, params.input_prefix, false);
+    std::vector<llama_token> inp_sfx = ::llama_tokenize(ctx, params.input_suffix, false);
+    const int space_token = 29871;
+    if (suff_rm_leading_spc && inp_sfx[0] == space_token) {
+        inp_sfx.erase(inp_sfx.begin());
+    }
     inp_pfx.insert(inp_pfx.begin(), llama_token_prefix(ctx));
+    if (add_bos) {
+        inp_pfx.insert(inp_pfx.begin(), llama_token_bos(ctx));
+    }
     inp_sfx.insert(inp_sfx.begin(), llama_token_suffix(ctx));
     embd_inp = inp_pfx;
     embd_inp.insert(embd_inp.end(), inp_sfx.begin(), inp_sfx.end());
@@ -257,9 +270,9 @@ int main(int argc, char ** argv) {
     int guidance_offset = 0;
     int original_prompt_len = 0;
     if (ctx_guidance) {
-        LOG("cfg_negative_prompt: \"%s\"\n", log_tostr(params.cfg_negative_prompt));
+        LOG("cfg_negative_prompt: \"%s\"\n", log_tostr(sparams.cfg_negative_prompt));
 
-        guidance_inp = ::llama_tokenize(ctx_guidance, params.cfg_negative_prompt, add_bos);
+        guidance_inp = ::llama_tokenize(ctx_guidance, sparams.cfg_negative_prompt, add_bos);
         LOG("guidance_inp tokenized: %s\n", LOG_TOKENS_TOSTR_PRETTY(ctx_guidance, guidance_inp));
 
         std::vector<llama_token> original_inp = ::llama_tokenize(ctx, params.prompt, add_bos);
@@ -300,7 +313,7 @@ int main(int argc, char ** argv) {
 
         if (ctx_guidance) {
             LOG_TEE("\n");
-            LOG_TEE("%s: negative prompt: '%s'\n", __func__, params.cfg_negative_prompt.c_str());
+            LOG_TEE("%s: negative prompt: '%s'\n", __func__, sparams.cfg_negative_prompt.c_str());
             LOG_TEE("%s: number of tokens in negative prompt = %zu\n", __func__, guidance_inp.size());
             for (int i = 0; i < (int) guidance_inp.size(); i++) {
                 LOG_TEE("%6d -> '%s'\n", guidance_inp[i], llama_token_to_piece(ctx, guidance_inp[i]).c_str());
@@ -346,7 +359,7 @@ int main(int argc, char ** argv) {
         }
     }
     LOG_TEE("sampling: repeat_last_n = %d, repeat_penalty = %f, presence_penalty = %f, frequency_penalty = %f, top_k = %d, tfs_z = %f, top_p = %f, typical_p = %f, temp = %f, mirostat = %d, mirostat_lr = %f, mirostat_ent = %f\n",
-            params.repeat_last_n, params.repeat_penalty, params.presence_penalty, params.frequency_penalty, params.top_k, params.tfs_z, params.top_p, params.typical_p, params.temp, params.mirostat, params.mirostat_eta, params.mirostat_tau);
+            sparams.repeat_last_n, sparams.repeat_penalty, sparams.presence_penalty, sparams.frequency_penalty, sparams.top_k, sparams.tfs_z, sparams.top_p, sparams.typical_p, sparams.temp, sparams.mirostat, sparams.mirostat_eta, sparams.mirostat_tau);
     LOG_TEE("generate: n_ctx = %d, n_batch = %d, n_predict = %d, n_keep = %d\n", n_ctx, params.n_batch, params.n_predict, params.n_keep);
     LOG_TEE("\n\n");
 
@@ -364,8 +377,8 @@ int main(int argc, char ** argv) {
         LOG_TEE("\n");
 
         {
-            auto it = params.logit_bias.find(llama_token_eos(ctx));
-            if (it != params.logit_bias.end() && it->second == -INFINITY) {
+            auto it = sparams.logit_bias.find(llama_token_eos(ctx));
+            if (it != sparams.logit_bias.end() && it->second == -INFINITY) {
                 LOG_TEE("%s: warning: EOS token is disabled, which will cause most grammars to fail\n", __func__);
             }
         }
@@ -422,6 +435,7 @@ int main(int argc, char ** argv) {
 
     const int n_vocab = llama_n_vocab(model);
 
+    llama_sampling_context ctx_sampling = llama_sampling_context_init(params, grammar);
     std::vector<llama_token_data> candidates;
     candidates.reserve(n_vocab);
 
@@ -540,7 +554,7 @@ int main(int argc, char ** argv) {
 
         if ((int) embd_inp.size() <= n_consumed && !is_interacting) {
 
-            const llama_token id = llama_sample_token(ctx, ctx_guidance, grammar, params, last_tokens, candidates);
+            const llama_token id = llama_sampling_sample(ctx, ctx_guidance, ctx_sampling, last_tokens, candidates);
 
             last_tokens.erase(last_tokens.begin());
             last_tokens.push_back(id);
@@ -627,10 +641,27 @@ int main(int argc, char ** argv) {
                 buffer.clear();
                 // done taking input, reset color
                 console::set_display(console::reset);
+
+                if (params.escape) {
+                    //process escape sequences, for the initial prompt this is done in common.cpp when we load the params, but for the interactive mode we need to do it here
+                    process_escapes(params.input_prefix);
+                    process_escapes(params.input_suffix);
+                }
+                suff_rm_leading_spc = params.escape;
+                if (suff_rm_leading_spc && params.input_suffix.find_first_of(" ") == 0 && params.input_suffix.size() > 1) {
+                    params.input_suffix.erase(0, 1);
+                    suff_rm_leading_spc = false;
+                }
                 // tokenize new prefix and suffix
-                std::vector<llama_token> inp_pfx = ::llama_tokenize(ctx, params.input_prefix, add_bos);
-                std::vector<llama_token> inp_sfx = ::llama_tokenize(ctx, params.input_suffix, add_bos);
+                std::vector<llama_token> inp_pfx = ::llama_tokenize(ctx, params.input_prefix, false);
+                std::vector<llama_token> inp_sfx = ::llama_tokenize(ctx, params.input_suffix, false);
+                if (suff_rm_leading_spc && inp_sfx[0] == space_token) {
+                    inp_sfx.erase(inp_sfx.begin());
+                }
                 inp_pfx.insert(inp_pfx.begin(), llama_token_prefix(ctx));
+                if (add_bos) {
+                    inp_pfx.insert(inp_pfx.begin(), llama_token_bos(ctx));
+                }
                 inp_sfx.insert(inp_sfx.begin(), llama_token_suffix(ctx));
                 embd_inp = inp_pfx;
                 embd_inp.insert(embd_inp.end(), inp_sfx.begin(), inp_sfx.end());

examples/llava/CMakeLists.txt

@@ -0,0 +1,20 @@
+set(TARGET clip)
+add_library(${TARGET} clip.cpp clip.h)
+install(TARGETS ${TARGET} LIBRARY)
+target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT})
+target_compile_features(${TARGET} PRIVATE cxx_std_11)
+if (NOT MSVC)
+    target_compile_options(${TARGET} PRIVATE -Wno-cast-qual) # stb_image.h
+    endif()
+if(TARGET BUILD_INFO)
+    add_dependencies(${TARGET} BUILD_INFO)
+endif()
+
+set(TARGET llava)
+add_executable(${TARGET} llava.cpp)
+install(TARGETS ${TARGET} RUNTIME)
+target_link_libraries(${TARGET} PRIVATE common llama clip ${CMAKE_THREAD_LIBS_INIT})
+target_compile_features(${TARGET} PRIVATE cxx_std_11)
+if(TARGET BUILD_INFO)
+    add_dependencies(${TARGET} BUILD_INFO)
+endif()

examples/llava/README.md

@@ -0,0 +1,57 @@
+# LLaVA
+
+Currently this implementation supports [llava-v1.5](https://huggingface.co/liuhaotian/llava-v1.5-7b) variants.
+
+The pre-converted [7b](https://huggingface.co/mys/ggml_llava-v1.5-7b)
+and [13b](https://huggingface.co/mys/ggml_llava-v1.5-13b)
+models are available.
+
+After API is confirmed, more models will be supported / uploaded.
+
+## Usage
+Build with cmake or run `make llava` to build it.
+
+After building, run: `./llava` to see the usage. For example:
+
+```sh
+./llava -m llava-v1.5-7b/ggml-model-q5_k.gguf --mmproj llava-v1.5-7b/mmproj-model-f16.gguf --image path/to/an/image.jpg
+```
+
+**note**: A lower temperature like 0.1 is recommended for better quality. add `--temp 0.1` to the command to do so.
+
+## Model conversion
+
+- Clone `llava-v15-7b`` and `clip-vit-large-patch14-336`` locally:
+
+```sh
+git clone https://huggingface.co/liuhaotian/llava-v1.5-7b
+
+git clone https://huggingface.co/openai/clip-vit-large-patch14-336
+```
+
+2. Use `llava-surgery.py` to split the LLaVA model to LLaMA and multimodel projector constituents:
+
+```sh
+python ./examples/llava/llava-surgery.py -m ../llava-v1.5-7b
+```
+
+3. Use `convert-image-encoder-to-gguf.py` to convert the LLaVA image encoder to GGUF:
+
+```sh
+python ./examples/llava/convert-image-encoder-to-gguf -m ../clip-vit-large-patch14-336 --llava-projector ../llava-v1.5-7b/llava.projector --output-dir ../llava-v1.5-7b
+```
+
+4. Use `convert.py` to convert the LLaMA part of LLaVA to GGUF:
+
+```sh
+python ./convert.py ../llava-v1.5-7b
+```
+
+Now both the LLaMA part and the image encoder is in the `llava-v1.5-7b` directory.
+
+## TODO
+
+- [ ] Support server mode.
+- [ ] Support non-CPU backend for the image encoding part.
+- [ ] Support different sampling methods.
+- [ ] Support more model variants.

examples/llava/clip.h

@@ -0,0 +1,73 @@
+#ifndef CLIP_H
+#define CLIP_H
+
+#include "ggml.h"
+
+struct clip_ctx;
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+struct clip_vision_hparams {
+    int32_t image_size;
+    int32_t patch_size;
+    int32_t hidden_size;
+    int32_t n_intermediate;
+    int32_t projection_dim;
+    int32_t n_head;
+    int32_t n_layer;
+    float eps;
+};
+
+struct clip_ctx * clip_model_load(const char * fname, const int verbosity);
+
+void clip_free(struct clip_ctx * ctx);
+
+size_t clip_embd_nbytes(struct clip_ctx * ctx);
+int clip_n_patches(struct clip_ctx * ctx);
+int clip_n_mmproj_embd(struct clip_ctx * ctx);
+
+// RGB uint8 image
+struct clip_image_u8 {
+    int nx;
+    int ny;
+    uint8_t * data;
+    size_t size;
+};
+
+// RGB float32 image (NHWC)
+// Memory layout: RGBRGBRGB...
+struct clip_image_f32 {
+    int nx;
+    int ny;
+    float * data;
+    size_t size;
+};
+
+struct clip_image_u8_batch {
+    struct clip_image_u8 * data;
+    size_t size;
+};
+
+struct clip_image_f32_batch {
+    struct clip_image_f32 * data;
+    size_t size;
+};
+
+struct clip_image_u8 * make_clip_image_u8();
+struct clip_image_f32 * make_clip_image_f32();
+bool clip_image_load_from_file(const char * fname, struct clip_image_u8 * img);
+bool clip_image_preprocess(const struct clip_ctx * ctx, const struct clip_image_u8 * img, struct clip_image_f32 * res, const bool pad2square);
+bool clip_image_encode(const struct clip_ctx * ctx, const int n_threads, struct clip_image_f32 * img, float * vec);
+
+bool clip_image_batch_encode(const struct clip_ctx * ctx, const int n_threads, const struct clip_image_f32_batch * imgs,
+                             float * vec);
+
+bool clip_model_quantize(const char * fname_inp, const char * fname_out, const int itype);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif // CLIP_H

examples/llava/convert-image-encoder-to-gguf.py

@@ -0,0 +1,250 @@
+import argparse
+import os
+import json
+
+import torch
+import numpy as np
+from gguf import *
+from transformers import CLIPModel, CLIPProcessor
+
+TEXT = "clip.text"
+VISION = "clip.vision"
+
+
+def k(raw_key: str, arch: str) -> str:
+    return raw_key.format(arch=arch)
+
+
+def should_skip_tensor(name: str, has_text: bool, has_vision: bool, has_llava: bool) -> bool:
+    if name in (
+        "logit_scale",
+        "text_model.embeddings.position_ids",
+        "vision_model.embeddings.position_ids",
+    ):
+        return True
+
+    if has_llava and name in ["visual_projection.weight", "vision_model.post_layernorm.weight", "vision_model.post_layernorm.bias"]:
+        return True
+
+    if name.startswith("v") and not has_vision:
+        return True
+
+    if name.startswith("t") and not has_text:
+        return True
+
+    return False
+
+
+def get_tensor_name(name: str) -> str:
+    if "projection" in name:
+        return name
+
+    if "mm_projector" in name:
+        return name.replace("model.mm_projector", "mm")
+
+    return name.replace("text_model", "t").replace("vision_model", "v").replace("encoder.layers", "blk").replace("embeddings.", "").replace("_proj", "").replace("self_attn.", "attn_").replace("layer_norm", "ln").replace("layernorm", "ln").replace("mlp.fc1", "ffn_down").replace("mlp.fc2", "ffn_up").replace("embedding", "embd").replace("final", "post").replace("layrnorm", "ln")
+
+
+def bytes_to_unicode():
+    """
+    Returns list of utf-8 byte and a corresponding list of unicode strings.
+    The reversible bpe codes work on unicode strings.
+    This means you need a large # of unicode characters in your vocab if you want to avoid UNKs.
+    When you're at something like a 10B token dataset you end up needing around 5K for decent coverage.
+    This is a signficant percentage of your normal, say, 32K bpe vocab.
+    To avoid that, we want lookup tables between utf-8 bytes and unicode strings.
+    And avoids mapping to whitespace/control characters the bpe code barfs on.
+    """
+    bs = (
+        list(range(ord("!"), ord("~") + 1))
+        + list(range(ord("¡"), ord("¬") + 1))
+        + list(range(ord("®"), ord("ÿ") + 1))
+    )
+    cs = bs[:]
+    n = 0
+    for b in range(2**8):
+        if b not in bs:
+            bs.append(b)
+            cs.append(2**8 + n)
+            n += 1
+    cs = [chr(n) for n in cs]
+    return dict(zip(bs, cs))
+
+
+ap = argparse.ArgumentParser(prog="convert_hf_to_gguf.py")
+ap.add_argument("-m", "--model-dir", help="Path to model directory cloned from HF Hub", required=True)
+ap.add_argument("--use-f32", action="store_true", default=False, help="Use f32 instead of f16")
+ap.add_argument("--text-only", action="store_true", required=False,
+                help="Save a text-only model. It can't be used to encode images")
+ap.add_argument("--vision-only", action="store_true", required=False,
+                help="Save a vision-only model. It can't be used to encode texts")
+ap.add_argument("--llava-projector", help="Path to llava.projector file. If specified, save an image encoder for LLaVA models.")
+ap.add_argument("--image-mean", nargs=3, type=float, required=False, help="Override image mean values")
+ap.add_argument("--image-std", nargs=3, type=float, required=False, help="Override image std values")
+ap.add_argument("-o", "--output-dir", help="Directory to save GGUF files. Default is the original model directory", default=None)
+
+args = ap.parse_args()
+
+
+if args.text_only and args.vision_only:
+    print("--text-only and --image-only arguments cannot be specified at the same time.")
+    exit(1)
+
+if args.use_f32:
+    print("WARNING: Weights for the convolution op is always saved in f16, as the convolution op in GGML does not support 32-bit kernel weights yet.")
+
+# output in the same directory as the model if output_dir is None
+dir_model = args.model_dir
+
+
+with open(dir_model + "/vocab.json", "r", encoding="utf-8") as f:
+    vocab = json.load(f)
+    tokens = [key for key in vocab]
+
+with open(dir_model + "/config.json", "r", encoding="utf-8") as f:
+    config = json.load(f)
+    v_hparams = config["vision_config"]
+    t_hparams = config["text_config"]
+
+# possible data types
+#   ftype == 0 -> float32
+#   ftype == 1 -> float16
+#
+# map from ftype to string
+ftype_str = ["f32", "f16"]
+
+ftype = 1
+if args.use_f32:
+    ftype = 0
+
+
+model = CLIPModel.from_pretrained(dir_model)
+processor = CLIPProcessor.from_pretrained(dir_model)
+
+fname_middle = None
+has_text_encoder = True
+has_vision_encoder = True
+has_llava_projector = False
+if args.text_only:
+    fname_middle = "text-"
+    has_vision_encoder = False
+elif args.vision_only:
+    fname_middle = "vision-"
+    has_text_encoder = False
+elif args.llava_projector is not None:
+    fname_middle = "mmproj-"
+    has_text_encoder = False
+    has_llava_projector = True
+else:
+    fname_middle = ""
+
+output_dir = args.output_dir if args.output_dir is not None else dir_model
+os.makedirs(output_dir, exist_ok=True)
+output_prefix = os.path.basename(output_dir).replace("ggml_", "")
+fname_out = os.path.join(output_dir, f"{fname_middle}model-{ftype_str[ftype]}.gguf")
+fout = GGUFWriter(path=fname_out, arch="clip")
+
+fout.add_bool("clip.has_text_encoder", has_text_encoder)
+fout.add_bool("clip.has_vision_encoder", has_vision_encoder)
+fout.add_bool("clip.has_llava_projector", has_llava_projector)
+fout.add_file_type(ftype)
+model_name = config["_name_or_path"] if "_name_or_path" in config else os.path.basename(dir_model)
+fout.add_name(model_name)
+if args.text_only:
+    fout.add_description("text-only CLIP model")
+elif args.vision_only and not has_llava_projector:
+    fout.add_description("vision-only CLIP model")
+elif has_llava_projector:
+    fout.add_description("image encoder for LLaVA")
+else:
+    fout.add_description("two-tower CLIP model")
+
+if has_text_encoder:
+    # text_model hparams
+    fout.add_uint32(k(KEY_CONTEXT_LENGTH, TEXT), t_hparams["max_position_embeddings"])
+    fout.add_uint32(k(KEY_EMBEDDING_LENGTH, TEXT), t_hparams["hidden_size"])
+    fout.add_uint32(k(KEY_FEED_FORWARD_LENGTH, TEXT), t_hparams["intermediate_size"])
+    fout.add_uint32("clip.text.projection_dim", t_hparams.get("projection_dim", config["projection_dim"]))
+    fout.add_uint32(k(KEY_ATTENTION_HEAD_COUNT, TEXT), t_hparams["num_attention_heads"])
+    fout.add_float32(k(KEY_ATTENTION_LAYERNORM_EPS, TEXT), t_hparams["layer_norm_eps"])
+    fout.add_uint32(k(KEY_BLOCK_COUNT, TEXT), t_hparams["num_hidden_layers"])
+    fout.add_token_list(tokens)
+
+if has_vision_encoder:
+    # vision_model hparams
+    fout.add_uint32("clip.vision.image_size", v_hparams["image_size"])
+    fout.add_uint32("clip.vision.patch_size", v_hparams["patch_size"])
+    fout.add_uint32(k(KEY_EMBEDDING_LENGTH, VISION), v_hparams["hidden_size"])
+    fout.add_uint32(k(KEY_FEED_FORWARD_LENGTH, VISION), v_hparams["intermediate_size"])
+    fout.add_uint32("clip.vision.projection_dim", v_hparams.get("projection_dim", config["projection_dim"]))
+    fout.add_uint32(k(KEY_ATTENTION_HEAD_COUNT, VISION), v_hparams["num_attention_heads"])
+    fout.add_float32(k(KEY_ATTENTION_LAYERNORM_EPS, VISION), v_hparams["layer_norm_eps"])
+    block_count = v_hparams["num_hidden_layers"] - 1 if has_llava_projector else v_hparams["num_hidden_layers"]
+    fout.add_uint32(k(KEY_BLOCK_COUNT, VISION), block_count)
+
+    image_mean = processor.image_processor.image_mean if args.image_mean is None else args.image_mean
+    image_std = processor.image_processor.image_std if args.image_std is None else args.image_std
+    fout.add_array("clip.vision.image_mean", image_mean)
+    fout.add_array("clip.vision.image_std", image_std)
+
+use_gelu = v_hparams["hidden_act"] == "gelu"
+fout.add_bool("clip.use_gelu", use_gelu)
+
+
+if has_llava_projector:
+    model.vision_model.encoder.layers.pop(-1)
+    projector = torch.load(args.llava_projector)
+    for name, data in projector.items():
+        name = get_tensor_name(name)
+        if data.ndim == 2:
+            data = data.squeeze().numpy().astype(np.float16)
+        else:
+            data = data.squeeze().numpy().astype(np.float32)
+
+        fout.add_tensor(name, data)
+
+    print("Projector tensors added\n")
+
+state_dict = model.state_dict()
+for name, data in state_dict.items():
+    if should_skip_tensor(name, has_text_encoder, has_vision_encoder, has_llava_projector):
+        # we don't need this
+        print(f"skipping parameter: {name}")
+        continue
+
+    name = get_tensor_name(name)
+    data = data.squeeze().numpy()
+
+    n_dims = len(data.shape)
+
+    # ftype == 0 -> float32, ftype == 1 -> float16
+    ftype_cur = 0
+    if n_dims == 4:
+        print(f"tensor {name} is always saved in f16")
+        data = data.astype(np.float16)
+        ftype_cur = 1
+    elif ftype == 1:
+        if name[-7:] == ".weight" and n_dims == 2:
+            print("  Converting to float16")
+            data = data.astype(np.float16)
+            ftype_cur = 1
+        else:
+            print("  Converting to float32")
+            data = data.astype(np.float32)
+            ftype_cur = 0
+    else:
+        if data.dtype != np.float32:
+            print("  Converting to float32")
+            data = data.astype(np.float32)
+            ftype_cur = 0
+
+    print(f"{name} - {ftype_str[ftype_cur]} - shape = {data.shape}")
+    fout.add_tensor(name, data)
+
+
+fout.write_header_to_file()
+fout.write_kv_data_to_file()
+fout.write_tensors_to_file()
+fout.close()
+
+print("Done. Output file: " + fname_out)

examples/llava/llava-surgery.py

@@ -0,0 +1,30 @@
+import argparse
+import glob
+import os
+import torch
+
+
+ap = argparse.ArgumentParser()
+ap.add_argument("-m", "--model", help="Path to LLaVA v1.5 model")
+args = ap.parse_args()
+
+# find the model part that includes the the multimodal projector weights
+path = sorted(glob.glob(f"{args.model}/pytorch_model*.bin"))[-1]
+checkpoint = torch.load(path)
+
+# get a list of mm tensor names
+mm_tensors = [k for k, v in checkpoint.items() if k.startswith("model.mm_projector")]
+
+# store these tensors in a new dictionary and torch.save them
+projector = {name: checkpoint[name] for name in mm_tensors}
+torch.save(projector, f"{args.model}/llava.projector")
+
+# remove these tensors from the checkpoint and save it again
+for name in mm_tensors:
+    del checkpoint[name]
+
+torch.save(checkpoint, path)
+
+print("Done!")
+print(f"Now you can convert {args.model} to a a regular LLaMA GGUF file.")
+print(f"Also, use {args.model}/llava.projector to prepare a llava-encoder.gguf file.")

examples/llava/llava-utils.h

@@ -0,0 +1,145 @@
+#pragma once
+
+// this one and clip lib will be eventually merged to a single lib, let's keep it this way for now
+
+#include "common.h"
+#include "llama.h"
+
+#include <cstdio>
+#include <cstdlib>
+#include <vector>
+
+inline bool eval_image_embd(llama_context * ctx_llama, float * embd, int N, int n_batch, int * n_past) {
+    int n_embd  = llama_n_embd(llama_get_model(ctx_llama));
+
+    for (int i = 0; i < N; i += n_batch) {
+        int n_eval = N - i;
+        if (n_eval > n_batch) {
+            n_eval = n_batch;
+        }
+        llama_batch batch = {int32_t(n_eval), nullptr, (embd+i*n_embd), nullptr, nullptr, nullptr, *n_past, 1, 0, };
+        if (llama_decode(ctx_llama, batch)) {
+            fprintf(stderr, "%s : failed to eval\n", __func__);
+            return false;
+        }
+        *n_past += n_eval;
+    }
+    return true;
+}
+
+inline bool eval_tokens(struct llama_context * ctx_llama, std::vector<llama_token> tokens, int n_batch, int * n_past) {
+    int N = (int) tokens.size();
+    for (int i = 0; i < N; i += n_batch) {
+        int n_eval = (int) tokens.size() - i;
+        if (n_eval > n_batch) {
+            n_eval = n_batch;
+        }
+        if (llama_decode(ctx_llama, llama_batch_get_one(&tokens[i], n_eval, *n_past, 0))) {
+            fprintf(stderr, "%s : failed to eval\n", __func__);
+            return false;
+        }
+        *n_past += n_eval;
+    }
+    return true;
+}
+
+inline bool eval_id(struct llama_context * ctx_llama, int id, int * n_past) {
+    std::vector<llama_token> tokens;
+    tokens.push_back(id);
+    return eval_tokens(ctx_llama, tokens, 1, n_past);
+}
+
+inline bool eval_string(struct llama_context * ctx_llama, const char* str, int n_batch, int * n_past){
+    std::string              str2     = str;
+    std::vector<llama_token> embd_inp = ::llama_tokenize(ctx_llama, str2, true);
+    eval_tokens(ctx_llama, embd_inp, n_batch, n_past);
+    return true;
+}
+
+// TODO: use common/sampling.h
+inline llama_token sample_id(llama_context * ctx_llama, gpt_params & params) {
+      // out of user input, sample next token
+    const float   temp      = params.sampling_params.temp;
+    const int32_t top_k     = params.sampling_params.top_k <= 0 ? llama_n_vocab(llama_get_model(ctx_llama)) : params.sampling_params.top_k;
+    const float   top_p     = params.sampling_params.top_p;
+    const float   tfs_z     = params.sampling_params.tfs_z;
+    const float   typical_p = params.sampling_params.typical_p;
+      // const int32_t repeat_last_n   = params.sampling_params.repeat_last_n < 0 ? n_ctx : params.sampling_params.repeat_last_n;
+      // const float   repeat_penalty  = params.sampling_params.repeat_penalty;
+      // const float   alpha_presence  = params.sampling_params.presence_penalty;
+      // const float   alpha_frequency = params.sampling_params.frequency_penalty;
+    const int     mirostat     = params.sampling_params.mirostat;
+    const float   mirostat_tau = params.sampling_params.mirostat_tau;
+    const float   mirostat_eta = params.sampling_params.mirostat_eta;
+      // const bool    penalize_nl     = params.sampling_params.penalize_nl;
+
+    llama_token id = 0;
+    {
+        auto logits  = llama_get_logits(ctx_llama);
+        auto n_vocab = llama_n_vocab(llama_get_model(ctx_llama));
+
+          // Apply params.logit_bias map
+        for (auto it = params.sampling_params.logit_bias.begin(); it != params.sampling_params.logit_bias.end(); it++) {
+            logits[it->first] += it->second;
+        }
+
+        std::vector<llama_token_data> candidates;
+        candidates.reserve(n_vocab);
+        for (llama_token token_id = 0; token_id < n_vocab; token_id++) {
+            candidates.emplace_back(llama_token_data{token_id, logits[token_id], 0.0f});
+        }
+
+        llama_token_data_array candidates_p = { candidates.data(), candidates.size(), false };
+
+          // TODO: Apply penalties
+          // float nl_logit = logits[llama_token_nl(ctx)];
+          // auto last_n_repeat = std::min(std::min((int)last_n_tokens.size(), repeat_last_n), n_ctx);
+          // llama_sample_repetition_penalty(ctx, &candidates_p,
+          //      last_n_tokens.data() + last_n_tokens.size() - last_n_repeat,
+          //      last_n_repeat, repeat_penalty);
+          // llama_sample_frequency_and_presence_penalties(ctx, &candidates_p,
+          // last_n_tokens.data() + last_n_tokens.size() - last_n_repeat,
+          // last_n_repeat, alpha_frequency, alpha_presence);
+          // if (!penalize_nl) {
+          //     logits[llama_token_nl(ctx)] = nl_logit;
+          // }
+
+        if (temp <= 0) {
+              // Greedy sampling
+            id = llama_sample_token_greedy(ctx_llama, &candidates_p);
+        } else {
+            if (mirostat == 1) {
+                static float mirostat_mu = 2.0f * mirostat_tau;
+                const  int mirostat_m    = 100;
+                llama_sample_temp(ctx_llama, &candidates_p, temp);
+                id = llama_sample_token_mirostat(ctx_llama, &candidates_p, mirostat_tau, mirostat_eta, mirostat_m, &mirostat_mu);
+            } else if (mirostat == 2) {
+                static float mirostat_mu = 2.0f * mirostat_tau;
+                llama_sample_temp(ctx_llama, &candidates_p, temp);
+                id = llama_sample_token_mirostat_v2(ctx_llama, &candidates_p, mirostat_tau, mirostat_eta, &mirostat_mu);
+            } else {
+                  // Temperature sampling
+                llama_sample_top_k(ctx_llama, &candidates_p, top_k, 1);
+                llama_sample_tail_free(ctx_llama, &candidates_p, tfs_z, 1);
+                llama_sample_typical(ctx_llama, &candidates_p, typical_p, 1);
+                llama_sample_top_p(ctx_llama, &candidates_p, top_p, 1);
+                llama_sample_temp(ctx_llama, &candidates_p, temp);
+                id = llama_sample_token(ctx_llama, &candidates_p);
+            }
+        }
+    }
+
+    return id;
+}
+
+inline const char * sample(struct llama_context * ctx_llama, gpt_params & params, int * n_past) {
+    int id = sample_id(ctx_llama, params);
+    static std::string ret;
+    if (id == llama_token_eos(ctx_llama)) {
+        ret = "</s>";
+    } else {
+        ret = llama_token_to_piece(ctx_llama, id);
+    }
+    eval_id(ctx_llama, id, n_past);
+    return ret.c_str();
+}

examples/llava/llava.cpp

@@ -0,0 +1,156 @@
+#include "clip.h"
+#include "llava-utils.h"
+#include "common.h"
+#include "llama.h"
+
+#include <cstdio>
+#include <cstdlib>
+#include <vector>
+
+static void show_additional_info(int /*argc*/, char ** argv) {
+    printf("\n example usage: %s -m <llava-v1.5-7b/ggml-model-q5_k.gguf> --mmproj <llava-v1.5-7b/mmproj-model-f16.gguf> --image <path/to/an/image.jpg> [--temp 0.1] [-p \"describe the image in detail.\"]\n", argv[0]);
+    printf("  note: a lower temperature value like 0.1 is recommended for better quality.\n");
+}
+
+int main(int argc, char ** argv) {
+    ggml_time_init();
+
+    gpt_params params;
+
+    if (!gpt_params_parse(argc, argv, params)) {
+        show_additional_info(argc, argv);
+        return 1;
+    }
+
+    if (params.mmproj.empty() || params.image.empty()) {
+        gpt_print_usage(argc, argv, params);
+        show_additional_info(argc, argv);
+        return 1;
+    }
+
+    const char * clip_path = params.mmproj.c_str();
+    const char * img_path = params.image.c_str();
+
+    if (params.prompt.empty()) {
+        params.prompt = "describe the image in detail.";
+    }
+
+    auto ctx_clip = clip_model_load(clip_path, /*verbosity=*/ 1);
+
+    // load and preprocess the image
+    clip_image_u8 img;
+    clip_image_f32 img_res;
+
+    if (!clip_image_load_from_file(img_path, &img)) {
+        fprintf(stderr, "%s: is %s really an image file?\n", __func__, img_path);
+
+        clip_free(ctx_clip);
+        return 1;
+    }
+
+    if (!clip_image_preprocess(ctx_clip, &img, &img_res, /*pad2square =*/ true)) {
+        fprintf(stderr, "%s: unable to preprocess %s\n", __func__, img_path);
+
+        clip_free(ctx_clip);
+        return 1;
+    }
+
+    int n_img_pos  = clip_n_patches(ctx_clip);
+    int n_img_embd = clip_n_mmproj_embd(ctx_clip);
+
+    float * image_embd = (float *)malloc(clip_embd_nbytes(ctx_clip));
+
+    if (!image_embd) {
+        fprintf(stderr, "Unable to allocate memory for image embeddings\n");
+
+        return 1;
+    }
+
+    const int64_t t_img_enc_start_us = ggml_time_us();
+    if (!clip_image_encode(ctx_clip, params.n_threads, &img_res, image_embd)) {
+        fprintf(stderr, "Unable to encode image\n");
+
+        return 1;
+    }
+    const int64_t t_img_enc_end_us = ggml_time_us();
+
+    // we get the embeddings, free up the memory required for CLIP
+    clip_free(ctx_clip);
+
+    llama_backend_init(params.numa);
+
+    llama_model_params model_params = llama_model_default_params();
+    llama_model * model = llama_load_model_from_file(params.model.c_str(), model_params);
+    if (model == NULL) {
+        fprintf(stderr , "%s: error: unable to load model\n" , __func__);
+        return 1;
+    }
+
+    llama_context_params ctx_params = llama_context_default_params();
+
+    ctx_params.n_ctx           = params.n_ctx < 2048 ? 2048 : params.n_ctx; // we need a longer context size to process image embeddings
+    ctx_params.n_threads       = params.n_threads;
+    ctx_params.n_threads_batch = params.n_threads_batch == -1 ? params.n_threads : params.n_threads_batch;
+
+    llama_context * ctx_llama = llama_new_context_with_model(model, ctx_params);
+
+    if (ctx_llama == NULL) {
+        fprintf(stderr , "%s: error: failed to create the llama_context\n" , __func__);
+        return 1;
+    }
+
+    // make sure that the correct mmproj was used, i.e., compare apples to apples
+    int n_llama_embd = llama_n_embd(llama_get_model(ctx_llama));
+    if (n_img_embd != n_llama_embd) {
+        printf("%s: embedding dim of the multimodal projector (%d) is not equal to that of LLaMA (%d). Make sure that you use the correct mmproj file.\n", __func__, n_img_embd, n_llama_embd);
+
+        llama_free(ctx_llama);
+        llama_free_model(model);
+        llama_backend_free();
+        free(image_embd);
+
+        return 1;
+    }
+
+    // process the prompt
+    // llava chat format is "<system_prompt>USER: <image_embeddings>\n<textual_prompt>\nASSISTANT:"
+
+    int n_past = 0;
+
+    const int max_tgt_len = params.n_predict < 0 ? 256 : params.n_predict;
+
+    // GG: are we sure that the should be a trailing whitespace at the end of this string?
+    eval_string(ctx_llama, "A chat between a curious human and an artificial intelligence assistant.  The assistant gives helpful, detailed, and polite answers to the human's questions.\nUSER: ", params.n_batch, &n_past);
+    eval_image_embd(ctx_llama, image_embd, n_img_pos, params.n_batch, &n_past);
+    eval_string(ctx_llama, params.prompt.c_str(), params.n_batch, &n_past);
+    eval_string(ctx_llama, "\nASSISTANT:",        params.n_batch, &n_past);
+
+    // generate the response
+
+    printf("\n");
+
+    for (int i = 0; i < max_tgt_len; i++) {
+        const char * tmp = sample(ctx_llama, params, &n_past);
+        if (strcmp(tmp, "</s>") == 0) break;
+
+        printf("%s", tmp);
+        fflush(stdout);
+    }
+
+    printf("\n");
+
+    {
+        const float t_img_enc_ms = (t_img_enc_end_us - t_img_enc_start_us) / 1000.0;
+
+        printf("\n%s: image encoded in %8.2f ms by CLIP (%8.2f ms per image patch)\n", __func__, t_img_enc_ms, t_img_enc_ms / n_img_pos);
+    }
+
+    llama_print_timings(ctx_llama);
+
+    llama_free(ctx_llama);
+    llama_free_model(model);
+    llama_backend_free();
+    free(image_embd);
+
+    return 0;
+}

examples/main/main.cpp

@@ -109,6 +109,7 @@ int main(int argc, char ** argv) {
     if (!gpt_params_parse(argc, argv, params)) {
         return 1;
     }
+    llama_sampling_params & sparams = params.sampling_params;
 
 #ifndef LOG_DISABLE_LOGS
     log_set_target(log_filename_generator("main", "log"));
@@ -179,7 +180,7 @@ int main(int argc, char ** argv) {
     // load the model and apply lora adapter, if any
     LOG("%s: load the model and apply lora adapter, if any\n", __func__);
     std::tie(model, ctx) = llama_init_from_gpt_params(params);
-    if (params.cfg_scale > 1.f) {
+    if (sparams.cfg_scale > 1.f) {
         struct llama_context_params lparams = llama_context_params_from_gpt_params(params);
         ctx_guidance = llama_new_context_with_model(model, lparams);
     }
@@ -257,9 +258,9 @@ int main(int argc, char ** argv) {
     int guidance_offset = 0;
     int original_prompt_len = 0;
     if (ctx_guidance) {
-        LOG("cfg_negative_prompt: \"%s\"\n", log_tostr(params.cfg_negative_prompt));
+        LOG("cfg_negative_prompt: \"%s\"\n", log_tostr(sparams.cfg_negative_prompt));
 
-        guidance_inp = ::llama_tokenize(ctx_guidance, params.cfg_negative_prompt, add_bos);
+        guidance_inp = ::llama_tokenize(ctx_guidance, sparams.cfg_negative_prompt, add_bos);
         LOG("guidance_inp tokenized: %s\n", LOG_TOKENS_TOSTR_PRETTY(ctx_guidance, guidance_inp));
 
         std::vector<llama_token> original_inp = ::llama_tokenize(ctx, params.prompt, add_bos);
@@ -296,6 +297,9 @@ int main(int argc, char ** argv) {
             LOG_TEE("%s: session file matches %zu / %zu tokens of prompt\n",
                 __func__, n_matching_session_tokens, embd_inp.size());
         }
+
+        // remove any "future" tokens that we might have inherited from the previous session
+        llama_kv_cache_tokens_rm(ctx, n_matching_session_tokens, -1);
     }
 
     LOGLN(
@@ -343,7 +347,7 @@ int main(int argc, char ** argv) {
 
         if (ctx_guidance) {
             LOG_TEE("\n");
-            LOG_TEE("%s: negative prompt: '%s'\n", __func__, params.cfg_negative_prompt.c_str());
+            LOG_TEE("%s: negative prompt: '%s'\n", __func__, sparams.cfg_negative_prompt.c_str());
             LOG_TEE("%s: number of tokens in negative prompt = %zu\n", __func__, guidance_inp.size());
             for (int i = 0; i < (int) guidance_inp.size(); i++) {
                 LOG_TEE("%6d -> '%s'\n", guidance_inp[i], llama_token_to_piece(ctx, guidance_inp[i]).c_str());
@@ -395,7 +399,7 @@ int main(int argc, char ** argv) {
         }
     }
     LOG_TEE("sampling: repeat_last_n = %d, repeat_penalty = %f, presence_penalty = %f, frequency_penalty = %f, top_k = %d, tfs_z = %f, top_p = %f, typical_p = %f, temp = %f, mirostat = %d, mirostat_lr = %f, mirostat_ent = %f\n",
-            params.repeat_last_n, params.repeat_penalty, params.presence_penalty, params.frequency_penalty, params.top_k, params.tfs_z, params.top_p, params.typical_p, params.temp, params.mirostat, params.mirostat_eta, params.mirostat_tau);
+            sparams.repeat_last_n, sparams.repeat_penalty, sparams.presence_penalty, sparams.frequency_penalty, sparams.top_k, sparams.tfs_z, sparams.top_p, sparams.typical_p, sparams.temp, sparams.mirostat, sparams.mirostat_eta, sparams.mirostat_tau);
     LOG_TEE("generate: n_ctx = %d, n_batch = %d, n_predict = %d, n_keep = %d\n", n_ctx, params.n_batch, params.n_predict, params.n_keep);
     LOG_TEE("\n\n");
 
@@ -413,8 +417,8 @@ int main(int argc, char ** argv) {
         LOG_TEE("\n");
 
         {
-            auto it = params.logit_bias.find(llama_token_eos(ctx));
-            if (it != params.logit_bias.end() && it->second == -INFINITY) {
+            auto it = sparams.logit_bias.find(llama_token_eos(ctx));
+            if (it != sparams.logit_bias.end() && it->second == -INFINITY) {
                 LOG_TEE("%s: warning: EOS token is disabled, which will cause most grammars to fail\n", __func__);
             }
         }
@@ -469,6 +473,7 @@ int main(int argc, char ** argv) {
 
     const int n_vocab = llama_n_vocab(model);
 
+    llama_sampling_context ctx_sampling = llama_sampling_context_init(params, grammar);
     std::vector<llama_token_data> candidates;
     candidates.reserve(n_vocab);
 
@@ -543,9 +548,6 @@ int main(int argc, char ** argv) {
                 if (i > 0) {
                     embd.erase(embd.begin(), embd.begin() + i);
                 }
-
-                // remove any "future" tokens that we might have inherited from the session from the KV cache
-                llama_kv_cache_tokens_rm(ctx, n_past, -1);
             }
 
             // evaluate tokens in batches
@@ -625,7 +627,7 @@ int main(int argc, char ** argv) {
                 LOG("saved session to %s\n", path_session.c_str());
             }
 
-            const llama_token id = llama_sample_token(ctx, ctx_guidance, grammar, params, last_tokens, candidates);
+            const llama_token id = llama_sampling_sample(ctx, ctx_guidance, ctx_sampling, last_tokens, candidates);
 
             last_tokens.erase(last_tokens.begin());
             last_tokens.push_back(id);

examples/parallel/parallel.cpp

@@ -125,6 +125,8 @@ int main(int argc, char ** argv) {
     params.logits_all = true;
     std::tie(model, ctx) = llama_init_from_gpt_params(params);
 
+    llama_sampling_context ctx_sampling = llama_sampling_context_init(params, NULL);
+
     // load the prompts from an external file if there are any
     if (params.prompt.empty()) {
         printf("\n\033[32mNo new questions so proceed with build-in defaults.\033[0m\n");
@@ -167,7 +169,7 @@ int main(int argc, char ** argv) {
 
     // the max batch size is as large as the context to handle cases where we get very long input prompt from multiple
     // users. regardless of the size, the main loop will chunk the batch into a maximum of params.n_batch tokens at a time
-    llama_batch batch = llama_batch_init(params.n_ctx, 0);
+    llama_batch batch = llama_batch_init(n_ctx, 0);
 
     int32_t n_total_prompt = 0;
     int32_t n_total_gen    = 0;
@@ -339,7 +341,7 @@ int main(int argc, char ** argv) {
                 //printf("client %d, seq %d, token %d, pos %d, batch %d\n",
                 //        client.id, client.seq_id, client.sampled, client.n_decoded, client.i_batch);
 
-                const llama_token id = llama_sample_token(ctx, NULL, NULL, params, client.tokens_prev, candidates, client.i_batch - i);
+                const llama_token id = llama_sampling_sample(ctx, NULL, ctx_sampling, client.tokens_prev, candidates, client.i_batch - i, client.seq_id);
 
                 if (client.n_decoded == 1) {
                     // start measuring generation time after the first token to make sure all concurrent clients
@@ -384,7 +386,7 @@ int main(int argc, char ** argv) {
 
                     n_total_prompt += client.n_prompt;
                     n_total_gen    += client.n_decoded;
-
+                    llama_sampling_context_reset(ctx_sampling, client.seq_id);
                     client.seq_id = -1;
                 }
 

examples/save-load-state/save-load-state.cpp

@@ -8,9 +8,10 @@
 
 int main(int argc, char ** argv) {
     gpt_params params;
+    llama_sampling_params & sparams = params.sampling_params;
     params.seed = 42;
     params.n_threads = 4;
-    params.repeat_last_n = 64;
+    sparams.repeat_last_n = 64;
     params.prompt = "The quick brown fox";
 
     if (!gpt_params_parse(argc, argv, params)) {
@@ -24,7 +25,7 @@ int main(int argc, char ** argv) {
     }
 
     auto n_past = 0;
-    auto last_n_tokens_data = std::vector<llama_token>(params.repeat_last_n, 0);
+    auto last_n_tokens_data = std::vector<llama_token>(sparams.repeat_last_n, 0);
 
     // init
     llama_model * model;

examples/server/api_like_OAI.py

@@ -27,10 +27,10 @@ def is_present(json, key):
         buf = json[key]
     except KeyError:
         return False
+    if json[key] == None:
+        return False
     return True
 
-
-
 #convert chat to prompt
 def convert_chat(messages):
     prompt = "" + args.chat_prompt.replace("\\n", "\n")

examples/server/public/index.html

@@ -136,6 +136,11 @@
       display: block;
     }
 
+    fieldset label.slim {
+      margin: 0 0.5em;
+      display: inline;
+    }
+
     header, footer {
       text-align: center;
     }
@@ -145,6 +150,14 @@
       color: #888;
     }
 
+    .mode-chat textarea[name=prompt] {
+      height: 4.5em;
+    }
+
+    .mode-completion textarea[name=prompt] {
+      height: 10em;
+    }
+
 
     @keyframes loading-bg-wipe {
       0% {
@@ -187,7 +200,7 @@
       template: "{{prompt}}\n\n{{history}}\n{{char}}:",
       historyTemplate: "{{name}}: {{message}}",
       transcript: [],
-      type: "chat",
+      type: "chat",  // "chat" | "completion"
       char: "Llama",
       user: "User",
     })
@@ -365,13 +378,44 @@
       return String(str).replaceAll(/\{\{(.*?)\}\}/g, (_, key) => template(settings[key]));
     }
 
+    async function runLlama(prompt, llamaParams, char) {
+      const currentMessages = [];
+      const history = session.value.transcript;
+      if (controller.value) {
+        throw new Error("already running");
+      }
+      controller.value = new AbortController();
+      for await (const chunk of llama(prompt, llamaParams, {controller: controller.value})) {
+        const data = chunk.data;
+
+        if (data.stop) {
+          while (
+            currentMessages.length > 0 &&
+            currentMessages[currentMessages.length - 1].content.match(/\n$/) != null
+            ) {
+            currentMessages.pop();
+          }
+          transcriptUpdate([...history, [char, currentMessages]])
+          console.log("Completion finished: '", currentMessages.map(msg => msg.content).join(''), "', summary: ", data);
+        } else {
+          currentMessages.push(data);
+          transcriptUpdate([...history, [char, currentMessages]])
+        }
+
+        if (data.timings) {
+          llamaStats.value = data.timings;
+        }
+      }
+
+      controller.value = null;
+    }
+
     // send message to server
     const chat = async (msg) => {
       if (controller.value) {
         console.log('already running...');
         return;
       }
-      controller.value = new AbortController();
 
       transcriptUpdate([...session.value.transcript, ["{{user}}", msg]])
 
@@ -391,42 +435,25 @@
         ).join("\n"),
       });
 
-      const currentMessages = [];
-      const history = session.value.transcript
-
-      const llamaParams = {
+      await runLlama(prompt, {
         ...params.value,
         stop: ["</s>", template("{{char}}:"), template("{{user}}:")],
+      }, "{{char}}");
     }
 
-      for await (const chunk of llama(prompt, llamaParams, { controller: controller.value })) {
-        const data = chunk.data;
-
-        if (data.stop) {
-          while (
-            currentMessages.length > 0 &&
-            currentMessages[currentMessages.length - 1].content.match(/\n$/) != null
-          ) {
-            currentMessages.pop();
+    const runCompletion = async () => {
+      if (controller.value) {
+        console.log('already running...');
+        return;
       }
-          transcriptUpdate([...history, ["{{char}}", currentMessages]])
-          console.log("Completion finished: '", currentMessages.map(msg => msg.content).join(''), "', summary: ", data);
-        } else {
-          currentMessages.push(data);
-          transcriptUpdate([...history, ["{{char}}", currentMessages]])
+      const {prompt} = session.value;
+      transcriptUpdate([...session.value.transcript, ["", prompt]]);
+      await runLlama(prompt, {
+        ...params.value,
+        stop: [],
+      }, "");
     }
 
-        if (data.timings) {
-          llamaStats.value = data.timings;
-        }
-      }
-
-      controller.value = null;
-    }
-
-    function MessageInput() {
-      const message = useSignal("")
-
     const stop = (e) => {
       e.preventDefault();
       if (controller.value) {
@@ -440,6 +467,9 @@
       transcriptUpdate([]);
     }
 
+    function MessageInput() {
+      const message = useSignal("")
+
       const submit = (e) => {
         stop(e);
         chat(message.value);
@@ -474,6 +504,19 @@
       `
     }
 
+    function CompletionControls() {
+      const submit = (e) => {
+        stop(e);
+        runCompletion();
+      }
+      return html`
+        <div>
+          <button onclick=${submit} type="button" disabled=${generating.value}>Start</button>
+          <button onclick=${stop} disabled=${!generating.value}>Stop</button>
+          <button onclick=${reset}>Reset</button>
+        </div>`;
+    }
+
     const ChatLog = (props) => {
       const messages = session.value.transcript;
       const container = useRef(null)
@@ -497,7 +540,11 @@
             data;
           message = html`<${Markdownish} text=${template(text)} />`
         }
+        if(user) {
           return html`<p key=${index}><strong>${template(user)}:</strong> ${message}</p>`
+        } else {
+          return html`<p key=${index}>${message}</p>`
+        }
       };
 
       return html`
@@ -574,18 +621,31 @@
         userTemplateAutosave()
       }, [session.value, params.value])
 
-      return html`
-        <form>
-          <fieldset>
-            <${UserTemplateResetButton}/>
-          </fieldset>
+      const GrammarControl = () => (
+        html`
+          <div>
+            <label for="template">Grammar</label>
+            <textarea id="grammar" name="grammar" placeholder="Use gbnf or JSON Schema+convert" value="${params.value.grammar}" rows=4 oninput=${updateParams}/>
+            <input type="text" name="prop-order" placeholder="order: prop1,prop2,prop3" oninput=${updateGrammarJsonSchemaPropOrder} />
+            <button type="button" onclick=${convertJSONSchemaGrammar}>Convert JSON Schema</button>
+          </div>
+          `
+      );
 
+      const PromptControlFieldSet = () => (
+        html`
         <fieldset>
           <div>
-              <label for="prompt">Prompt</label>
-              <textarea type="text" name="prompt" value="${session.value.prompt}" rows=4 oninput=${updateSession}/>
+            <label htmlFor="prompt">Prompt</label>
+            <textarea type="text" name="prompt" value="${session.value.prompt}" oninput=${updateSession}/>
           </div>
         </fieldset>
+        `
+      );
+
+      const ChatConfigForm = () => (
+        html`
+          ${PromptControlFieldSet()}
 
           <fieldset class="two">
             <div>
@@ -609,15 +669,30 @@
               <label for="template">Chat history template</label>
               <textarea id="template" name="historyTemplate" value="${session.value.historyTemplate}" rows=1 oninput=${updateSession}/>
             </div>
+            ${GrammarControl()}
+          </fieldset>
+      `
+    );
 
+      const CompletionConfigForm = () => (
+        html`
+          ${PromptControlFieldSet()}
+          <fieldset>${GrammarControl()}</fieldset>
+        `
+      );
+
+      return html`
+        <form>
+          <fieldset class="two">
+            <${UserTemplateResetButton}/>
             <div>
-              <label for="template">Grammar</label>
-              <textarea id="grammar" name="grammar" placeholder="Use gbnf or JSON Schema+convert" value="${params.value.grammar}" rows=4 oninput=${updateParams}/>
-              <input type="text" name="prop-order" placeholder="order: prop1,prop2,prop3" oninput=${updateGrammarJsonSchemaPropOrder} />
-              <button type="button" onclick=${convertJSONSchemaGrammar}>Convert JSON Schema</button>
+              <label class="slim"><input type="radio" name="type" value="chat" checked=${session.value.type === "chat"} oninput=${updateSession} /> Chat</label>
+              <label class="slim"><input type="radio" name="type" value="completion" checked=${session.value.type === "completion"} oninput=${updateSession} /> Completion</label>
             </div>
           </fieldset>
 
+          ${session.value.type === 'chat' ? ChatConfigForm() : CompletionConfigForm()}
+
           <fieldset class="two">
             ${IntField({label: "Predictions", max: 2048, min: -1, name: "n_predict", value: params.value.n_predict})}
             ${FloatField({label: "Temperature", max: 1.5, min: 0.0, name: "temperature", step: 0.01, value: params.value.temperature})}
@@ -851,7 +926,7 @@
     function App(props) {
 
       return html`
-        <div>
+        <div class="mode-${session.value.type}">
           <header>
             <h1>llama.cpp</h1>
           </header>
@@ -861,7 +936,7 @@
           </main>
 
           <section id="write">
-            <${MessageInput} />
+            <${session.value.type === 'chat' ? MessageInput : CompletionControls} />
           </section>
 
           <footer>

examples/server/server.cpp

@@ -200,6 +200,7 @@ struct llama_server_context
     llama_model *model = nullptr;
     llama_context *ctx = nullptr;
     gpt_params params;
+    llama_sampling_context ctx_sampling;
     int n_ctx;
 
     grammar_parser::parse_state parsed_grammar;
@@ -254,6 +255,7 @@ struct llama_server_context
         if (grammar != nullptr) {
             llama_grammar_free(grammar);
             grammar = nullptr;
+            ctx_sampling = llama_sampling_context_init(params, NULL);
         }
     }
 
@@ -329,8 +331,8 @@ struct llama_server_context
             grammar_parser::print_grammar(stderr, parsed_grammar);
 
             {
-                auto it = params.logit_bias.find(llama_token_eos(ctx));
-                if (it != params.logit_bias.end() && it->second == -INFINITY) {
+                auto it = params.sampling_params.logit_bias.find(llama_token_eos(ctx));
+                if (it != params.sampling_params.logit_bias.end() && it->second == -INFINITY) {
                     LOG_WARNING("EOS token is disabled, which will cause most grammars to fail", {});
                 }
             }
@@ -339,14 +341,26 @@ struct llama_server_context
             grammar = llama_grammar_init(
                 grammar_rules.data(), grammar_rules.size(), parsed_grammar.symbol_ids.at("root"));
         }
+        ctx_sampling = llama_sampling_context_init(params, grammar);
         return true;
     }
 
     void loadInfill()
     {
-        auto prefix_tokens = tokenize(params.input_prefix, true);  // always add BOS
-        auto suffix_tokens = tokenize(params.input_suffix, true);  // always add BOS
+        bool suff_rm_leading_spc = true;
+        if (params.input_suffix.find_first_of(" ") == 0 && params.input_suffix.size() > 1) {
+            params.input_suffix.erase(0, 1);
+            suff_rm_leading_spc = false;
+        }
+
+        auto prefix_tokens = tokenize(params.input_prefix, false);
+        auto suffix_tokens = tokenize(params.input_suffix, false);
+        const int space_token = 29871;
+        if (suff_rm_leading_spc  && suffix_tokens[0] == space_token) {
+            suffix_tokens.erase(suffix_tokens.begin());
+        }
         prefix_tokens.insert(prefix_tokens.begin(), llama_token_prefix(ctx));
+        prefix_tokens.insert(prefix_tokens.begin(), llama_token_bos(ctx)); // always add BOS
         prefix_tokens.insert(prefix_tokens.end(), llama_token_suffix(ctx));
         prefix_tokens.insert(prefix_tokens.end(), suffix_tokens.begin(), suffix_tokens.end());
         prefix_tokens.push_back(llama_token_middle(ctx));
@@ -391,6 +405,7 @@ struct llama_server_context
         // compare the evaluated prompt with the new prompt
         n_past = common_part(embd, prompt_tokens);
         embd = prompt_tokens;
+
         if (n_past == num_prompt_tokens)
         {
             // we have to evaluate at least 1 token to generate logits.
@@ -398,6 +413,9 @@ struct llama_server_context
             n_past--;
         }
 
+        // since #3228 we now have to manually manage the KV cache
+        llama_kv_cache_seq_rm(ctx, 0, n_past, -1);
+
         LOG_VERBOSE("prompt ingested", {
                                            {"n_past", n_past},
                                            {"cached", tokens_to_str(ctx, embd.cbegin(), embd.cbegin() + n_past)},
@@ -447,9 +465,6 @@ struct llama_server_context
         // compare the evaluated prompt with the new prompt
         n_past = common_part(embd, prompt_tokens);
 
-        // since #3228 we now have to manually manage the KV cache
-        llama_kv_cache_seq_rm(ctx, 0, n_past, -1);
-
         embd = prompt_tokens;
         if (n_past == num_prompt_tokens)
         {
@@ -457,6 +472,9 @@ struct llama_server_context
             n_past--;
         }
 
+        // since #3228 we now have to manually manage the KV cache
+        llama_kv_cache_seq_rm(ctx, 0, n_past, -1);
+
         LOG_VERBOSE("prompt ingested", {
                                            {"n_past", n_past},
                                            {"cached", tokens_to_str(ctx, embd.cbegin(), embd.cbegin() + n_past)},
@@ -539,12 +557,12 @@ struct llama_server_context
             std::vector<llama_token_data> candidates;
             candidates.reserve(llama_n_vocab(model));
 
-            result.tok = llama_sample_token(ctx, NULL, grammar, params, last_n_tokens, candidates);
+            result.tok = llama_sampling_sample(ctx, NULL, ctx_sampling, last_n_tokens, candidates);
 
             llama_token_data_array candidates_p = { candidates.data(), candidates.size(), false };
 
-            const int32_t n_probs = params.n_probs;
-            if (params.temp <= 0 && n_probs > 0)
+            const int32_t n_probs = params.sampling_params.n_probs;
+            if (params.sampling_params.temp <= 0 && n_probs > 0)
             {
                 // For llama_sample_token_greedy we need to sort candidates
                 llama_sample_softmax(ctx, &candidates_p);
@@ -619,7 +637,7 @@ struct llama_server_context
         const std::string token_text = token_with_probs.tok == -1 ? "" : llama_token_to_piece(ctx, token_with_probs.tok);
         generated_text += token_text;
 
-        if (params.n_probs > 0)
+        if (params.sampling_params.n_probs > 0)
         {
             generated_token_probs.push_back(token_with_probs);
         }
@@ -703,6 +721,7 @@ static void server_print_usage(const char *argv0, const gpt_params &params,
     printf("  -h, --help                show this help message and exit\n");
     printf("  -v, --verbose             verbose output (default: %s)\n", server_verbose ? "enabled" : "disabled");
     printf("  -t N, --threads N         number of threads to use during computation (default: %d)\n", params.n_threads);
+    printf("  -tb N, --threads-batch N  number of threads to use during batch and prompt processing (default: same as --threads)\n");
     printf("  -c N, --ctx-size N        size of the prompt context (default: %d)\n", params.n_ctx);
     printf("  --rope-scaling {none,linear,yarn}\n");
     printf("                            RoPE frequency scaling method, defaults to linear unless specified by the model\n");
@@ -904,6 +923,15 @@ static void server_params_parse(int argc, char **argv, server_params &sparams,
             }
             params.n_threads = std::stoi(argv[i]);
         }
+        else if (arg == "--threads-batch" || arg == "-tb")
+        {
+            if (++i >= argc)
+            {
+                invalid_param = true;
+                break;
+            }
+            params.n_threads_batch = std::stoi(argv[i]);
+        }
         else if (arg == "-b" || arg == "--batch-size")
         {
             if (++i >= argc)
@@ -1058,34 +1086,35 @@ static void server_params_parse(int argc, char **argv, server_params &sparams,
 
 static json format_generation_settings(llama_server_context &llama)
 {
-    const auto eos_bias = llama.params.logit_bias.find(llama_token_eos(llama.ctx));
-    const bool ignore_eos = eos_bias != llama.params.logit_bias.end() &&
+    const auto & sparams = llama.params.sampling_params;
+    const auto eos_bias = sparams.logit_bias.find(llama_token_eos(llama.ctx));
+    const bool ignore_eos = eos_bias != sparams.logit_bias.end() &&
                             eos_bias->second < 0.0f && std::isinf(eos_bias->second);
 
     return json{
         {"n_ctx", llama.n_ctx},
         {"model", llama.params.model_alias},
         {"seed", llama.params.seed},
-        {"temp", llama.params.temp},
-        {"top_k", llama.params.top_k},
-        {"top_p", llama.params.top_p},
-        {"tfs_z", llama.params.tfs_z},
-        {"typical_p", llama.params.typical_p},
-        {"repeat_last_n", llama.params.repeat_last_n},
-        {"repeat_penalty", llama.params.repeat_penalty},
-        {"presence_penalty", llama.params.presence_penalty},
-        {"frequency_penalty", llama.params.frequency_penalty},
-        {"mirostat", llama.params.mirostat},
-        {"mirostat_tau", llama.params.mirostat_tau},
-        {"mirostat_eta", llama.params.mirostat_eta},
-        {"penalize_nl", llama.params.penalize_nl},
+        {"temp", sparams.temp},
+        {"top_k", sparams.top_k},
+        {"top_p", sparams.top_p},
+        {"tfs_z", sparams.tfs_z},
+        {"typical_p", sparams.typical_p},
+        {"repeat_last_n", sparams.repeat_last_n},
+        {"repeat_penalty", sparams.repeat_penalty},
+        {"presence_penalty", sparams.presence_penalty},
+        {"frequency_penalty", sparams.frequency_penalty},
+        {"mirostat", sparams.mirostat},
+        {"mirostat_tau", sparams.mirostat_tau},
+        {"mirostat_eta", sparams.mirostat_eta},
+        {"penalize_nl", sparams.penalize_nl},
         {"stop", llama.params.antiprompt},
         {"n_predict", llama.params.n_predict},
         {"n_keep", llama.params.n_keep},
         {"ignore_eos", ignore_eos},
         {"stream", llama.stream},
-        {"logit_bias", llama.params.logit_bias},
-        {"n_probs", llama.params.n_probs},
+        {"logit_bias", sparams.logit_bias},
+        {"n_probs", sparams.n_probs},
         {"grammar", llama.params.grammar},
     };
 }
@@ -1134,7 +1163,7 @@ static json format_final_response(llama_server_context &llama, const std::string
         {"timings", format_timings(llama)},
     };
 
-    if (llama.params.n_probs > 0)
+    if (llama.params.sampling_params.n_probs > 0)
     {
         res["completion_probabilities"] = probs_vector_to_json(llama.ctx, probs);
     }
@@ -1150,7 +1179,7 @@ static json format_partial_response(
         {"stop", false},
     };
 
-    if (llama.params.n_probs > 0)
+    if (llama.params.sampling_params.n_probs > 0)
     {
         res["completion_probabilities"] = probs_vector_to_json(llama.ctx, probs);
     }
@@ -1182,26 +1211,28 @@ static T json_value(const json &body, const std::string &key, const T &default_v
 static void parse_options_completion(const json &body, llama_server_context &llama)
 {
     gpt_params default_params;
+    const auto & default_sparams = default_params.sampling_params;
+    auto & sparams = llama.params.sampling_params;
 
     llama.stream = json_value(body, "stream", false);
     llama.params.n_predict = json_value(body, "n_predict", default_params.n_predict);
-    llama.params.top_k = json_value(body, "top_k", default_params.top_k);
-    llama.params.top_p = json_value(body, "top_p", default_params.top_p);
-    llama.params.tfs_z = json_value(body, "tfs_z", default_params.tfs_z);
-    llama.params.typical_p = json_value(body, "typical_p", default_params.typical_p);
-    llama.params.repeat_last_n = json_value(body, "repeat_last_n", default_params.repeat_last_n);
-    llama.params.temp = json_value(body, "temperature", default_params.temp);
-    llama.params.repeat_penalty = json_value(body, "repeat_penalty", default_params.repeat_penalty);
-    llama.params.presence_penalty = json_value(body, "presence_penalty", default_params.presence_penalty);
-    llama.params.frequency_penalty = json_value(body, "frequency_penalty", default_params.frequency_penalty);
-    llama.params.mirostat = json_value(body, "mirostat", default_params.mirostat);
-    llama.params.mirostat_tau = json_value(body, "mirostat_tau", default_params.mirostat_tau);
-    llama.params.mirostat_eta = json_value(body, "mirostat_eta", default_params.mirostat_eta);
-    llama.params.penalize_nl = json_value(body, "penalize_nl", default_params.penalize_nl);
+    sparams.top_k = json_value(body, "top_k", default_sparams.top_k);
+    sparams.top_p = json_value(body, "top_p", default_sparams.top_p);
+    sparams.tfs_z = json_value(body, "tfs_z", default_sparams.tfs_z);
+    sparams.typical_p = json_value(body, "typical_p", default_sparams.typical_p);
+    sparams.repeat_last_n = json_value(body, "repeat_last_n", default_sparams.repeat_last_n);
+    sparams.temp = json_value(body, "temperature", default_sparams.temp);
+    sparams.repeat_penalty = json_value(body, "repeat_penalty", default_sparams.repeat_penalty);
+    sparams.presence_penalty = json_value(body, "presence_penalty", default_sparams.presence_penalty);
+    sparams.frequency_penalty = json_value(body, "frequency_penalty", default_sparams.frequency_penalty);
+    sparams.mirostat = json_value(body, "mirostat", default_sparams.mirostat);
+    sparams.mirostat_tau = json_value(body, "mirostat_tau", default_sparams.mirostat_tau);
+    sparams.mirostat_eta = json_value(body, "mirostat_eta", default_sparams.mirostat_eta);
+    sparams.penalize_nl = json_value(body, "penalize_nl", default_sparams.penalize_nl);
     llama.params.n_keep = json_value(body, "n_keep", default_params.n_keep);
     llama.params.seed = json_value(body, "seed", default_params.seed);
     llama.params.grammar = json_value(body, "grammar", default_params.grammar);
-    llama.params.n_probs = json_value(body, "n_probs", default_params.n_probs);
+    sparams.n_probs = json_value(body, "n_probs", default_sparams.n_probs);
 
     if (body.count("prompt") != 0)
     {
@@ -1212,10 +1243,10 @@ static void parse_options_completion(const json &body, llama_server_context &lla
         llama.prompt = "";
     }
 
-    llama.params.logit_bias.clear();
+    sparams.logit_bias.clear();
     if (json_value(body, "ignore_eos", false))
     {
-        llama.params.logit_bias[llama_token_eos(llama.ctx)] = -INFINITY;
+        sparams.logit_bias[llama_token_eos(llama.ctx)] = -INFINITY;
     }
 
     const auto &logit_bias = body.find("logit_bias");
@@ -1231,11 +1262,11 @@ static void parse_options_completion(const json &body, llama_server_context &lla
                 {
                     if (el[1].is_number())
                     {
-                        llama.params.logit_bias[tok] = el[1].get<float>();
+                        sparams.logit_bias[tok] = el[1].get<float>();
                     }
                     else if (el[1].is_boolean() && !el[1].get<bool>())
                     {
-                        llama.params.logit_bias[tok] = -INFINITY;
+                        sparams.logit_bias[tok] = -INFINITY;
                     }
                 }
             }
@@ -1255,6 +1286,8 @@ static void parse_options_completion(const json &body, llama_server_context &lla
         }
     }
 
+    llama.ctx_sampling = llama_sampling_context_init(llama.params, llama.grammar);
+
     LOG_VERBOSE("completion parameters parsed", format_generation_settings(llama));
 }
 
@@ -1463,7 +1496,7 @@ int main(int argc, char **argv)
             }
 
             auto probs = llama.generated_token_probs;
-            if (llama.params.n_probs > 0 && llama.stopped_word) {
+            if (llama.params.sampling_params.n_probs > 0 && llama.stopped_word) {
                 const std::vector<llama_token> stop_word_toks = llama_tokenize(llama.ctx, llama.stopping_word, false);
                 probs = std::vector<completion_token_output>(llama.generated_token_probs.begin(), llama.generated_token_probs.end() - stop_word_toks.size());
             }
@@ -1515,7 +1548,7 @@ int main(int argc, char **argv)
 
                         std::vector<completion_token_output> probs_output = {};
 
-                        if (llama.params.n_probs > 0) {
+                        if (llama.params.sampling_params.n_probs > 0) {
                             const std::vector<llama_token> to_send_toks = llama_tokenize(llama.ctx, to_send, false);
                             size_t probs_pos = std::min(sent_token_probs_index, llama.generated_token_probs.size());
                             size_t probs_stop_pos = std::min(sent_token_probs_index + to_send_toks.size(), llama.generated_token_probs.size());
@@ -1636,7 +1669,7 @@ int main(int argc, char **argv)
 
                     std::vector<completion_token_output> probs_output = {};
 
-                    if (llama.params.n_probs > 0) {
+                    if (llama.params.sampling_params.n_probs > 0) {
                         const std::vector<llama_token> to_send_toks = llama_tokenize(llama.ctx, to_send, false);
                         size_t probs_pos = std::min(sent_token_probs_index, llama.generated_token_probs.size());
                         size_t probs_stop_pos = std::min(sent_token_probs_index + to_send_toks.size(), llama.generated_token_probs.size());

examples/speculative/speculative.cpp

@@ -125,6 +125,8 @@ int main(int argc, char ** argv) {
         grammar_tgt = llama_grammar_init(grammar_rules.data(), grammar_rules.size(), parsed_grammar.symbol_ids.at("root"));
     }
 
+    llama_sampling_context ctx_sampling = llama_sampling_context_init(params, grammar_tgt);
+
     const auto t_dec_start = ggml_time_us();
 
     while (true) {
@@ -134,7 +136,7 @@ int main(int argc, char ** argv) {
 
         while (true) {
             // sample from the target model
-            llama_token id = llama_sample_token(ctx_tgt, NULL, grammar_tgt, params, last_tokens, candidates, i_dft);
+            llama_token id = llama_sampling_sample(ctx_tgt, NULL, ctx_sampling, last_tokens, candidates, i_dft);
 
             // remember which tokens were sampled - used for repetition penalties during sampling
             last_tokens.erase(last_tokens.begin());
@@ -211,7 +213,13 @@ int main(int argc, char ** argv) {
             if (grammar_dft) {
                 llama_grammar_free(grammar_dft);
             }
-            grammar_dft = llama_grammar_copy(grammar_tgt);
+            // Note: Hardcoded to sequence id 0, if this ever supports parallel generation
+            //       that will need to change.
+            auto it = ctx_sampling.sequence_contexts.find(0);
+            GGML_ASSERT(it != ctx_sampling.sequence_contexts.end());
+            // This is necessary because each sequence id in sequence_contexts
+            // uses a copy of the original grammar.
+            grammar_dft = llama_grammar_copy(it->second.grammar);
 
             LOG("copied target grammar to draft grammar\n");
         }

ggml-alloc.c

@@ -1,4 +1,5 @@
 #include "ggml-alloc.h"
+#include "ggml-backend.h"
 #include "ggml.h"
 #include <assert.h>
 #include <stdarg.h>
@@ -6,25 +7,6 @@
 #include <stdlib.h>
 #include <string.h>
 
-#ifdef __has_include
-    #if __has_include(<unistd.h>)
-        #include <unistd.h>
-        #if defined(_POSIX_MAPPED_FILES)
-            #include <sys/types.h>
-            #include <sys/mman.h>
-        #endif
-    #endif
-#endif
-
-#if defined(_WIN32)
-    #define WIN32_LEAN_AND_MEAN
-    #ifndef NOMINMAX
-        #define NOMINMAX
-    #endif
-    #include <windows.h>
-    #include <memoryapi.h>
-#endif
-
 
 #define UNUSED(x) (void)(x)
 #define MAX(a, b) ((a) > (b) ? (a) : (b))
@@ -80,8 +62,9 @@ struct free_block {
 #define MAX_FREE_BLOCKS 256
 
 struct ggml_allocr {
+    struct ggml_backend_buffer * buffer;
+    bool buffer_owned;
     void * data;
-    size_t size;
     size_t alignment;
     int n_free_blocks;
     struct free_block free_blocks[MAX_FREE_BLOCKS];
@@ -119,16 +102,9 @@ static void remove_allocated_tensor(struct ggml_allocr * alloc, struct ggml_tens
 }
 #endif
 
-static size_t ggml_allocr_get_alloc_size(struct ggml_allocr * alloc, struct ggml_tensor * tensor) {
-    return ggml_nbytes(tensor);
-
-    UNUSED(alloc);
-}
-
 // check if a tensor is allocated by this buffer
 static bool ggml_allocr_is_own(struct ggml_allocr * alloc, const struct ggml_tensor * tensor) {
-    void * ptr = tensor->data;
-    return ptr >= alloc->data && (char *)ptr < (char *)alloc->data + alloc->max_size;
+    return tensor->buffer == alloc->buffer;
 }
 
 static bool ggml_is_view(struct ggml_tensor * t) {
@@ -136,11 +112,10 @@ static bool ggml_is_view(struct ggml_tensor * t) {
 }
 
 void ggml_allocr_alloc(struct ggml_allocr * alloc, struct ggml_tensor * tensor) {
-#ifdef GGML_ALLOCATOR_DEBUG
     GGML_ASSERT(!ggml_is_view(tensor)); // views generally get data pointer from one of their sources
     GGML_ASSERT(tensor->data == NULL); // avoid allocating tensor which already has memory allocated
-#endif
-    size_t size = ggml_allocr_get_alloc_size(alloc, tensor);
+
+    size_t size = ggml_backend_buffer_get_alloc_size(alloc->buffer, tensor);
     size = aligned_offset(NULL, size, alloc->alignment);
 
     AT_PRINTF("%s: allocating %s (%zu bytes) - ", __func__, tensor->name, size);
@@ -188,6 +163,8 @@ void ggml_allocr_alloc(struct ggml_allocr * alloc, struct ggml_tensor * tensor)
 
     tensor->data = addr;
     AT_PRINTF("%s: allocated data at %p\n", __func__, tensor->data);
+    tensor->buffer = alloc->buffer;
+    ggml_backend_buffer_init_tensor(alloc->buffer, tensor);
 
 #ifdef GGML_ALLOCATOR_DEBUG
     add_allocated_tensor(alloc, tensor);
@@ -208,19 +185,21 @@ void ggml_allocr_alloc(struct ggml_allocr * alloc, struct ggml_tensor * tensor)
 
 // this is a very naive implementation, but for our case the number of free blocks should be very small
 static void ggml_allocr_free_tensor(struct ggml_allocr * alloc, struct ggml_tensor * tensor) {
-    void * ptr = tensor->data;
-
     if (ggml_allocr_is_own(alloc, tensor) == false) {
         // the tensor was not allocated in this buffer
         // this can happen because the graph allocator will try to free weights and other tensors from different buffers
         // the easiest way to deal with this is just to ignore it
+        AT_PRINTF("ignoring %s (their buffer: %p, our buffer: %p)\n", tensor->name, (void *)tensor->buffer, (void *)alloc->buffer);
         return;
     }
 
-    size_t size = ggml_allocr_get_alloc_size(alloc, tensor);
+    void * ptr = tensor->data;
+
+    size_t size = ggml_backend_buffer_get_alloc_size(alloc->buffer, tensor);
     size = aligned_offset(NULL, size, alloc->alignment);
     AT_PRINTF("%s: freeing %s at %p (%zu bytes) - n_free_blocks = %d\n", __func__, tensor->name, ptr, size, alloc->n_free_blocks);
-    AT_PRINTF("%s: alloc->data = %p alloc->data+alloc->size = %p alloc->data+alloc->max_size = %p\n", __func__, alloc->data, (char*)alloc->data + alloc->size, (char*)alloc->data + alloc->max_size);
+
+    ggml_backend_buffer_free_tensor(alloc->buffer, tensor);
 
 #ifdef GGML_ALLOCATOR_DEBUG
     remove_allocated_tensor(alloc, tensor);
@@ -285,15 +264,18 @@ void ggml_allocr_reset(struct ggml_allocr * alloc) {
     alloc->n_free_blocks = 1;
     size_t align_offset = aligned_offset(alloc->data, 0, alloc->alignment);
     alloc->free_blocks[0].addr = (char *)alloc->data + align_offset;
-    alloc->free_blocks[0].size = alloc->size - align_offset;
+    alloc->free_blocks[0].size = ggml_backend_buffer_get_size(alloc->buffer) - align_offset;
 }
 
 struct ggml_allocr * ggml_allocr_new(void * data, size_t size, size_t alignment) {
-    struct ggml_allocr * alloc = (struct ggml_allocr *)malloc(sizeof(struct ggml_allocr) /* + n_free_blocks * sizeof(struct free_block) */);
+    struct ggml_backend_buffer * buffer = ggml_backend_cpu_buffer_from_ptr(NULL, data, size);
+
+    struct ggml_allocr * alloc = (struct ggml_allocr *)malloc(sizeof(struct ggml_allocr));
 
     *alloc = (struct ggml_allocr){
-        /*.data          = */ data,
-        /*.size          = */ size,
+        /*.buffer        = */ buffer,
+        /*.buffer_owned  = */ true,
+        /*.base          = */ ggml_backend_buffer_get_base(buffer),
         /*.alignment     = */ alignment,
         /*.n_free_blocks = */ 0,
         /*.free_blocks   = */ {{0}},
@@ -312,74 +294,26 @@ struct ggml_allocr * ggml_allocr_new(void * data, size_t size, size_t alignment)
     return alloc;
 }
 
-// OS specific functions to allocate and free uncommitted virtual memory
-static void * alloc_vmem(size_t size) {
-#if defined(_WIN32)
-    return VirtualAlloc(NULL, size, MEM_RESERVE, PAGE_NOACCESS);
-#elif defined(_POSIX_MAPPED_FILES)
-    void * ptr = mmap(NULL, size, PROT_NONE, MAP_PRIVATE | MAP_ANON, -1, 0);
-    if (ptr == MAP_FAILED) {
-        return NULL;
-    }
-    return ptr;
-#else
-    // use a fixed address for other platforms
-    uintptr_t base_addr = (uintptr_t)-size - 0x100;
-    return (void *)base_addr;
-#endif
-}
-
-static void free_vmem(void * base_addr, size_t size) {
-#if defined(_WIN32)
-    VirtualFree(base_addr, 0, MEM_RELEASE);
-    UNUSED(size);
-#elif defined(_POSIX_MAPPED_FILES)
-    munmap(base_addr, size);
-#else
-    // nothing to do
-    UNUSED(base_addr);
-    UNUSED(size);
-#endif
-}
-
-// allocate uncommitted virtual memory to measure the size of the graph
-static void alloc_measure_vmem(void ** base_addr, size_t * size) {
-    // 128GB for 64-bit, 1GB for 32-bit
-    *size = sizeof(void *) == 4 ? 1ULL<<30 : 1ULL<<37;
-    do {
-        *base_addr = alloc_vmem(*size);
-        if (*base_addr != NULL) {
-            AT_PRINTF("allocated %.2f GB of virtual memory for measure buffer at %p\n", *size / 1024.0 / 1024.0 / 1024.0, *base_addr);
-            return;
-        }
-        // try again with half the size
-        *size /= 2;
-    } while (*size > 0);
-
-    GGML_ASSERT(!"failed to allocate virtual memory for measure buffer");
-}
-
-static void free_measure_vmem(void * base_addr, size_t size) {
-    free_vmem(base_addr, size);
-}
-
 struct ggml_allocr * ggml_allocr_new_measure(size_t alignment) {
-    struct ggml_allocr * alloc = (struct ggml_allocr *)malloc(sizeof(struct ggml_allocr) /* + n_free_blocks * sizeof(struct free_block) */);
+    struct ggml_allocr * alloc = ggml_allocr_new((void *)0x1000, (size_t)-0x1001, alignment);
+    alloc->measure = true;
 
-    void * base_addr;
-    size_t size;
+    return alloc;
+}
 
-    alloc_measure_vmem(&base_addr, &size);
+struct ggml_allocr * ggml_allocr_new_from_buffer(struct ggml_backend_buffer * buffer) {
+    struct ggml_allocr * alloc = (struct ggml_allocr *)malloc(sizeof(struct ggml_allocr));
 
     *alloc = (struct ggml_allocr){
-        /*.data          = */ base_addr,
-        /*.size          = */ size,
-        /*.alignment     = */ alignment,
+        /*.buffer        = */ buffer,
+        /*.buffer_owned  = */ false,
+        /*.base          = */ ggml_backend_buffer_get_base(buffer),
+        /*.alignment     = */ ggml_backend_buffer_get_alignment(buffer),
         /*.n_free_blocks = */ 0,
         /*.free_blocks   = */ {{0}},
         /*.hash_table    = */ {{0}},
         /*.max_size      = */ 0,
-        /*.measure       = */ true,
+        /*.measure       = */ false,
         /*.parse_seq     = */ {0},
         /*.parse_seq_len = */ 0,
 #ifdef GGML_ALLOCATOR_DEBUG
@@ -393,8 +327,8 @@ struct ggml_allocr * ggml_allocr_new_measure(size_t alignment) {
 }
 
 void ggml_allocr_free(struct ggml_allocr * alloc) {
-    if (alloc->measure) {
-        free_measure_vmem(alloc->data, alloc->size);
+    if (alloc->buffer_owned) {
+        ggml_backend_buffer_free(alloc->buffer);
     }
     free(alloc);
 }
@@ -437,7 +371,6 @@ static bool ggml_op_can_inplace(enum ggml_op op) {
         case GGML_OP_ROPE:
         case GGML_OP_RMS_NORM:
         case GGML_OP_SOFT_MAX:
-        case GGML_OP_CONT:
             return true;
 
         default:
@@ -445,12 +378,23 @@ static bool ggml_op_can_inplace(enum ggml_op op) {
     }
 }
 
+static void init_view(struct ggml_allocr * alloc, struct ggml_tensor * view) {
+    assert(view->view_src != NULL && view->view_src->data != NULL);
+    view->backend = view->view_src->backend;
+    view->buffer  = view->view_src->buffer;
+    view->data    = (char *)view->view_src->data + view->view_offs;
+
+    // FIXME: the view should be initialized by the owning buffer, but currently this breaks the CUDA backend
+    // due to the ggml_tensor_extra_gpu ring buffer overwriting the KV cache extras
+    assert(ggml_allocr_is_measure(alloc) || !view->buffer || view->buffer->backend == alloc->buffer->backend);
+    ggml_backend_buffer_init_tensor(alloc->buffer, view);
+}
+
 static void allocate_node(struct ggml_allocr * alloc, struct ggml_tensor * node) {
     struct hash_node * ht = alloc->hash_table;
     if (node->data == NULL) {
         if (ggml_is_view(node)) {
-            assert(node->view_src->data != NULL);
-            node->data = (char *)node->view_src->data + node->view_offs;
+            init_view(alloc, node);
         } else {
             // see if we can reuse a parent's buffer (inplace)
             if (ggml_op_can_inplace(node->op)) {
@@ -478,13 +422,17 @@ static void allocate_node(struct ggml_allocr * alloc, struct ggml_tensor * node)
                                 // adding a view_src pointer to the tensor would solve this and simplify the code dealing with views
                                 // for now, we only reuse the parent's data if the offset is zero (view_src->data == parent->data)
                                 AT_PRINTF("reusing view parent %s (%s) for %s\n", parent->name, view_src->name, node->name);
-                                node->data = parent->data;
+                                node->view_src = view_src;
+                                view_src_hn->n_views += 1;
+                                init_view(alloc, node);
                                 return;
                             }
                         }
                         else {
                             AT_PRINTF("reusing parent %s for %s\n", parent->name, node->name);
-                            node->data = parent->data;
+                            node->view_src = parent;
+                            p_hn->n_views += 1;
+                            init_view(alloc, node);
                             return;
                         }
                     }
@@ -495,7 +443,7 @@ static void allocate_node(struct ggml_allocr * alloc, struct ggml_tensor * node)
     }
 }
 
-static size_t ggml_allocr_alloc_graph_tensors_n(
+size_t ggml_allocr_alloc_graph_n(
     struct ggml_allocr * alloc,
     struct ggml_cgraph ** graphs, int n_graphs,
     struct ggml_tensor *** inputs, struct ggml_tensor *** outputs) {
@@ -513,6 +461,10 @@ static size_t ggml_allocr_alloc_graph_tensors_n(
             if (ggml_is_view(node)) {
                 struct ggml_tensor * view_src = node->view_src;
                 hash_get(ht, view_src)->n_views += 1;
+                if (node->buffer == NULL && node->data != NULL) {
+                    // view of a pre-allocated tensor, didn't call init_view() yet
+                    init_view(alloc, node);
+                }
             }
 
             for (int j = 0; j < GGML_MAX_SRC; j++) {
@@ -521,6 +473,9 @@ static size_t ggml_allocr_alloc_graph_tensors_n(
                     break;
                 }
                 hash_get(ht, parent)->n_children += 1;
+                if (ggml_is_view(parent) && parent->buffer == NULL && parent->data != NULL) {
+                    init_view(alloc, parent);
+                }
             }
         }
     }
@@ -631,7 +586,7 @@ static size_t ggml_allocr_alloc_graph_tensors_n(
 }
 
 size_t ggml_allocr_alloc_graph(struct ggml_allocr * alloc, struct ggml_cgraph * graph) {
-    return ggml_allocr_alloc_graph_tensors_n(alloc, &graph, 1, NULL, NULL);
+    return ggml_allocr_alloc_graph_n(alloc, &graph, 1, NULL, NULL);
 }
 
 size_t ggml_allocr_max_size(struct ggml_allocr * alloc) {

ggml-alloc.h

@@ -6,9 +6,11 @@
 extern "C" {
 #endif
 
+struct ggml_backend_buffer;
 
 GGML_API struct ggml_allocr * ggml_allocr_new(void * data, size_t size, size_t alignment);
 GGML_API struct ggml_allocr * ggml_allocr_new_measure(size_t alignment);
+GGML_API struct ggml_allocr * ggml_allocr_new_from_buffer(struct ggml_backend_buffer * buffer);
 
 // tell the allocator to parse nodes following the order described in the list
 // you should call this if your graph are optimized to execute out-of-order
@@ -21,6 +23,10 @@ GGML_API void   ggml_allocr_alloc(struct ggml_allocr * alloc, struct ggml_tensor
 GGML_API size_t ggml_allocr_alloc_graph(struct ggml_allocr * alloc, struct ggml_cgraph * graph);
 GGML_API size_t ggml_allocr_max_size   (struct ggml_allocr * alloc);
 
+GGML_API size_t ggml_allocr_alloc_graph_n(
+                    struct ggml_allocr * alloc,
+                    struct ggml_cgraph ** graphs, int n_graphs,
+                    struct ggml_tensor *** inputs, struct ggml_tensor *** outputs);
 
 #ifdef  __cplusplus
 }

ggml-backend.c

@@ -0,0 +1,385 @@
+#include "ggml-backend.h"
+#include "ggml-alloc.h"
+
+#include <assert.h>
+#include <stdarg.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#define UNUSED GGML_UNUSED
+
+#define MAX(a, b) ((a) > (b) ? (a) : (b))
+
+// backend buffer
+
+ggml_backend_buffer_t ggml_backend_buffer_init(
+        struct ggml_backend                  * backend,
+        struct ggml_backend_buffer_i           iface,
+               ggml_backend_buffer_context_t   context,
+               size_t                          size) {
+    ggml_backend_buffer_t buffer = malloc(sizeof(struct ggml_backend_buffer));
+
+    GGML_ASSERT(iface.get_base != NULL);
+
+    (*buffer) = (struct ggml_backend_buffer) {
+        /* .interface = */ iface,
+        /* .backend   = */ backend,
+        /* .context   = */ context,
+        /* .size      = */ size,
+    };
+
+    return buffer;
+}
+
+void ggml_backend_buffer_free(ggml_backend_buffer_t buffer) {
+    if (buffer->iface.free_buffer != NULL) {
+        buffer->iface.free_buffer(buffer);
+    }
+    free(buffer);
+}
+
+size_t ggml_backend_buffer_get_alignment(ggml_backend_buffer_t buffer) {
+    return ggml_backend_get_alignment(buffer->backend);
+}
+
+void * ggml_backend_buffer_get_base(ggml_backend_buffer_t buffer) {
+    return buffer->iface.get_base(buffer);
+}
+
+size_t ggml_backend_buffer_get_size(ggml_backend_buffer_t buffer) {
+    return buffer->size;
+}
+
+size_t ggml_backend_buffer_get_alloc_size(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor) {
+    if (buffer->iface.get_alloc_size) {
+        return buffer->iface.get_alloc_size(buffer, tensor);
+    }
+    return ggml_nbytes(tensor);
+}
+
+void ggml_backend_buffer_init_tensor(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor) {
+    if (buffer->iface.init_tensor) {
+        buffer->iface.init_tensor(buffer, tensor);
+    }
+}
+
+void ggml_backend_buffer_free_tensor(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor) {
+    if (buffer->iface.free_tensor) {
+        buffer->iface.free_tensor(buffer, tensor);
+    }
+}
+
+// backend
+
+ggml_backend_t ggml_get_backend(const struct ggml_tensor * tensor) {
+    return tensor->buffer->backend;
+}
+
+const char * ggml_backend_name(ggml_backend_t backend) {
+    return backend->iface.get_name(backend);
+}
+
+void ggml_backend_free(ggml_backend_t backend) {
+    backend->iface.free(backend);
+}
+
+ggml_backend_buffer_t ggml_backend_alloc_buffer(ggml_backend_t backend, size_t size) {
+    return backend->iface.alloc_buffer(backend, size);
+}
+
+size_t ggml_backend_get_alignment(ggml_backend_t backend) {
+    return backend->iface.get_alignment(backend);
+}
+
+void ggml_backend_tensor_set_async(struct ggml_tensor * tensor, const void * data, size_t offset, size_t size) {
+    ggml_get_backend(tensor)->iface.set_tensor_async(ggml_get_backend(tensor), tensor, data, offset, size);
+}
+
+void ggml_backend_tensor_get_async(const struct ggml_tensor * tensor, void * data, size_t offset, size_t size) {
+    ggml_get_backend(tensor)->iface.get_tensor_async(ggml_get_backend(tensor), tensor, data, offset, size);
+}
+
+void ggml_backend_tensor_set(struct ggml_tensor * tensor, const void * data, size_t offset, size_t size) {
+    ggml_get_backend(tensor)->iface.set_tensor_async(ggml_get_backend(tensor), tensor, data, offset, size);
+    ggml_get_backend(tensor)->iface.synchronize(ggml_get_backend(tensor));
+}
+
+void ggml_backend_tensor_get(const struct ggml_tensor * tensor, void * data, size_t offset, size_t size) {
+    ggml_get_backend(tensor)->iface.get_tensor_async(ggml_get_backend(tensor), tensor, data, offset, size);
+    ggml_get_backend(tensor)->iface.synchronize(ggml_get_backend(tensor));
+}
+
+void ggml_backend_synchronize(ggml_backend_t backend) {
+    backend->iface.synchronize(backend);
+}
+
+ggml_backend_graph_plan_t ggml_backend_graph_plan_create(ggml_backend_t backend, struct ggml_cgraph * cgraph) {
+    return backend->iface.graph_plan_create(backend, cgraph);
+}
+
+void ggml_backend_graph_plan_free(ggml_backend_t backend, ggml_backend_graph_plan_t plan) {
+    backend->iface.graph_plan_free(backend, plan);
+}
+
+void ggml_backend_graph_plan_compute(ggml_backend_t backend, ggml_backend_graph_plan_t plan) {
+    backend->iface.graph_plan_compute(backend, plan);
+}
+
+void ggml_backend_graph_compute(ggml_backend_t backend, struct ggml_cgraph * cgraph) {
+    backend->iface.graph_compute(backend, cgraph);
+}
+
+bool ggml_backend_supports_op(ggml_backend_t backend, const struct ggml_tensor * op) {
+    return backend->iface.supports_op(backend, op);
+}
+
+// backend copy
+
+static bool ggml_are_same_layout(const struct ggml_tensor * a, const struct ggml_tensor * b) {
+    if (a->type != b->type) {
+        return false;
+    }
+    for (int i = 0; i < GGML_MAX_DIMS; i++) {
+        if (a->ne[i] != b->ne[i]) {
+            return false;
+        }
+        if (a->nb[i] != b->nb[i]) {
+            return false;
+        }
+    }
+    return true;
+}
+
+void ggml_backend_tensor_copy(struct ggml_tensor * src, struct ggml_tensor * dst) {
+    //printf("src: %s ne: [%d %d %d %d] nb: [%d %d %d %d]\n", src->name, (int)src->ne[0], (int)src->ne[1], (int)src->ne[2], (int)src->ne[3], (int)src->nb[0], (int)src->nb[1], (int)src->nb[2], (int)src->nb[3]);
+    //printf("dst: %s ne: [%d %d %d %d] nb: [%d %d %d %d]\n", dst->name, (int)dst->ne[0], (int)dst->ne[1], (int)dst->ne[2], (int)dst->ne[3], (int)dst->nb[0], (int)dst->nb[1], (int)dst->nb[2], (int)dst->nb[3]);
+    GGML_ASSERT(ggml_are_same_layout(src, dst) && "cannot copy tensors with different layouts");
+
+    // printf("cpy tensor %s from %s to %s (%lu bytes)\n", src->name, ggml_backend_name(src->backend), ggml_backend_name(dst->backend), ggml_nbytes(src));
+
+    if (src == dst) {
+        return;
+    }
+
+    // TODO: allow backends to support copy to/from same backend
+
+    if (ggml_get_backend(dst)->iface.cpy_tensor_from != NULL) {
+        ggml_get_backend(dst)->iface.cpy_tensor_from(ggml_get_backend(dst)->context, src, dst);
+    } else if (ggml_get_backend(src)->iface.cpy_tensor_to != NULL) {
+        ggml_get_backend(src)->iface.cpy_tensor_to(ggml_get_backend(src)->context, src, dst);
+    } else {
+        // shouldn't be hit when copying from/to CPU
+        #ifndef NDEBUG
+        fprintf(stderr, "ggml_backend_tensor_copy: neither cpy_tensor_from nor cpy_tensor_to are implemented for backends %s and %s, falling back to get/set\n", ggml_backend_name(src->buffer->backend), ggml_backend_name(dst->buffer->backend));
+        #endif
+        size_t nbytes = ggml_nbytes(src);
+        void * data = malloc(nbytes);
+        ggml_backend_tensor_get(src, data, 0, nbytes);
+        ggml_backend_tensor_set(dst, data, 0, nbytes);
+        free(data);
+    }
+}
+
+// backend CPU
+
+struct ggml_backend_cpu_context {
+    int n_threads;
+    void * work_data;
+    size_t work_size;
+};
+
+static const char * ggml_backend_cpu_name(ggml_backend_t backend) {
+    return "CPU";
+
+    UNUSED(backend);
+}
+
+static void ggml_backend_cpu_free(ggml_backend_t backend) {
+    struct ggml_backend_cpu_context * cpu_ctx = (struct ggml_backend_cpu_context *)backend->context;
+    free(cpu_ctx->work_data);
+    free(cpu_ctx);
+    free(backend);
+}
+
+static void * ggml_backend_cpu_buffer_get_base(ggml_backend_buffer_t buffer) {
+    return (void *)buffer->context;
+}
+
+static void ggml_backend_cpu_buffer_free_buffer(ggml_backend_buffer_t buffer) {
+    free(buffer->context);
+    UNUSED(buffer);
+}
+
+static struct ggml_backend_buffer_i cpu_backend_buffer_i = {
+    /* .free_buffer    = */ ggml_backend_cpu_buffer_free_buffer,
+    /* .get_base       = */ ggml_backend_cpu_buffer_get_base,
+    /* .get_alloc_size = */ NULL, // defaults to ggml_nbytes
+    /* .init_tensor    = */ NULL, // no initialization required
+    /* .free_tensor    = */ NULL, // no cleanup required
+};
+
+// for buffers from ptr, free is not called
+static struct ggml_backend_buffer_i cpu_backend_buffer_i_from_ptr = {
+    /* .free_buffer    = */ NULL, // ptr is not owned by the buffer, so it does not need to be freed
+    /* .get_base       = */ ggml_backend_cpu_buffer_get_base,
+    /* .get_alloc_size = */ NULL, // defaults to ggml_nbytes
+    /* .init_tensor    = */ NULL,
+    /* .free_tensor    = */ NULL,
+};
+
+static const size_t TENSOR_ALIGNMENT = 64; // should be enough for AVX 512
+
+static ggml_backend_buffer_t ggml_backend_cpu_alloc_buffer(ggml_backend_t backend, size_t size) {
+    size += TENSOR_ALIGNMENT;   // malloc may return an address that is not aligned
+    void * data = malloc(size); // TODO: maybe use GGML_ALIGNED_MALLOC?
+
+    return ggml_backend_buffer_init(backend, cpu_backend_buffer_i, data, size);
+}
+
+static size_t ggml_backend_cpu_get_alignment(ggml_backend_t backend) {
+    return TENSOR_ALIGNMENT;
+    UNUSED(backend);
+}
+
+static void ggml_backend_cpu_set_tensor_async(ggml_backend_t backend, struct ggml_tensor * tensor, const void * data, size_t offset, size_t size) {
+    GGML_ASSERT(offset + size <= ggml_nbytes(tensor) && "tensor write out of bounds");
+    GGML_ASSERT(tensor->data != NULL && "tensor not allocated");
+
+    memcpy((char *)tensor->data + offset, data, size);
+
+    UNUSED(backend);
+}
+
+static void ggml_backend_cpu_get_tensor_async(ggml_backend_t backend, const struct ggml_tensor * tensor, void * data, size_t offset, size_t size) {
+    GGML_ASSERT(offset + size <= ggml_nbytes(tensor) && "tensor read out of bounds");
+    GGML_ASSERT(tensor->data != NULL && "tensor not allocated");
+
+    memcpy(data, (const char *)tensor->data + offset, size);
+
+    UNUSED(backend);
+}
+
+static void ggml_backend_cpu_synchronize(ggml_backend_t backend) {
+    UNUSED(backend);
+}
+
+static void ggml_backend_cpu_cpy_tensor_from(ggml_backend_t backend, struct ggml_tensor * src, struct ggml_tensor * dst) {
+    ggml_backend_tensor_get(src, dst->data, 0, ggml_nbytes(src));
+
+    UNUSED(backend);
+}
+
+static void ggml_backend_cpu_cpy_tensor_to(ggml_backend_t backend, struct ggml_tensor * src, struct ggml_tensor * dst) {
+    // for a backend such as CUDA that can queue async calls, it is ok to do this asynchronously, but it may not be the case for other backends
+    ggml_backend_tensor_set_async(dst, src->data, 0, ggml_nbytes(src));
+
+    UNUSED(backend);
+}
+
+struct ggml_backend_plan_cpu {
+    struct ggml_cplan cplan;
+    struct ggml_cgraph cgraph;
+};
+
+static ggml_backend_graph_plan_t ggml_backend_cpu_graph_plan_create(ggml_backend_t backend, struct ggml_cgraph * cgraph) {
+    struct ggml_backend_cpu_context * cpu_ctx = (struct ggml_backend_cpu_context *)backend->context;
+
+    struct ggml_backend_plan_cpu * cpu_plan = malloc(sizeof(struct ggml_backend_plan_cpu));
+
+    cpu_plan->cplan = ggml_graph_plan(cgraph, cpu_ctx->n_threads);
+    cpu_plan->cgraph = *cgraph;
+
+    if (cpu_plan->cplan.work_size > 0) {
+        cpu_plan->cplan.work_data = malloc(cpu_plan->cplan.work_size);
+    }
+
+    return cpu_plan;
+}
+
+static void ggml_backend_cpu_graph_plan_free(ggml_backend_t backend, ggml_backend_graph_plan_t plan) {
+    struct ggml_backend_plan_cpu * cpu_plan = (struct ggml_backend_plan_cpu *)plan;
+
+    free(cpu_plan->cplan.work_data);
+    free(cpu_plan);
+
+    UNUSED(backend);
+}
+
+static void ggml_backend_cpu_graph_plan_compute(ggml_backend_t backend, ggml_backend_graph_plan_t plan) {
+    struct ggml_backend_plan_cpu * cpu_plan = (struct ggml_backend_plan_cpu *)plan;
+
+    ggml_graph_compute(&cpu_plan->cgraph, &cpu_plan->cplan);
+
+    UNUSED(backend);
+}
+
+static void ggml_backend_cpu_graph_compute(ggml_backend_t backend, struct ggml_cgraph * cgraph) {
+    struct ggml_backend_cpu_context * cpu_ctx = (struct ggml_backend_cpu_context *)backend->context;
+
+    struct ggml_cplan cplan = ggml_graph_plan(cgraph, cpu_ctx->n_threads);
+
+    if (cpu_ctx->work_size < cplan.work_size) {
+        // TODO: may be faster to free and use malloc to avoid the copy
+        cpu_ctx->work_data = realloc(cpu_ctx->work_data, cplan.work_size);
+        cpu_ctx->work_size = cplan.work_size;
+    }
+
+    cplan.work_data = cpu_ctx->work_data;
+
+    ggml_graph_compute(cgraph, &cplan);
+}
+
+static bool ggml_backend_cpu_supports_op(ggml_backend_t backend, const struct ggml_tensor * op) {
+    return true;
+    UNUSED(backend);
+    UNUSED(op);
+}
+
+static struct ggml_backend_i cpu_backend_i = {
+    /* .get_name            = */ ggml_backend_cpu_name,
+    /* .free                = */ ggml_backend_cpu_free,
+    /* .alloc_buffer        = */ ggml_backend_cpu_alloc_buffer,
+    /* .get_alignment       = */ ggml_backend_cpu_get_alignment,
+    /* .set_tensor_async    = */ ggml_backend_cpu_set_tensor_async,
+    /* .get_tensor_async    = */ ggml_backend_cpu_get_tensor_async,
+    /* .synchronize         = */ ggml_backend_cpu_synchronize,
+    /* .cpy_tensor_from     = */ ggml_backend_cpu_cpy_tensor_from,
+    /* .cpy_tensor_to       = */ ggml_backend_cpu_cpy_tensor_to,
+    /* .graph_plan_create   = */ ggml_backend_cpu_graph_plan_create,
+    /* .graph_plan_free     = */ ggml_backend_cpu_graph_plan_free,
+    /* .graph_plan_compute  = */ ggml_backend_cpu_graph_plan_compute,
+    /* .graph_compute       = */ ggml_backend_cpu_graph_compute,
+    /* .supports_op         = */ ggml_backend_cpu_supports_op,
+};
+
+ggml_backend_t ggml_backend_cpu_init(void) {
+    struct ggml_backend_cpu_context * ctx = malloc(sizeof(struct ggml_backend_cpu_context));
+
+    ctx->n_threads = GGML_DEFAULT_N_THREADS;
+    ctx->work_data = NULL;
+    ctx->work_size = 0;
+
+    ggml_backend_t cpu_backend = malloc(sizeof(struct ggml_backend));
+
+    *cpu_backend = (struct ggml_backend) {
+        /* .interface = */ cpu_backend_i,
+        /* .context   = */ ctx
+    };
+    return cpu_backend;
+}
+
+bool ggml_backend_is_cpu(ggml_backend_t backend) {
+    return backend->iface.get_name == ggml_backend_cpu_name;
+}
+
+void ggml_backend_cpu_set_n_threads(ggml_backend_t backend_cpu, int n_threads) {
+    GGML_ASSERT(ggml_backend_is_cpu(backend_cpu));
+
+    struct ggml_backend_cpu_context * ctx = (struct ggml_backend_cpu_context *)backend_cpu->context;
+    ctx->n_threads = n_threads;
+}
+
+ggml_backend_buffer_t ggml_backend_cpu_buffer_from_ptr(ggml_backend_t backend_cpu, void * ptr, size_t size) {
+    return ggml_backend_buffer_init(backend_cpu, cpu_backend_buffer_i_from_ptr, ptr, size);
+}

ggml-backend.h

@@ -0,0 +1,143 @@
+#pragma once
+
+#include "ggml.h"
+
+#ifdef  __cplusplus
+extern "C" {
+#endif
+    struct ggml_backend;
+    struct ggml_backend_buffer;
+
+    // type-erased backend-specific types / wrappers
+    typedef void * ggml_backend_context_t;
+    typedef void * ggml_backend_graph_plan_t;
+    typedef void * ggml_backend_buffer_context_t;
+
+    // avoid accessing internals of these types
+    typedef struct ggml_backend        * ggml_backend_t;
+    typedef struct ggml_backend_buffer * ggml_backend_buffer_t;
+
+    //
+    // backend buffer
+    //
+
+    struct ggml_backend_buffer_i {
+        void   (*free_buffer)   (ggml_backend_buffer_t buffer);
+        void * (*get_base)      (ggml_backend_buffer_t buffer); // get base pointer
+        size_t (*get_alloc_size)(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor); // pre-allocation callback
+        void   (*init_tensor)   (ggml_backend_buffer_t buffer, struct ggml_tensor * tensor); // post-allocation callback
+        void   (*free_tensor)   (ggml_backend_buffer_t buffer, struct ggml_tensor * tensor); // pre-free callback
+    };
+
+    // TODO: hide behind API
+    struct ggml_backend_buffer {
+        struct ggml_backend_buffer_i iface;
+
+        ggml_backend_t                backend;
+        ggml_backend_buffer_context_t context;
+
+        size_t size;
+    };
+
+    // backend buffer functions
+    GGML_API ggml_backend_buffer_t ggml_backend_buffer_init(
+            struct ggml_backend                  * backend,
+            struct ggml_backend_buffer_i           iface,
+                   ggml_backend_buffer_context_t   context,
+                   size_t                          size);
+
+    GGML_API void   ggml_backend_buffer_free          (ggml_backend_buffer_t buffer);
+    GGML_API size_t ggml_backend_buffer_get_alignment (ggml_backend_buffer_t buffer);
+    GGML_API void * ggml_backend_buffer_get_base      (ggml_backend_buffer_t buffer);
+    GGML_API size_t ggml_backend_buffer_get_size      (ggml_backend_buffer_t buffer);
+    GGML_API size_t ggml_backend_buffer_get_alloc_size(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor);
+    GGML_API void   ggml_backend_buffer_init_tensor   (ggml_backend_buffer_t buffer, struct ggml_tensor * tensor);
+    GGML_API void   ggml_backend_buffer_free_tensor   (ggml_backend_buffer_t buffer, struct ggml_tensor * tensor);
+
+    //
+    // backend
+    //
+
+    struct ggml_backend_i {
+        const char * (*get_name)(ggml_backend_t backend);
+
+        void (*free)(ggml_backend_t backend);
+
+        // buffer allocation
+        ggml_backend_buffer_t (*alloc_buffer)(ggml_backend_t backend, size_t size);
+
+        // get buffer alignment
+        size_t (*get_alignment)(ggml_backend_t backend);
+
+        // tensor data access
+        // these functions can be asynchronous, helper functions are provided for synchronous access that automatically call synchronize
+        void (*set_tensor_async)(ggml_backend_t backend,       struct ggml_tensor * tensor, const void * data, size_t offset, size_t size);
+        void (*get_tensor_async)(ggml_backend_t backend, const struct ggml_tensor * tensor,       void * data, size_t offset, size_t size);
+        void (*synchronize)     (ggml_backend_t backend);
+
+        // (optional) copy tensor between different backends, allow for single-copy tranfers
+        void (*cpy_tensor_from)(ggml_backend_t backend, struct ggml_tensor * src, struct ggml_tensor * dst);
+        void (*cpy_tensor_to)  (ggml_backend_t backend, struct ggml_tensor * src, struct ggml_tensor * dst);
+
+        // compute graph with a plan
+        ggml_backend_graph_plan_t (*graph_plan_create) (ggml_backend_t backend, struct ggml_cgraph * cgraph);
+        void                      (*graph_plan_free)   (ggml_backend_t backend, ggml_backend_graph_plan_t plan);
+        void                      (*graph_plan_compute)(ggml_backend_t backend, ggml_backend_graph_plan_t plan);
+
+        // compute graph without a plan
+        void (*graph_compute)(ggml_backend_t backend, struct ggml_cgraph * cgraph);
+
+        // check if the backend supports an operation
+        bool (*supports_op)(ggml_backend_t backend, const struct ggml_tensor * op);
+    };
+
+    // TODO: hide behind API
+    struct ggml_backend {
+        struct ggml_backend_i iface;
+
+        ggml_backend_context_t context;
+    };
+
+    // backend helper functions
+    GGML_API ggml_backend_t ggml_get_backend(const struct ggml_tensor * tensor);
+
+    GGML_API const char * ggml_backend_name(ggml_backend_t backend);
+    GGML_API void         ggml_backend_free(ggml_backend_t backend);
+
+    GGML_API ggml_backend_buffer_t ggml_backend_alloc_buffer(ggml_backend_t backend, size_t size);
+
+    GGML_API size_t ggml_backend_get_alignment(ggml_backend_t backend);
+
+    GGML_API void ggml_backend_tensor_set_async(      struct ggml_tensor * tensor, const void * data, size_t offset, size_t size);
+    GGML_API void ggml_backend_tensor_get_async(const struct ggml_tensor * tensor,       void * data, size_t offset, size_t size);
+
+    GGML_API void ggml_backend_tensor_set(      struct ggml_tensor * tensor, const void * data, size_t offset, size_t size);
+    GGML_API void ggml_backend_tensor_get(const struct ggml_tensor * tensor,       void * data, size_t offset, size_t size);
+
+    GGML_API void ggml_backend_synchronize(ggml_backend_t backend);
+
+    GGML_API ggml_backend_graph_plan_t ggml_backend_graph_plan_create (ggml_backend_t backend, struct ggml_cgraph * cgraph);
+
+    GGML_API void ggml_backend_graph_plan_free   (ggml_backend_t backend, ggml_backend_graph_plan_t plan);
+    GGML_API void ggml_backend_graph_plan_compute(ggml_backend_t backend, ggml_backend_graph_plan_t plan);
+    GGML_API void ggml_backend_graph_compute     (ggml_backend_t backend, struct ggml_cgraph * cgraph);
+    GGML_API bool ggml_backend_supports_op       (ggml_backend_t backend, const struct ggml_tensor * op);
+
+    // tensor copy between different backends
+    GGML_API void ggml_backend_tensor_copy(struct ggml_tensor * src, struct ggml_tensor * dst);
+
+    //
+    // CPU backend
+    //
+
+    GGML_API ggml_backend_t ggml_backend_cpu_init(void);
+
+    GGML_API bool ggml_backend_is_cpu(ggml_backend_t backend);
+
+    GGML_API void ggml_backend_cpu_set_n_threads(ggml_backend_t backend_cpu, int n_threads);
+
+    GGML_API ggml_backend_buffer_t ggml_backend_cpu_buffer_from_ptr(ggml_backend_t backend_cpu, void * ptr, size_t size);
+
+#ifdef  __cplusplus
+}
+#endif

ggml-cuda.cu

@@ -62,6 +62,7 @@
 #define cudaMemcpyHostToDevice hipMemcpyHostToDevice
 #define cudaMemcpyKind hipMemcpyKind
 #define cudaMemset hipMemset
+#define cudaMemsetAsync hipMemsetAsync
 #define cudaOccupancyMaxPotentialBlockSize hipOccupancyMaxPotentialBlockSize
 #define cudaSetDevice hipSetDevice
 #define cudaStreamCreateWithFlags hipStreamCreateWithFlags
@@ -414,11 +415,13 @@ static_assert(sizeof(block_q6_K) == sizeof(ggml_fp16_t) + 13*QK_K/16, "wrong q6_
 #define CUDA_SILU_BLOCK_SIZE 256
 #define CUDA_CPY_BLOCK_SIZE 32
 #define CUDA_SCALE_BLOCK_SIZE 256
+#define CUDA_CLAMP_BLOCK_SIZE 256
 #define CUDA_ROPE_BLOCK_SIZE 256
 #define CUDA_ALIBI_BLOCK_SIZE 32
 #define CUDA_DIAG_MASK_INF_BLOCK_SIZE 32
 #define CUDA_QUANTIZE_BLOCK_SIZE 256
 #define CUDA_DEQUANTIZE_BLOCK_SIZE 256
+#define CUDA_GET_ROWS_BLOCK_SIZE 256
 
 // dmmv = dequantize_mul_mat_vec
 #ifndef GGML_CUDA_DMMV_X
@@ -1574,6 +1577,34 @@ static __global__ void quantize_q8_1(const float * __restrict__ x, void * __rest
     reinterpret_cast<half&>(y[ib].ds.y) = sum;
 }
 
+template<int qk, int qr, dequantize_kernel_t dequantize_kernel, typename dst_t>
+static __global__ void k_get_rows(const void * x, const int32_t * y, dst_t * dst, const int ncols) {
+    const int col = (blockIdx.x*blockDim.x + threadIdx.x)*2;
+    const int row = blockDim.y*blockIdx.y + threadIdx.y;
+
+    if (col >= ncols) {
+        return;
+    }
+
+    const int r = y[row];
+
+    // copy x[r*ncols + col] to dst[row*ncols + col]
+    const int xi = r*ncols + col;
+    const int di = row*ncols + col;
+
+    const int ib = xi/qk; // block index
+    const int iqs = (xi%qk)/qr; // quant index
+    const int iybs = di - di%qk; // y block start index
+    const int y_offset = qr == 1 ? 1 : qk/2;
+
+    // dequantize
+    dfloat2 v;
+    dequantize_kernel(x, ib, iqs, v);
+
+    dst[iybs + iqs + 0]        = v.x;
+    dst[iybs + iqs + y_offset] = v.y;
+}
+
 template <int qk, int qr, dequantize_kernel_t dequantize_kernel, typename dst_t>
 static __global__ void dequantize_block(const void * __restrict__ vx, dst_t * __restrict__ y, const int k) {
     const int i = blockDim.x*blockIdx.x + 2*threadIdx.x;
@@ -4586,6 +4617,24 @@ static __global__ void scale_f32(const float * x, float * dst, const float scale
     dst[i] = scale * x[i];
 }
 
+static __global__ void clamp_f32(const float * x, float * dst, const float min, const float max, const int k) {
+    const int i = blockDim.x*blockIdx.x + threadIdx.x;
+
+    if (i >= k) {
+        return;
+    }
+
+    dst[i] = x[i] < min ? min : (x[i] > max ? max : x[i]);
+}
+
+template<int qk, int qr, dequantize_kernel_t dq>
+static void get_rows_cuda(const void * x, const int32_t * y, float * dst, const int nrows, const int ncols, cudaStream_t stream) {
+    const dim3 block_dims(CUDA_GET_ROWS_BLOCK_SIZE, 1, 1);
+    const int block_num_x = (ncols + 2*CUDA_GET_ROWS_BLOCK_SIZE - 1) / (2*CUDA_GET_ROWS_BLOCK_SIZE);
+    const dim3 block_nums(block_num_x, nrows, 1);
+    k_get_rows<qk, qr, dq><<<block_nums, block_dims, 0, stream>>>(x, y, dst, ncols);
+}
+
 static void add_f32_cuda(const float * x, const float * y, float * dst, const int kx, const int ky, cudaStream_t stream) {
     const int num_blocks = (kx + CUDA_ADD_BLOCK_SIZE - 1) / CUDA_ADD_BLOCK_SIZE;
     add_f32<<<num_blocks, CUDA_ADD_BLOCK_SIZE, 0, stream>>>(x, y, dst, kx, ky);
@@ -5467,6 +5516,11 @@ static void scale_f32_cuda(const float * x, float * dst, const float scale, cons
     scale_f32<<<num_blocks, CUDA_SCALE_BLOCK_SIZE, 0, stream>>>(x, dst, scale, k);
 }
 
+static void clamp_f32_cuda(const float * x, float * dst, const float min, const float max, const int k, cudaStream_t stream) {
+    const int num_blocks = (k + CUDA_CLAMP_BLOCK_SIZE - 1) / CUDA_CLAMP_BLOCK_SIZE;
+    clamp_f32<<<num_blocks, CUDA_CLAMP_BLOCK_SIZE, 0, stream>>>(x, dst, min, max, k);
+}
+
 template<typename T>
 static void rope_cuda(
     const T * x, T * dst, int ncols, int nrows, const int32_t * pos, float freq_scale, int p_delta_rows,
@@ -5740,7 +5794,7 @@ static cudaError_t ggml_cuda_cpy_tensor_2d(
     } else if (src->backend == GGML_BACKEND_GPU || src->backend == GGML_BACKEND_GPU_SPLIT) {
         GGML_ASSERT(src->backend != GGML_BACKEND_GPU_SPLIT || (i1_low == 0 && i1_high == src->ne[1]));
         kind = cudaMemcpyDeviceToDevice;
-        struct ggml_tensor_extra_gpu * extra = (ggml_tensor_extra_gpu *) src->extra;
+        ggml_tensor_extra_gpu * extra = (ggml_tensor_extra_gpu *) src->extra;
         int id;
         CUDA_CHECK(cudaGetDevice(&id));
         src_ptr = (char *) extra->data_device[id];
@@ -5776,6 +5830,107 @@ static cudaError_t ggml_cuda_cpy_tensor_2d(
     }
 }
 
+static void ggml_cuda_op_repeat(
+    const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst,
+    const float * src0_d, const float * src1_d, float * dst_d, const cudaStream_t & stream) {
+    // guaranteed to be an integer due to the check in ggml_can_repeat
+    const int64_t ne0 = dst->ne[0];
+    const int64_t ne1 = dst->ne[1];
+    const int64_t ne2 = dst->ne[2];
+    const int64_t ne3 = dst->ne[3];
+
+    const int64_t ne00 = src0->ne[0];
+    const int64_t ne01 = src0->ne[1];
+    const int64_t ne02 = src0->ne[2];
+    const int64_t ne03 = src0->ne[3];
+
+    const size_t nb0 = dst->nb[0];
+    const size_t nb1 = dst->nb[1];
+    const size_t nb2 = dst->nb[2];
+    const size_t nb3 = dst->nb[3];
+
+    const size_t nb00 = src0->nb[0];
+    const size_t nb01 = src0->nb[1];
+    const size_t nb02 = src0->nb[2];
+    const size_t nb03 = src0->nb[3];
+
+    const int nr0 = (int)(ne0/ne00);
+    const int nr1 = (int)(ne1/ne01);
+    const int nr2 = (int)(ne2/ne02);
+    const int nr3 = (int)(ne3/ne03);
+
+    // TODO: support for transposed / permuted tensors
+    GGML_ASSERT(nb0  == sizeof(float));
+    GGML_ASSERT(nb00 == sizeof(float));
+
+    // TODO: very inefficient, implement in a kernel, or fewer cudaMemcpyAsync calls for contiguous tensors
+    for                         (int i3 = 0; i3 < nr3;  i3++) {
+        for                     (int k3 = 0; k3 < ne03; k3++) {
+            for                 (int i2 = 0; i2 < nr2;  i2++) {
+                for             (int k2 = 0; k2 < ne02; k2++) {
+                    for         (int i1 = 0; i1 < nr1;  i1++) {
+                        for     (int k1 = 0; k1 < ne01; k1++) {
+                            for (int i0 = 0; i0 < nr0;  i0++) {
+                                CUDA_CHECK(cudaMemcpyAsync(
+                                              (char *)  dst_d + (i3*ne03 + k3)*nb3  + (i2*ne02 + k2)*nb2  + (i1*ne01 + k1)*nb1  + (i0*ne00)*nb0,
+                                        (const char *) src0_d + (          k3)*nb03 + (          k2)*nb02 + (          k1)*nb01,
+                                        ne00*nb0, cudaMemcpyDeviceToDevice, stream));
+                            }
+                        }
+                    }
+                }
+            }
+        }
+    }
+
+    (void) src1;
+    (void) src1_d;
+}
+
+static void ggml_cuda_op_get_rows(
+    const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst,
+    const float * src0_d, const float * src1_d, float * dst_d, const cudaStream_t & stream) {
+
+    GGML_ASSERT(src1->type == GGML_TYPE_I32);
+    GGML_ASSERT(dst->type == GGML_TYPE_F32);
+    GGML_ASSERT(ggml_is_contiguous(src0));
+    GGML_ASSERT(ggml_is_contiguous(src1));
+    GGML_ASSERT(ggml_is_contiguous(dst));
+
+    const int ncols = src0->ne[0];
+    const int nrows = ggml_nelements(src1);
+
+    const int32_t * src1_i32 = (const int32_t *) src1_d;
+
+    switch (src0->type) {
+        case GGML_TYPE_F16:
+            get_rows_cuda<1, 1, convert_f16>(src0_d, src1_i32, dst_d, nrows, ncols, stream);
+            break;
+        case GGML_TYPE_F32:
+            get_rows_cuda<1, 1, convert_f32>(src0_d, src1_i32, dst_d, nrows, ncols, stream);
+            break;
+        case GGML_TYPE_Q4_0:
+            get_rows_cuda<QK4_0, QR4_0, dequantize_q4_0>(src0_d, src1_i32, dst_d, nrows, ncols, stream);
+            break;
+        case GGML_TYPE_Q4_1:
+            get_rows_cuda<QK4_1, QR4_1, dequantize_q4_1>(src0_d, src1_i32, dst_d, nrows, ncols, stream);
+            break;
+        case GGML_TYPE_Q5_0:
+            get_rows_cuda<QK5_0, QR5_0, dequantize_q5_0>(src0_d, src1_i32, dst_d, nrows, ncols, stream);
+            break;
+        case GGML_TYPE_Q5_1:
+            get_rows_cuda<QK5_1, QR5_1, dequantize_q5_1>(src0_d, src1_i32, dst_d, nrows, ncols, stream);
+            break;
+        case GGML_TYPE_Q8_0:
+            get_rows_cuda<QK8_0, QR8_0, dequantize_q8_0>(src0_d, src1_i32, dst_d, nrows, ncols, stream);
+            break;
+        default:
+            // TODO: k-quants
+            GGML_ASSERT(false);
+            break;
+    }
+}
+
 inline void ggml_cuda_op_add(
     const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst,
     const float * src0_dd, const float * src1_dd, float * dst_dd, const cudaStream_t & main_stream) {
@@ -6330,12 +6485,12 @@ inline void ggml_cuda_op_alibi(
     const int64_t ne02 = src0->ne[2];
     const int64_t nrows = ggml_nrows(src0);
 
-    const int n_past = ((int32_t *) dst->op_params)[0];
+    //const int n_past = ((int32_t *) dst->op_params)[0];
     const int n_head = ((int32_t *) dst->op_params)[1];
     float max_bias;
     memcpy(&max_bias, (int32_t *) dst->op_params + 2, sizeof(float));
 
-    GGML_ASSERT(ne01 + n_past == ne00);
+    //GGML_ASSERT(ne01 + n_past == ne00);
     GGML_ASSERT(n_head == ne02);
 
     const int n_heads_log2_floor = 1 << (int) floor(log2(n_head));
@@ -6394,7 +6549,14 @@ inline void ggml_cuda_op_scale(
     GGML_ASSERT(src1->type == GGML_TYPE_F32);
     GGML_ASSERT( dst->type == GGML_TYPE_F32);
 
-    const float scale = ((float *) src1->data)[0];
+    float scale;
+    // HACK: support for ggml backend interface
+    if (src1->backend == GGML_BACKEND_CPU) {
+        scale = ((float *) src1->data)[0];
+    } else {
+        // TODO: pass pointer to kernel instead of copying to host
+        CUDA_CHECK(cudaMemcpy(&scale, src1->data, sizeof(float), cudaMemcpyDeviceToHost));
+    }
 
     scale_f32_cuda(src0_dd, dst_dd, scale, ggml_nelements(src0), main_stream);
     CUDA_CHECK(cudaGetLastError());
@@ -6404,6 +6566,24 @@ inline void ggml_cuda_op_scale(
     (void) src1_dd;
 }
 
+inline void ggml_cuda_op_clamp(
+    const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst,
+    const float * src0_dd, const float * src1_dd, float * dst_dd, const cudaStream_t & main_stream) {
+
+    GGML_ASSERT(src0->type == GGML_TYPE_F32);
+    GGML_ASSERT( dst->type == GGML_TYPE_F32);
+
+    const float min = ((float *) dst->op_params)[0];
+    const float max = ((float *) dst->op_params)[1];
+
+    clamp_f32_cuda(src0_dd, dst_dd, min, max, ggml_nelements(src0), main_stream);
+    CUDA_CHECK(cudaGetLastError());
+
+    (void) src1;
+    (void) dst;
+    (void) src1_dd;
+}
+
 static void ggml_cuda_op_flatten(const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst, const ggml_cuda_op_flatten_t op) {
     const int64_t nrows0 = ggml_nrows(src0);
 
@@ -6413,9 +6593,9 @@ static void ggml_cuda_op_flatten(const ggml_tensor * src0, const ggml_tensor * s
     GGML_ASSERT(!use_src1 || src1->backend != GGML_BACKEND_GPU_SPLIT);
     GGML_ASSERT(              dst->backend != GGML_BACKEND_GPU_SPLIT);
 
-    struct ggml_tensor_extra_gpu * src0_extra =            (ggml_tensor_extra_gpu *) src0->extra;
-    struct ggml_tensor_extra_gpu * src1_extra = use_src1 ? (ggml_tensor_extra_gpu *) src1->extra : nullptr;
-    struct ggml_tensor_extra_gpu * dst_extra  =            (ggml_tensor_extra_gpu *)  dst->extra;
+    ggml_tensor_extra_gpu * src0_extra =            (ggml_tensor_extra_gpu *) src0->extra;
+    ggml_tensor_extra_gpu * src1_extra = use_src1 ? (ggml_tensor_extra_gpu *) src1->extra : nullptr;
+    ggml_tensor_extra_gpu * dst_extra  =            (ggml_tensor_extra_gpu *)  dst->extra;
 
     const bool src0_on_device =             src0->backend == GGML_BACKEND_GPU || src0->backend == GGML_BACKEND_GPU_SPLIT;
     const bool src1_on_device = use_src1 && src1->backend == GGML_BACKEND_GPU;
@@ -6556,9 +6736,9 @@ static void ggml_cuda_op_mul_mat(
     const size_t q8_1_ts = sizeof(block_q8_1);
     const size_t q8_1_bs = QK8_1;
 
-    struct ggml_tensor_extra_gpu * src0_extra = (ggml_tensor_extra_gpu *) src0->extra;
-    struct ggml_tensor_extra_gpu * src1_extra = (ggml_tensor_extra_gpu *) src1->extra;
-    struct ggml_tensor_extra_gpu *  dst_extra = (ggml_tensor_extra_gpu *)  dst->extra;
+    ggml_tensor_extra_gpu * src0_extra = (ggml_tensor_extra_gpu *) src0->extra;
+    ggml_tensor_extra_gpu * src1_extra = (ggml_tensor_extra_gpu *) src1->extra;
+    ggml_tensor_extra_gpu *  dst_extra = (ggml_tensor_extra_gpu *)  dst->extra;
 
     const bool src0_on_device = src0->backend == GGML_BACKEND_GPU || src0->backend == GGML_BACKEND_GPU_SPLIT;
     const bool src0_is_contiguous = ggml_is_contiguous(src0);
@@ -6636,7 +6816,7 @@ static void ggml_cuda_op_mul_mat(
         if (convert_src1_to_q8_1) {
             src1_ddq[id] = (char *) ggml_cuda_pool_malloc(nrows1*src1_padded_col_size*q8_1_ts/q8_1_bs, &src1_asq[id]);
 
-            if (split && src1_on_device && src1_is_contiguous) {
+            if (src1_on_device && src1_is_contiguous) {
                 quantize_row_q8_1_cuda(src1_ddf[id], src1_ddq[id], ne10, nrows1, src1_padded_col_size, stream);
                 CUDA_CHECK(cudaGetLastError());
             }
@@ -6718,7 +6898,7 @@ static void ggml_cuda_op_mul_mat(
                     GGML_ASSERT(false);
                 }
 
-                if (convert_src1_to_q8_1 && src1->backend == GGML_BACKEND_CPU) {
+                if (convert_src1_to_q8_1 && (src1->backend == GGML_BACKEND_CPU || !src1_is_contiguous)) {
                     quantize_row_q8_1_cuda(src1_ddf_i, src1_ddq_i, ne10, src1_ncols, src1_padded_col_size, stream);
                     CUDA_CHECK(cudaGetLastError());
                 }
@@ -6809,6 +6989,14 @@ static void ggml_cuda_op_mul_mat(
     }
 }
 
+static void ggml_cuda_repeat(const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) {
+    ggml_cuda_op_flatten(src0, src1, dst, ggml_cuda_op_repeat);
+}
+
+static void ggml_cuda_get_rows(const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) {
+    ggml_cuda_op_flatten(src0, src1, dst, ggml_cuda_op_get_rows);
+}
+
 static void ggml_cuda_add(const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) {
     ggml_cuda_op_flatten(src0, src1, dst, ggml_cuda_op_add);
 }
@@ -6863,13 +7051,13 @@ static void ggml_cuda_mul_mat_vec_p021(const ggml_tensor * src0, const ggml_tens
     CUDA_CHECK(ggml_cuda_set_device(g_main_device));
     cudaStream_t main_stream = g_cudaStreams[g_main_device][0];
 
-    struct ggml_tensor_extra_gpu * src0_extra = (ggml_tensor_extra_gpu *) src0->extra;
+    ggml_tensor_extra_gpu * src0_extra = (ggml_tensor_extra_gpu *) src0->extra;
     void * src0_ddq = src0_extra->data_device[g_main_device];
 
-    struct ggml_tensor_extra_gpu * src1_extra = (ggml_tensor_extra_gpu *) src1->extra;
+    ggml_tensor_extra_gpu * src1_extra = (ggml_tensor_extra_gpu *) src1->extra;
     float * src1_ddf = (float *) src1_extra->data_device[g_main_device];
 
-    struct ggml_tensor_extra_gpu * dst_extra = (ggml_tensor_extra_gpu *) dst->extra;
+    ggml_tensor_extra_gpu * dst_extra = (ggml_tensor_extra_gpu *) dst->extra;
     float * dst_ddf = (float *) dst_extra->data_device[g_main_device];
 
     ggml_mul_mat_p021_f16_f32_cuda(src0_ddq, src1_ddf, dst_ddf, ne00, ne01, ne02, ne12, main_stream);
@@ -6894,13 +7082,13 @@ static void ggml_cuda_mul_mat_vec_nc(const ggml_tensor * src0, const ggml_tensor
     CUDA_CHECK(ggml_cuda_set_device(g_main_device));
     cudaStream_t main_stream = g_cudaStreams[g_main_device][0];
 
-    struct ggml_tensor_extra_gpu * src0_extra = (ggml_tensor_extra_gpu *) src0->extra;
+    ggml_tensor_extra_gpu * src0_extra = (ggml_tensor_extra_gpu *) src0->extra;
     void * src0_ddq = src0_extra->data_device[g_main_device];
 
-    struct ggml_tensor_extra_gpu * src1_extra = (ggml_tensor_extra_gpu *) src1->extra;
+    ggml_tensor_extra_gpu * src1_extra = (ggml_tensor_extra_gpu *) src1->extra;
     float * src1_ddf = (float *) src1_extra->data_device[g_main_device];
 
-    struct ggml_tensor_extra_gpu * dst_extra = (ggml_tensor_extra_gpu *) dst->extra;
+    ggml_tensor_extra_gpu * dst_extra = (ggml_tensor_extra_gpu *) dst->extra;
     float * dst_ddf = (float *) dst_extra->data_device[g_main_device];
 
     const int64_t row_stride_x = nb01 / sizeof(half);
@@ -6921,7 +7109,7 @@ static void ggml_cuda_mul_mat(const ggml_tensor * src0, const ggml_tensor * src1
         }
     }
 
-    if (all_on_device && ggml_is_permuted(src0) && ggml_is_permuted(src1) && src1->ne[1] == 1) {
+    if (all_on_device && src0->type == GGML_TYPE_F16 && ggml_is_permuted(src0) && ggml_is_permuted(src1) && src1->ne[1] == 1) {
         ggml_cuda_mul_mat_vec_p021(src0, src1, dst);
     } else if (all_on_device && !ggml_is_contiguous(src0) && ggml_is_contiguous(src1) && src1->ne[1] == 1) {
         ggml_cuda_mul_mat_vec_nc(src0, src1, dst);
@@ -6957,6 +7145,10 @@ static void ggml_cuda_scale(const ggml_tensor * src0, const ggml_tensor * src1,
     ggml_cuda_op_flatten(src0, src1, dst, ggml_cuda_op_scale);
 }
 
+static void ggml_cuda_clamp(const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) {
+    ggml_cuda_op_flatten(src0, src1, dst, ggml_cuda_op_clamp);
+}
+
 static void ggml_cuda_cpy(const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) {
     const int64_t ne = ggml_nelements(src0);
     GGML_ASSERT(ne == ggml_nelements(src1));
@@ -6986,8 +7178,8 @@ static void ggml_cuda_cpy(const ggml_tensor * src0, const ggml_tensor * src1, gg
     CUDA_CHECK(ggml_cuda_set_device(g_main_device));
     cudaStream_t main_stream = g_cudaStreams[g_main_device][0];
 
-    const struct ggml_tensor_extra_gpu * src0_extra = (ggml_tensor_extra_gpu *) src0->extra;
-    const struct ggml_tensor_extra_gpu * src1_extra = (ggml_tensor_extra_gpu *) src1->extra;
+    const ggml_tensor_extra_gpu * src0_extra = (ggml_tensor_extra_gpu *) src0->extra;
+    const ggml_tensor_extra_gpu * src1_extra = (ggml_tensor_extra_gpu *) src1->extra;
 
     char * src0_ddc = (char *) src0_extra->data_device[g_main_device];
     char * src1_ddc = (char *) src1_extra->data_device[g_main_device];
@@ -7042,8 +7234,8 @@ void ggml_cuda_transform_tensor(void * data, struct ggml_tensor * tensor) {
 
     const size_t nb1 = tensor->nb[1];
 
-    ggml_backend backend = tensor->backend;
-    struct ggml_tensor_extra_gpu * extra = new struct ggml_tensor_extra_gpu;
+    ggml_backend_type backend = tensor->backend;
+    ggml_tensor_extra_gpu * extra = new struct ggml_tensor_extra_gpu;
     memset(extra, 0, sizeof(*extra));
 
     for (int64_t id = 0; id < g_device_count; ++id) {
@@ -7097,7 +7289,6 @@ void ggml_cuda_transform_tensor(void * data, struct ggml_tensor * tensor) {
             CUDA_CHECK(cudaMemset(buf + original_size, 0, size - original_size));
         }
 
-
         CUDA_CHECK(cudaMemcpy(buf, buf_host, original_size, cudaMemcpyHostToDevice));
 
         extra->data_device[id] = buf;
@@ -7136,17 +7327,17 @@ void ggml_cuda_free_data(struct ggml_tensor * tensor) {
     delete extra;
 }
 
-static struct ggml_tensor_extra_gpu * g_temp_tensor_extras = nullptr;
+static ggml_tensor_extra_gpu * g_temp_tensor_extras = nullptr;
 static size_t g_temp_tensor_extra_index = 0;
 
-static struct ggml_tensor_extra_gpu * ggml_cuda_alloc_temp_tensor_extra() {
+static ggml_tensor_extra_gpu * ggml_cuda_alloc_temp_tensor_extra() {
     if (g_temp_tensor_extras == nullptr) {
         g_temp_tensor_extras = new ggml_tensor_extra_gpu[GGML_MAX_NODES];
     }
 
     size_t alloc_index = g_temp_tensor_extra_index;
     g_temp_tensor_extra_index = (g_temp_tensor_extra_index + 1) % GGML_MAX_NODES;
-    struct ggml_tensor_extra_gpu * extra = &g_temp_tensor_extras[alloc_index];
+    ggml_tensor_extra_gpu * extra = &g_temp_tensor_extras[alloc_index];
     memset(extra, 0, sizeof(*extra));
 
     return extra;
@@ -7174,7 +7365,7 @@ static void ggml_cuda_assign_buffers_impl(struct ggml_tensor * tensor, bool scra
         return;
     }
 
-    struct ggml_tensor_extra_gpu * extra;
+    ggml_tensor_extra_gpu * extra;
 
     const bool inplace = (tensor->src[0] != nullptr && tensor->src[0]->data == tensor->data) ||
         tensor->op == GGML_OP_VIEW ||
@@ -7183,7 +7374,7 @@ static void ggml_cuda_assign_buffers_impl(struct ggml_tensor * tensor, bool scra
 
     CUDA_CHECK(ggml_cuda_set_device(g_main_device));
     if (inplace && (tensor->src[0]->backend == GGML_BACKEND_GPU || tensor->src[0]->backend == GGML_BACKEND_GPU_SPLIT)) {
-        struct ggml_tensor_extra_gpu * src0_extra = (ggml_tensor_extra_gpu * ) tensor->src[0]->extra;
+        ggml_tensor_extra_gpu * src0_extra = (ggml_tensor_extra_gpu * ) tensor->src[0]->extra;
         char * src0_ddc = (char *) src0_extra->data_device[g_main_device];
         size_t offset = 0;
         if (tensor->op == GGML_OP_VIEW) {
@@ -7192,7 +7383,7 @@ static void ggml_cuda_assign_buffers_impl(struct ggml_tensor * tensor, bool scra
         extra = ggml_cuda_alloc_temp_tensor_extra();
         extra->data_device[g_main_device] = src0_ddc + offset;
     } else if (tensor->op == GGML_OP_CPY) {
-        struct ggml_tensor_extra_gpu * src1_extra = (ggml_tensor_extra_gpu * ) tensor->src[1]->extra;
+        ggml_tensor_extra_gpu * src1_extra = (ggml_tensor_extra_gpu * ) tensor->src[1]->extra;
         void * src1_ddv = src1_extra->data_device[g_main_device];
         extra = ggml_cuda_alloc_temp_tensor_extra();
         extra->data_device[g_main_device] = src1_ddv;
@@ -7234,13 +7425,13 @@ void ggml_cuda_assign_scratch_offset(struct ggml_tensor * tensor, size_t offset)
         CUDA_CHECK(cudaMalloc(&g_scratch_buffer, g_scratch_size));
     }
 
-    struct ggml_tensor_extra_gpu * extra = ggml_cuda_alloc_temp_tensor_extra();
+    ggml_tensor_extra_gpu * extra = ggml_cuda_alloc_temp_tensor_extra();
 
     const bool inplace = (tensor->src[0] != nullptr && tensor->src[0]->data == tensor->data) ||
         tensor->op == GGML_OP_VIEW;
 
     if (inplace && (tensor->src[0]->backend == GGML_BACKEND_GPU || tensor->src[0]->backend == GGML_BACKEND_GPU_SPLIT)) {
-        struct ggml_tensor_extra_gpu * src0_extra = (ggml_tensor_extra_gpu * ) tensor->src[0]->extra;
+        ggml_tensor_extra_gpu * src0_extra = (ggml_tensor_extra_gpu * ) tensor->src[0]->extra;
         char * src0_ddc = (char *) src0_extra->data_device[g_main_device];
         size_t view_offset = 0;
         if (tensor->op == GGML_OP_VIEW) {
@@ -7258,7 +7449,7 @@ void ggml_cuda_copy_to_device(struct ggml_tensor * tensor) {
     GGML_ASSERT(tensor->backend == GGML_BACKEND_GPU);
     GGML_ASSERT(ggml_is_contiguous(tensor));
 
-    struct ggml_tensor_extra_gpu * extra = (ggml_tensor_extra_gpu *) tensor->extra;
+    ggml_tensor_extra_gpu * extra = (ggml_tensor_extra_gpu *) tensor->extra;
     CUDA_CHECK(ggml_cuda_set_device(g_main_device));
     CUDA_CHECK(cudaMemcpy(extra->data_device[g_main_device], tensor->data, ggml_nbytes(tensor), cudaMemcpyHostToDevice));
 }
@@ -7321,52 +7512,41 @@ bool ggml_cuda_compute_forward(struct ggml_compute_params * params, struct ggml_
         || (tensor->src[0] != nullptr && (tensor->src[0]->backend == GGML_BACKEND_GPU || tensor->src[0]->backend == GGML_BACKEND_GPU_SPLIT))
         || (tensor->src[1] != nullptr && tensor->src[1]->backend == GGML_BACKEND_GPU);
 
-    switch (tensor->op) {
-        case GGML_OP_DUP:
-            if (!any_on_device) {
+    if (!any_on_device && tensor->op != GGML_OP_MUL_MAT) {
         return false;
     }
+
+    switch (tensor->op) {
+        case GGML_OP_REPEAT:
+            func = ggml_cuda_repeat;
+            break;
+        case GGML_OP_GET_ROWS:
+            func = ggml_cuda_get_rows;
+            break;
+        case GGML_OP_DUP:
             func = ggml_cuda_dup;
             break;
         case GGML_OP_ADD:
-            if (!any_on_device) {
-                return false;
-            }
             func = ggml_cuda_add;
             break;
         case GGML_OP_MUL:
-            if (!any_on_device) {
-                return false;
-            }
             func = ggml_cuda_mul;
             break;
         case GGML_OP_UNARY:
             switch (ggml_get_unary_op(tensor)) {
                 case GGML_UNARY_OP_GELU:
-                    if (!any_on_device) {
-                        return false;
-                    }
                     func = ggml_cuda_gelu;
                     break;
                 case GGML_UNARY_OP_SILU:
-                    if (!any_on_device) {
-                        return false;
-                    }
                     func = ggml_cuda_silu;
                     break;
                 default:
                     return false;
             } break;
         case GGML_OP_NORM:
-            if (!any_on_device) {
-                return false;
-            }
             func = ggml_cuda_norm;
             break;
         case GGML_OP_RMS_NORM:
-            if (!any_on_device) {
-                return false;
-            }
             func = ggml_cuda_rms_norm;
             break;
         case GGML_OP_MUL_MAT:
@@ -7376,54 +7556,36 @@ bool ggml_cuda_compute_forward(struct ggml_compute_params * params, struct ggml_
             func = ggml_cuda_mul_mat;
             break;
         case GGML_OP_SCALE:
-            if (!any_on_device) {
-                return false;
-            }
             func = ggml_cuda_scale;
             break;
-        case GGML_OP_CPY:
+        case GGML_OP_CLAMP:
             if (!any_on_device) {
                 return false;
             }
+            func = ggml_cuda_clamp;
+            break;
+        case GGML_OP_CPY:
             func = ggml_cuda_cpy;
             break;
         case GGML_OP_CONT:
-            if (!any_on_device) {
-                return false;
-            }
             func = ggml_cuda_dup;
             break;
         case GGML_OP_RESHAPE:
         case GGML_OP_VIEW:
         case GGML_OP_PERMUTE:
         case GGML_OP_TRANSPOSE:
-            if (!any_on_device) {
-                return false;
-            }
             func = ggml_cuda_nop;
             break;
         case GGML_OP_DIAG_MASK_INF:
-            if (!any_on_device) {
-                return false;
-            }
             func = ggml_cuda_diag_mask_inf;
             break;
         case GGML_OP_SOFT_MAX:
-            if (!any_on_device) {
-                return false;
-            }
             func = ggml_cuda_soft_max;
             break;
         case GGML_OP_ROPE:
-            if (!any_on_device) {
-                return false;
-            }
             func = ggml_cuda_rope;
             break;
         case GGML_OP_ALIBI:
-            if (!any_on_device) {
-                return false;
-            }
             func = ggml_cuda_alibi;
             break;
         default:
@@ -7451,3 +7613,263 @@ void ggml_cuda_get_device_description(int device, char * description, size_t des
     CUDA_CHECK(cudaGetDeviceProperties(&prop, device));
     snprintf(description, description_size, "%s", prop.name);
 }
+
+////////////////////////////////////////////////////////////////////////////////
+
+// backend interface
+
+#define UNUSED GGML_UNUSED
+
+struct ggml_backend_context_cuda {
+};
+
+static const char * ggml_backend_cuda_name(ggml_backend_t backend) {
+    return GGML_CUDA_NAME;
+
+    UNUSED(backend);
+}
+
+static void ggml_backend_cuda_free(ggml_backend_t backend) {
+    ggml_backend_context_cuda * cuda_ctx = (ggml_backend_context_cuda *)backend->context;
+    delete cuda_ctx;
+    delete backend;
+}
+
+struct ggml_backend_buffer_context_cuda {
+    void * device;
+
+    ggml_tensor_extra_gpu * temp_tensor_extras = nullptr;
+    size_t temp_tensor_extra_index = 0;
+
+    ~ggml_backend_buffer_context_cuda() {
+        delete[] temp_tensor_extras;
+    }
+
+    ggml_tensor_extra_gpu * ggml_cuda_alloc_temp_tensor_extra() {
+        if (temp_tensor_extras == nullptr) {
+            temp_tensor_extras = new ggml_tensor_extra_gpu[GGML_MAX_NODES];
+        }
+
+        size_t alloc_index = temp_tensor_extra_index;
+        temp_tensor_extra_index = (temp_tensor_extra_index + 1) % GGML_MAX_NODES;
+        ggml_tensor_extra_gpu * extra = &temp_tensor_extras[alloc_index];
+        memset(extra, 0, sizeof(*extra));
+
+        return extra;
+    }
+};
+
+static void ggml_backend_cuda_buffer_free_buffer(ggml_backend_buffer_t buffer) {
+    ggml_backend_buffer_context_cuda * ctx = (ggml_backend_buffer_context_cuda *)buffer->context;
+    CUDA_CHECK(cudaFree(ctx->device));
+    delete ctx;
+}
+
+static void * ggml_backend_cuda_buffer_get_base(ggml_backend_buffer_t buffer) {
+    ggml_backend_buffer_context_cuda * ctx = (ggml_backend_buffer_context_cuda *)buffer->context;
+    return ctx->device;
+}
+
+static size_t ggml_backend_cuda_buffer_get_alloc_size(ggml_backend_buffer_t buffer, ggml_tensor * tensor) {
+    int64_t row_low = 0;
+    int64_t row_high = ggml_nrows(tensor);
+    int64_t nrows_split = row_high - row_low;
+
+    size_t size = ggml_nbytes_split(tensor, nrows_split);
+
+    int64_t ne0 = tensor->ne[0];
+
+    if (ggml_is_quantized(tensor->type)) {
+        if (ne0 % MATRIX_ROW_PADDING != 0) {
+            size += (MATRIX_ROW_PADDING - ne0 % MATRIX_ROW_PADDING)
+                * ggml_type_size(tensor->type)/ggml_blck_size(tensor->type);
+        }
+    }
+
+    return size;
+
+    UNUSED(buffer);
+}
+
+static void ggml_backend_cuda_buffer_init_tensor(ggml_backend_buffer_t buffer, ggml_tensor * tensor) {
+    ggml_backend_buffer_context_cuda * ctx = (ggml_backend_buffer_context_cuda *)buffer->context;
+
+    if (tensor->view_src != NULL && tensor->view_offs == 0) {
+        assert(tensor->view_src->buffer->backend == buffer->backend);
+        tensor->backend = tensor->view_src->backend;
+        tensor->extra = tensor->view_src->extra;
+        return;
+    }
+
+    ggml_tensor_extra_gpu * extra = ctx->ggml_cuda_alloc_temp_tensor_extra();
+
+    extra->data_device[g_main_device] = tensor->data;
+
+    tensor->backend = GGML_BACKEND_GPU;
+    tensor->extra = extra;
+
+    if (ggml_is_quantized(tensor->type)) {
+        // initialize padding to 0 to avoid possible NaN values
+        int64_t row_low = 0;
+        int64_t row_high = ggml_nrows(tensor);
+        int64_t nrows_split = row_high - row_low;
+
+        size_t original_size = ggml_nbytes_split(tensor, nrows_split);
+        size_t padded_size = ggml_backend_cuda_buffer_get_alloc_size(tensor->buffer, tensor);
+
+        if (padded_size > original_size && tensor->view_src == nullptr) {
+            CUDA_CHECK(cudaMemsetAsync((char *)tensor->data + original_size, 0, padded_size - original_size, g_cudaStreams[g_main_device][0]));
+        }
+    }
+
+    UNUSED(buffer);
+}
+
+static struct ggml_backend_buffer_i cuda_backend_buffer_interface = {
+    /* .free_buffer    = */ ggml_backend_cuda_buffer_free_buffer,
+    /* .get_base       = */ ggml_backend_cuda_buffer_get_base,
+    /* .get_alloc_size = */ ggml_backend_cuda_buffer_get_alloc_size,
+    /* .init_tensor    = */ ggml_backend_cuda_buffer_init_tensor,
+    /* .free_tensor    = */ NULL,
+};
+
+static ggml_backend_buffer_t ggml_backend_cuda_alloc_buffer(ggml_backend_t backend, size_t size) {
+    ggml_cuda_set_device(g_main_device);
+
+    ggml_backend_buffer_context_cuda * ctx = new ggml_backend_buffer_context_cuda;
+    CUDA_CHECK(cudaMalloc(&ctx->device, size));
+    return ggml_backend_buffer_init(backend, cuda_backend_buffer_interface, ctx, size);
+}
+
+static size_t ggml_backend_cuda_get_alignment(ggml_backend_t backend) {
+    return 128;
+    UNUSED(backend);
+}
+
+static void ggml_backend_cuda_set_tensor_async(ggml_backend_t backend, ggml_tensor * tensor, const void * data, size_t offset, size_t size) {
+    GGML_ASSERT(offset + size <= ggml_nbytes(tensor) && "tensor write out of bounds");
+    GGML_ASSERT(tensor->data != NULL && "tensor not allocated");
+    GGML_ASSERT(tensor->backend == GGML_BACKEND_GPU);
+
+    CUDA_CHECK(cudaMemcpyAsync((char *)tensor->data + offset, data, size, cudaMemcpyHostToDevice, g_cudaStreams[g_main_device][0]));
+
+    UNUSED(backend);
+}
+
+static void ggml_backend_cuda_get_tensor_async(ggml_backend_t backend, const ggml_tensor * tensor, void * data, size_t offset, size_t size) {
+    GGML_ASSERT(offset + size <= ggml_nbytes(tensor) && "tensor read out of bounds");
+    GGML_ASSERT(tensor->data != NULL && "tensor not allocated");
+    GGML_ASSERT(tensor->backend == GGML_BACKEND_GPU);
+
+    CUDA_CHECK(cudaMemcpyAsync(data, (const char *)tensor->data + offset, size, cudaMemcpyDeviceToHost, g_cudaStreams[g_main_device][0]));
+
+    UNUSED(backend);
+}
+
+static void ggml_backend_cuda_synchronize(ggml_backend_t backend) {
+    CUDA_CHECK(cudaStreamSynchronize(g_cudaStreams[g_main_device][0]));
+
+    UNUSED(backend);
+}
+
+static ggml_backend_graph_plan_t ggml_backend_cuda_graph_plan_create(ggml_backend_t backend, ggml_cgraph * cgraph) {
+    GGML_ASSERT(!"not implemented");
+
+    return nullptr;
+
+    UNUSED(backend);
+    UNUSED(cgraph);
+}
+
+static void ggml_backend_cuda_graph_plan_free(ggml_backend_t backend, ggml_backend_graph_plan_t plan) {
+    GGML_ASSERT(!"not implemented");
+
+    UNUSED(backend);
+    UNUSED(plan);
+}
+
+static void ggml_backend_cuda_graph_plan_compute(ggml_backend_t backend, ggml_backend_graph_plan_t plan) {
+    GGML_ASSERT(!"not implemented");
+
+    UNUSED(backend);
+    UNUSED(plan);
+}
+
+static void ggml_backend_cuda_graph_compute(ggml_backend_t backend, ggml_cgraph * cgraph) {
+    ggml_cuda_set_device(g_main_device);
+
+    ggml_compute_params params = {};
+    params.type = GGML_TASK_COMPUTE;
+    params.ith = 0;
+    for (int i = 0; i < cgraph->n_nodes; i++) {
+        ggml_tensor * node = cgraph->nodes[i];
+
+        assert(node->backend == GGML_BACKEND_GPU);
+        for (int j = 0; j < GGML_MAX_SRC; j++) {
+            if (node->src[j] != nullptr) {
+                assert(node->src[j]->backend == GGML_BACKEND_GPU);
+            }
+        }
+
+        bool ok = ggml_cuda_compute_forward(&params, node);
+        if (!ok) {
+            fprintf(stderr, "%s: error: op not supported %s (%s)\n", __func__, node->name, ggml_op_name(node->op));
+        }
+        GGML_ASSERT(ok);
+
+#if 0
+        if (node->type == GGML_TYPE_F32) {
+            cudaDeviceSynchronize();
+            std::vector<float> tmp(ggml_nelements(node), 0.0f);
+            cudaMemcpy(tmp.data(), node->data, ggml_nelements(node)*sizeof(float), cudaMemcpyDeviceToHost);
+            printf("\n%s (%s) (%s %s) (%s %s): ", node->name, ggml_op_name(node->op),
+                ggml_type_name(node->src[0]->type),
+                node->src[1] ? ggml_type_name(node->src[1]->type) : "none",
+                node->src[0]->name,
+                node->src[1] ? node->src[1]->name : "none");
+            double sum = 0.0;
+            double sq_sum = 0.0;
+            for (int i = 0; i < ggml_nelements(node); i++) {
+                printf("%f ", tmp[i]);
+                sum += tmp[i];
+                sq_sum += tmp[i]*tmp[i];
+            }
+            printf("\n");
+            printf("sum: %f, ", sum);
+            printf("sq_sum: %f\n", sq_sum);
+        }
+#endif
+    }
+
+    UNUSED(backend);
+}
+
+static ggml_backend_i cuda_backend_i = {
+    /* .get_name            = */ ggml_backend_cuda_name,
+    /* .free                = */ ggml_backend_cuda_free,
+    /* .alloc_buffer        = */ ggml_backend_cuda_alloc_buffer,
+    /* .get_alignment       = */ ggml_backend_cuda_get_alignment,
+    /* .set_tensor_async    = */ ggml_backend_cuda_set_tensor_async,
+    /* .get_tensor_async    = */ ggml_backend_cuda_get_tensor_async,
+    /* .synchronize         = */ ggml_backend_cuda_synchronize,
+    /* .cpy_tensor_from     = */ nullptr,
+    /* .cpy_tensor_to       = */ nullptr,
+    /* .graph_plan_create   = */ ggml_backend_cuda_graph_plan_create,
+    /* .graph_plan_free     = */ ggml_backend_cuda_graph_plan_free,
+    /* .graph_plan_compute  = */ ggml_backend_cuda_graph_plan_compute,
+    /* .graph_compute       = */ ggml_backend_cuda_graph_compute,
+    /* .supports_op         = */ nullptr,
+};
+
+ggml_backend_t ggml_backend_cuda_init() {
+    ggml_init_cublas(); // TODO: remove from ggml.c
+
+    ggml_backend_context_cuda * ctx = new ggml_backend_context_cuda;
+
+    ggml_backend_t cuda_backend = new ggml_backend {
+        /* .interface = */ cuda_backend_i,
+        /* .context   = */ ctx
+    };
+
+    return cuda_backend;
+}

ggml-cuda.h

@@ -1,6 +1,7 @@
 #pragma once
 
 #include "ggml.h"
+#include "ggml-backend.h"
 
 #ifdef GGML_USE_HIPBLAS
 #define GGML_CUDA_NAME "ROCm"
@@ -42,6 +43,9 @@ GGML_API bool   ggml_cuda_compute_forward(struct ggml_compute_params * params, s
 GGML_API int    ggml_cuda_get_device_count(void);
 GGML_API void   ggml_cuda_get_device_description(int device, char * description, size_t description_size);
 
+// backend API
+GGML_API ggml_backend_t ggml_backend_cuda_init(void); // TODO: take a list of devices to use
+
 #ifdef  __cplusplus
 }
 #endif

ggml-metal.h

@@ -20,6 +20,7 @@
 #pragma once
 
 #include "ggml.h"
+#include "ggml-backend.h"
 
 #include <stddef.h>
 #include <stdbool.h>
@@ -35,10 +36,15 @@ struct ggml_cgraph;
 extern "C" {
 #endif
 
-void ggml_metal_log_set_callback(ggml_log_callback log_callback, void * user_data);
+//
+// internal API
+// temporary exposed to user-code
+//
 
 struct ggml_metal_context;
 
+void ggml_metal_log_set_callback(ggml_log_callback log_callback, void * user_data);
+
 // number of command buffers to use
 struct ggml_metal_context * ggml_metal_init(int n_cb);
 void ggml_metal_free(struct ggml_metal_context * ctx);
@@ -83,6 +89,17 @@ int * ggml_metal_get_concur_list(struct ggml_metal_context * ctx);
 // creates gf->n_threads command buffers in parallel
 void ggml_metal_graph_compute(struct ggml_metal_context * ctx, struct ggml_cgraph * gf);
 
+//
+// backend API
+// user-code should use only these functions
+//
+
+GGML_API ggml_backend_t ggml_backend_metal_init(void);
+
+GGML_API bool ggml_backend_is_metal(ggml_backend_t backend);
+
+GGML_API void ggml_backend_metal_set_n_cb(ggml_backend_t backend, int n_cb);
+
 #ifdef __cplusplus
 }
 #endif

ggml-metal.m

@@ -81,18 +81,18 @@ struct ggml_metal_context {
     GGML_METAL_DECL_KERNEL(get_rows_q6_K);
     GGML_METAL_DECL_KERNEL(rms_norm);
     GGML_METAL_DECL_KERNEL(norm);
-    GGML_METAL_DECL_KERNEL(mul_mat_f32_f32);
-    GGML_METAL_DECL_KERNEL(mul_mat_f16_f32);
-    GGML_METAL_DECL_KERNEL(mul_mat_f16_f32_1row);
-    GGML_METAL_DECL_KERNEL(mul_mat_f16_f32_l4);
-    GGML_METAL_DECL_KERNEL(mul_mat_q4_0_f32);
-    GGML_METAL_DECL_KERNEL(mul_mat_q4_1_f32);
-    GGML_METAL_DECL_KERNEL(mul_mat_q8_0_f32);
-    GGML_METAL_DECL_KERNEL(mul_mat_q2_K_f32);
-    GGML_METAL_DECL_KERNEL(mul_mat_q3_K_f32);
-    GGML_METAL_DECL_KERNEL(mul_mat_q4_K_f32);
-    GGML_METAL_DECL_KERNEL(mul_mat_q5_K_f32);
-    GGML_METAL_DECL_KERNEL(mul_mat_q6_K_f32);
+    GGML_METAL_DECL_KERNEL(mul_mv_f32_f32);
+    GGML_METAL_DECL_KERNEL(mul_mv_f16_f32);
+    GGML_METAL_DECL_KERNEL(mul_mv_f16_f32_1row);
+    GGML_METAL_DECL_KERNEL(mul_mv_f16_f32_l4);
+    GGML_METAL_DECL_KERNEL(mul_mv_q4_0_f32);
+    GGML_METAL_DECL_KERNEL(mul_mv_q4_1_f32);
+    GGML_METAL_DECL_KERNEL(mul_mv_q8_0_f32);
+    GGML_METAL_DECL_KERNEL(mul_mv_q2_K_f32);
+    GGML_METAL_DECL_KERNEL(mul_mv_q3_K_f32);
+    GGML_METAL_DECL_KERNEL(mul_mv_q4_K_f32);
+    GGML_METAL_DECL_KERNEL(mul_mv_q5_K_f32);
+    GGML_METAL_DECL_KERNEL(mul_mv_q6_K_f32);
     GGML_METAL_DECL_KERNEL(mul_mm_f32_f32);
     GGML_METAL_DECL_KERNEL(mul_mm_f16_f32);
     GGML_METAL_DECL_KERNEL(mul_mm_q4_0_f32);
@@ -262,18 +262,19 @@ struct ggml_metal_context * ggml_metal_init(int n_cb) {
         GGML_METAL_ADD_KERNEL(get_rows_q6_K);
         GGML_METAL_ADD_KERNEL(rms_norm);
         GGML_METAL_ADD_KERNEL(norm);
-        GGML_METAL_ADD_KERNEL(mul_mat_f32_f32);
-        GGML_METAL_ADD_KERNEL(mul_mat_f16_f32);
-        GGML_METAL_ADD_KERNEL(mul_mat_f16_f32_1row);
-        GGML_METAL_ADD_KERNEL(mul_mat_f16_f32_l4);
-        GGML_METAL_ADD_KERNEL(mul_mat_q4_0_f32);
-        GGML_METAL_ADD_KERNEL(mul_mat_q4_1_f32);
-        GGML_METAL_ADD_KERNEL(mul_mat_q8_0_f32);
-        GGML_METAL_ADD_KERNEL(mul_mat_q2_K_f32);
-        GGML_METAL_ADD_KERNEL(mul_mat_q3_K_f32);
-        GGML_METAL_ADD_KERNEL(mul_mat_q4_K_f32);
-        GGML_METAL_ADD_KERNEL(mul_mat_q5_K_f32);
-        GGML_METAL_ADD_KERNEL(mul_mat_q6_K_f32);
+        GGML_METAL_ADD_KERNEL(mul_mv_f32_f32);
+        GGML_METAL_ADD_KERNEL(mul_mv_f16_f32);
+        GGML_METAL_ADD_KERNEL(mul_mv_f16_f32_1row);
+        GGML_METAL_ADD_KERNEL(mul_mv_f16_f32_l4);
+        GGML_METAL_ADD_KERNEL(mul_mv_q4_0_f32);
+        GGML_METAL_ADD_KERNEL(mul_mv_q4_1_f32);
+        GGML_METAL_ADD_KERNEL(mul_mv_q8_0_f32);
+        GGML_METAL_ADD_KERNEL(mul_mv_q2_K_f32);
+        GGML_METAL_ADD_KERNEL(mul_mv_q3_K_f32);
+        GGML_METAL_ADD_KERNEL(mul_mv_q4_K_f32);
+        GGML_METAL_ADD_KERNEL(mul_mv_q5_K_f32);
+        GGML_METAL_ADD_KERNEL(mul_mv_q6_K_f32);
+        if ([ctx->device supportsFamily:MTLGPUFamilyApple7]) {
             GGML_METAL_ADD_KERNEL(mul_mm_f32_f32);
             GGML_METAL_ADD_KERNEL(mul_mm_f16_f32);
             GGML_METAL_ADD_KERNEL(mul_mm_q4_0_f32);
@@ -284,6 +285,7 @@ struct ggml_metal_context * ggml_metal_init(int n_cb) {
             GGML_METAL_ADD_KERNEL(mul_mm_q4_K_f32);
             GGML_METAL_ADD_KERNEL(mul_mm_q5_K_f32);
             GGML_METAL_ADD_KERNEL(mul_mm_q6_K_f32);
+        }
         GGML_METAL_ADD_KERNEL(rope_f32);
         GGML_METAL_ADD_KERNEL(rope_f16);
         GGML_METAL_ADD_KERNEL(alibi_f32);
@@ -296,8 +298,21 @@ struct ggml_metal_context * ggml_metal_init(int n_cb) {
 #undef GGML_METAL_ADD_KERNEL
     }
 
-    GGML_METAL_LOG_INFO("%s: hasUnifiedMemory              = %s\n",       __func__, ctx->device.hasUnifiedMemory ? "true" : "false");
 #if TARGET_OS_OSX
+    // print MTL GPU family:
+    GGML_METAL_LOG_INFO("%s: GPU name:   %s\n", __func__, [[ctx->device name] UTF8String]);
+
+    // determine max supported GPU family
+    // https://developer.apple.com/metal/Metal-Shading-Language-Specification.pdf
+    // https://developer.apple.com/metal/Metal-Feature-Set-Tables.pdf
+    for (int i = MTLGPUFamilyApple1 + 20; i >= MTLGPUFamilyApple1; --i) {
+        if ([ctx->device supportsFamily:i]) {
+            GGML_METAL_LOG_INFO("%s: GPU family: MTLGPUFamilyApple%d (%d)\n", __func__, i - MTLGPUFamilyApple1 + 1, i);
+            break;
+        }
+    }
+
+    GGML_METAL_LOG_INFO("%s: hasUnifiedMemory              = %s\n",       __func__, ctx->device.hasUnifiedMemory ? "true" : "false");
     GGML_METAL_LOG_INFO("%s: recommendedMaxWorkingSetSize  = %8.2f MB\n", __func__, ctx->device.recommendedMaxWorkingSetSize / 1024.0 / 1024.0);
     if (ctx->device.maxTransferRate != 0) {
         GGML_METAL_LOG_INFO("%s: maxTransferRate               = %8.2f MB/s\n", __func__, ctx->device.maxTransferRate / 1024.0 / 1024.0);
@@ -339,18 +354,19 @@ void ggml_metal_free(struct ggml_metal_context * ctx) {
     GGML_METAL_DEL_KERNEL(get_rows_q6_K);
     GGML_METAL_DEL_KERNEL(rms_norm);
     GGML_METAL_DEL_KERNEL(norm);
-    GGML_METAL_DEL_KERNEL(mul_mat_f32_f32);
-    GGML_METAL_DEL_KERNEL(mul_mat_f16_f32);
-    GGML_METAL_DEL_KERNEL(mul_mat_f16_f32_1row);
-    GGML_METAL_DEL_KERNEL(mul_mat_f16_f32_l4);
-    GGML_METAL_DEL_KERNEL(mul_mat_q4_0_f32);
-    GGML_METAL_DEL_KERNEL(mul_mat_q4_1_f32);
-    GGML_METAL_DEL_KERNEL(mul_mat_q8_0_f32);
-    GGML_METAL_DEL_KERNEL(mul_mat_q2_K_f32);
-    GGML_METAL_DEL_KERNEL(mul_mat_q3_K_f32);
-    GGML_METAL_DEL_KERNEL(mul_mat_q4_K_f32);
-    GGML_METAL_DEL_KERNEL(mul_mat_q5_K_f32);
-    GGML_METAL_DEL_KERNEL(mul_mat_q6_K_f32);
+    GGML_METAL_DEL_KERNEL(mul_mv_f32_f32);
+    GGML_METAL_DEL_KERNEL(mul_mv_f16_f32);
+    GGML_METAL_DEL_KERNEL(mul_mv_f16_f32_1row);
+    GGML_METAL_DEL_KERNEL(mul_mv_f16_f32_l4);
+    GGML_METAL_DEL_KERNEL(mul_mv_q4_0_f32);
+    GGML_METAL_DEL_KERNEL(mul_mv_q4_1_f32);
+    GGML_METAL_DEL_KERNEL(mul_mv_q8_0_f32);
+    GGML_METAL_DEL_KERNEL(mul_mv_q2_K_f32);
+    GGML_METAL_DEL_KERNEL(mul_mv_q3_K_f32);
+    GGML_METAL_DEL_KERNEL(mul_mv_q4_K_f32);
+    GGML_METAL_DEL_KERNEL(mul_mv_q5_K_f32);
+    GGML_METAL_DEL_KERNEL(mul_mv_q6_K_f32);
+    if ([ctx->device supportsFamily:MTLGPUFamilyApple7]) {
         GGML_METAL_DEL_KERNEL(mul_mm_f32_f32);
         GGML_METAL_DEL_KERNEL(mul_mm_f16_f32);
         GGML_METAL_DEL_KERNEL(mul_mm_q4_0_f32);
@@ -361,6 +377,7 @@ void ggml_metal_free(struct ggml_metal_context * ctx) {
         GGML_METAL_DEL_KERNEL(mul_mm_q4_K_f32);
         GGML_METAL_DEL_KERNEL(mul_mm_q5_K_f32);
         GGML_METAL_DEL_KERNEL(mul_mm_q6_K_f32);
+    }
     GGML_METAL_DEL_KERNEL(rope_f32);
     GGML_METAL_DEL_KERNEL(rope_f16);
     GGML_METAL_DEL_KERNEL(alibi_f32);
@@ -762,8 +779,8 @@ void ggml_metal_graph_compute(
                         } break;
                     case GGML_OP_CONCAT:
                         {
+                            const int64_t nb = ne00;
 
-                            int64_t nb = ne00;
                             [encoder setComputePipelineState:ctx->pipeline_concat];
                             [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0];
                             [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1];
@@ -795,6 +812,7 @@ void ggml_metal_graph_compute(
                             [encoder setBytes:&nb   length:sizeof(nb)   atIndex:27];
 
                             const int nth = MIN(1024, ne0);
+
                             [encoder dispatchThreadgroups:MTLSizeMake(ne1, ne2, ne3) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)];
                         } break;
                     case GGML_OP_ADD:
@@ -892,9 +910,10 @@ void ggml_metal_graph_compute(
                             [encoder setBuffer:id_dst  offset:offs_dst  atIndex:1];
                             [encoder setBytes:&scale length:sizeof(scale) atIndex:2];
 
-                            const int64_t n = ggml_nelements(dst)/4;
+                            const int64_t n = ggml_nelements(dst);
+                            GGML_ASSERT(n % 4 == 0);
 
-                            [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)];
+                            [encoder dispatchThreadgroups:MTLSizeMake(n/4, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)];
                         } break;
                     case GGML_OP_UNARY:
                         switch (ggml_get_unary_op(gf->nodes[i])) {
@@ -904,9 +923,10 @@ void ggml_metal_graph_compute(
                                     [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0];
                                     [encoder setBuffer:id_dst  offset:offs_dst  atIndex:1];
 
-                                    const int64_t n = ggml_nelements(dst)/4;
+                                    const int64_t n = ggml_nelements(dst);
+                                    GGML_ASSERT(n % 4 == 0);
 
-                                    [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)];
+                                    [encoder dispatchThreadgroups:MTLSizeMake(n/4, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)];
                                 } break;
                             case GGML_UNARY_OP_RELU:
                                 {
@@ -924,9 +944,10 @@ void ggml_metal_graph_compute(
                                     [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0];
                                     [encoder setBuffer:id_dst  offset:offs_dst  atIndex:1];
 
-                                    const int64_t n = ggml_nelements(dst)/4;
+                                    const int64_t n = ggml_nelements(dst);
+                                    GGML_ASSERT(n % 4 == 0);
 
-                                    [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)];
+                                    [encoder dispatchThreadgroups:MTLSizeMake(n/4, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)];
                                 } break;
                             default:
                                 {
@@ -986,21 +1007,46 @@ void ggml_metal_graph_compute(
                         } break;
                     case GGML_OP_MUL_MAT:
                         {
-                            // TODO: needs to be updated after PR: https://github.com/ggerganov/ggml/pull/224
-
                             GGML_ASSERT(ne00 == ne10);
-                            // GGML_ASSERT(ne02 == ne12); // Should be checked on individual data types until broadcast is implemented everywhere
-                            uint gqa = ne12/ne02;
                             GGML_ASSERT(ne03 == ne13);
 
+                            const uint gqa = ne12/ne02;
+
+                            // find the break-even point where the matrix-matrix kernel becomes more efficient compared
+                            // to the matrix-vector kernel
+                            int ne11_mm_min = 1;
+
+#if 0
+                            // the numbers below are measured on M2 Ultra for 7B and 13B models
+                            // these numbers do not translate to other devices or model sizes
+                            // TODO: need to find a better approach
+                            if ([ctx->device.name isEqualToString:@"Apple M2 Ultra"]) {
+                                switch (src0t) {
+                                    case GGML_TYPE_F16:  ne11_mm_min = 2;  break;
+                                    case GGML_TYPE_Q8_0: ne11_mm_min = 7;  break;
+                                    case GGML_TYPE_Q2_K: ne11_mm_min = 15; break;
+                                    case GGML_TYPE_Q3_K: ne11_mm_min = 7;  break;
+                                    case GGML_TYPE_Q4_0:
+                                    case GGML_TYPE_Q4_1: ne11_mm_min = 15; break;
+                                    case GGML_TYPE_Q4_K: ne11_mm_min = 11; break;
+                                    case GGML_TYPE_Q5_0:                          // not tested yet
+                                    case GGML_TYPE_Q5_1: ne11_mm_min = 13; break; // not tested yet
+                                    case GGML_TYPE_Q5_K: ne11_mm_min = 7;  break;
+                                    case GGML_TYPE_Q6_K: ne11_mm_min = 7;  break;
+                                    default:             ne11_mm_min = 1;  break;
+                                }
+                            }
+#endif
+
                             // for now the matrix-matrix multiplication kernel only works on A14+/M1+ SoCs
                             // AMD GPU and older A-chips will reuse matrix-vector multiplication kernel
-                            if (!ggml_is_transposed(src0) &&
+                            if ([ctx->device supportsFamily:MTLGPUFamilyApple7] &&
+                                !ggml_is_transposed(src0) &&
                                 !ggml_is_transposed(src1) &&
                                 src1t == GGML_TYPE_F32 &&
-                                [ctx->device supportsFamily:MTLGPUFamilyApple7] &&
-                                ne00%32 == 0 &&
-                                ne11 > 2) {
+                                ne00 % 32 == 0 && ne00 >= 64 &&
+                                ne11 > ne11_mm_min) {
+                                //printf("matrix: ne00 = %6d, ne01 = %6d, ne02 = %6d, ne11 = %6d, ne12 = %6d\n", ne00, ne01, ne02, ne11, ne12);
                                 switch (src0->type) {
                                     case GGML_TYPE_F32:  [encoder setComputePipelineState:ctx->pipeline_mul_mm_f32_f32];  break;
                                     case GGML_TYPE_F16:  [encoder setComputePipelineState:ctx->pipeline_mul_mm_f16_f32];  break;
@@ -1034,12 +1080,13 @@ void ggml_metal_graph_compute(
                                 int nth0 = 32;
                                 int nth1 = 1;
                                 int nrows = 1;
+                                //printf("vector: ne00 = %6d, ne01 = %6d, ne02 = %6d, ne11 = %6d, ne12 = %6d\n", ne00, ne01, ne02, ne11, ne12);
 
                                 // use custom matrix x vector kernel
                                 switch (src0t) {
                                     case GGML_TYPE_F32:
                                         {
-                                            [encoder setComputePipelineState:ctx->pipeline_mul_mat_f32_f32];
+                                            [encoder setComputePipelineState:ctx->pipeline_mul_mv_f32_f32];
                                             nrows = 4;
                                         } break;
                                     case GGML_TYPE_F16:
@@ -1047,12 +1094,12 @@ void ggml_metal_graph_compute(
                                             nth0 = 32;
                                             nth1 = 1;
                                             if (ne11 * ne12 < 4) {
-                                                [encoder setComputePipelineState:ctx->pipeline_mul_mat_f16_f32_1row];
+                                                [encoder setComputePipelineState:ctx->pipeline_mul_mv_f16_f32_1row];
                                             } else if (ne00 >= 128 && ne01 >= 8 && ne00%4 == 0) {
-                                                [encoder setComputePipelineState:ctx->pipeline_mul_mat_f16_f32_l4];
+                                                [encoder setComputePipelineState:ctx->pipeline_mul_mv_f16_f32_l4];
                                                 nrows = ne11;
                                             } else {
-                                                [encoder setComputePipelineState:ctx->pipeline_mul_mat_f16_f32];
+                                                [encoder setComputePipelineState:ctx->pipeline_mul_mv_f16_f32];
                                                 nrows = 4;
                                             }
                                         } break;
@@ -1063,7 +1110,7 @@ void ggml_metal_graph_compute(
 
                                             nth0 = 8;
                                             nth1 = 8;
-                                            [encoder setComputePipelineState:ctx->pipeline_mul_mat_q4_0_f32];
+                                            [encoder setComputePipelineState:ctx->pipeline_mul_mv_q4_0_f32];
                                         } break;
                                     case GGML_TYPE_Q4_1:
                                         {
@@ -1072,7 +1119,7 @@ void ggml_metal_graph_compute(
 
                                             nth0 = 8;
                                             nth1 = 8;
-                                            [encoder setComputePipelineState:ctx->pipeline_mul_mat_q4_1_f32];
+                                            [encoder setComputePipelineState:ctx->pipeline_mul_mv_q4_1_f32];
                                         } break;
                                     case GGML_TYPE_Q8_0:
                                         {
@@ -1081,7 +1128,7 @@ void ggml_metal_graph_compute(
 
                                             nth0 = 8;
                                             nth1 = 8;
-                                            [encoder setComputePipelineState:ctx->pipeline_mul_mat_q8_0_f32];
+                                            [encoder setComputePipelineState:ctx->pipeline_mul_mv_q8_0_f32];
                                         } break;
                                     case GGML_TYPE_Q2_K:
                                         {
@@ -1090,7 +1137,7 @@ void ggml_metal_graph_compute(
 
                                             nth0 = 2;
                                             nth1 = 32;
-                                            [encoder setComputePipelineState:ctx->pipeline_mul_mat_q2_K_f32];
+                                            [encoder setComputePipelineState:ctx->pipeline_mul_mv_q2_K_f32];
                                         } break;
                                     case GGML_TYPE_Q3_K:
                                         {
@@ -1099,7 +1146,7 @@ void ggml_metal_graph_compute(
 
                                             nth0 = 2;
                                             nth1 = 32;
-                                            [encoder setComputePipelineState:ctx->pipeline_mul_mat_q3_K_f32];
+                                            [encoder setComputePipelineState:ctx->pipeline_mul_mv_q3_K_f32];
                                         } break;
                                     case GGML_TYPE_Q4_K:
                                         {
@@ -1108,7 +1155,7 @@ void ggml_metal_graph_compute(
 
                                             nth0 = 4; //1;
                                             nth1 = 8; //32;
-                                            [encoder setComputePipelineState:ctx->pipeline_mul_mat_q4_K_f32];
+                                            [encoder setComputePipelineState:ctx->pipeline_mul_mv_q4_K_f32];
                                         } break;
                                     case GGML_TYPE_Q5_K:
                                         {
@@ -1117,7 +1164,7 @@ void ggml_metal_graph_compute(
 
                                             nth0 = 2;
                                             nth1 = 32;
-                                            [encoder setComputePipelineState:ctx->pipeline_mul_mat_q5_K_f32];
+                                            [encoder setComputePipelineState:ctx->pipeline_mul_mv_q5_K_f32];
                                         } break;
                                     case GGML_TYPE_Q6_K:
                                         {
@@ -1126,7 +1173,7 @@ void ggml_metal_graph_compute(
 
                                             nth0 = 2;
                                             nth1 = 32;
-                                            [encoder setComputePipelineState:ctx->pipeline_mul_mat_q6_K_f32];
+                                            [encoder setComputePipelineState:ctx->pipeline_mul_mv_q6_K_f32];
                                         } break;
                                     default:
                                         {
@@ -1208,6 +1255,8 @@ void ggml_metal_graph_compute(
                         } break;
                     case GGML_OP_RMS_NORM:
                         {
+                            GGML_ASSERT(ne00 % 4 == 0);
+
                             float eps;
                             memcpy(&eps, dst->op_params, sizeof(float));
 
@@ -1250,7 +1299,7 @@ void ggml_metal_graph_compute(
 
                             const int nth = MIN(1024, ne00);
 
-                            const int n_past = ((int32_t *) dst->op_params)[0]; UNUSED(n_past);
+                            //const int n_past = ((int32_t *) dst->op_params)[0];
                             const int n_head = ((int32_t *) dst->op_params)[1];
                             float max_bias;
                             memcpy(&max_bias, (int32_t *) dst->op_params + 2, sizeof(float));
@@ -1421,3 +1470,140 @@ void ggml_metal_graph_compute(
 
     }
 }
+
+////////////////////////////////////////////////////////////////////////////////
+
+// backend interface
+
+static const char * ggml_backend_metal_name(ggml_backend_t backend) {
+    return "Metal";
+
+    UNUSED(backend);
+}
+
+static void ggml_backend_metal_free(ggml_backend_t backend) {
+    struct ggml_metal_context * ctx = (struct ggml_metal_context *)backend->context;
+    ggml_metal_free(ctx);
+    free(backend);
+}
+
+static void * ggml_backend_metal_buffer_get_base(ggml_backend_buffer_t buffer) {
+    return (void *)buffer->context;
+}
+
+static void ggml_backend_metal_buffer_free_buffer(ggml_backend_buffer_t buffer) {
+    free(buffer->context);
+    UNUSED(buffer);
+}
+
+static struct ggml_backend_buffer_i metal_backend_buffer_i = {
+    /* .free_buffer    = */ ggml_backend_metal_buffer_free_buffer,
+    /* .get_base       = */ ggml_backend_metal_buffer_get_base,
+    /* .get_alloc_size = */ NULL, // defaults to ggml_nbytes
+    /* .init_tensor    = */ NULL, // no initialization required
+    /* .free_tensor    = */ NULL, // no cleanup required
+};
+
+static ggml_backend_buffer_t ggml_backend_metal_alloc_buffer(ggml_backend_t backend, size_t size) {
+    struct ggml_metal_context * ctx = (struct ggml_metal_context *)backend->context;
+
+    void * data = ggml_metal_host_malloc(size);
+
+    // TODO: set proper name of the buffers
+    ggml_metal_add_buffer(ctx, "backend", data, size, 0);
+
+    return ggml_backend_buffer_init(backend, metal_backend_buffer_i, data, size);
+}
+
+static size_t ggml_backend_metal_get_alignment(ggml_backend_t backend) {
+    return 32;
+    UNUSED(backend);
+}
+
+static void ggml_backend_metal_set_tensor_async(ggml_backend_t backend, struct ggml_tensor * tensor, const void * data, size_t offset, size_t size) {
+    GGML_ASSERT(offset + size <= ggml_nbytes(tensor) && "tensor write out of bounds");
+    GGML_ASSERT(tensor->data != NULL && "tensor not allocated");
+
+    memcpy((char *)tensor->data + offset, data, size);
+
+    UNUSED(backend);
+}
+
+static void ggml_backend_metal_get_tensor_async(ggml_backend_t backend, const struct ggml_tensor * tensor, void * data, size_t offset, size_t size) {
+    GGML_ASSERT(offset + size <= ggml_nbytes(tensor) && "tensor read out of bounds");
+    GGML_ASSERT(tensor->data != NULL && "tensor not allocated");
+
+    memcpy(data, (const char *)tensor->data + offset, size);
+
+    UNUSED(backend);
+}
+
+static void ggml_backend_metal_synchronize(ggml_backend_t backend) {
+    UNUSED(backend);
+}
+
+static void ggml_backend_metal_cpy_tensor_from(ggml_backend_t backend, struct ggml_tensor * src, struct ggml_tensor * dst) {
+    ggml_backend_tensor_get(src, dst->data, 0, ggml_nbytes(src));
+
+    UNUSED(backend);
+}
+
+static void ggml_backend_metal_cpy_tensor_to(ggml_backend_t backend, struct ggml_tensor * src, struct ggml_tensor * dst) {
+    ggml_backend_tensor_set_async(dst, src->data, 0, ggml_nbytes(src));
+
+    UNUSED(backend);
+}
+
+static void ggml_backend_metal_graph_compute(ggml_backend_t backend, struct ggml_cgraph * cgraph) {
+    struct ggml_metal_context * metal_ctx = (struct ggml_metal_context *)backend->context;
+
+    ggml_metal_graph_compute(metal_ctx, cgraph);
+}
+
+static bool ggml_backend_metal_supports_op(ggml_backend_t backend, const struct ggml_tensor * op) {
+    return true;
+    UNUSED(backend);
+    UNUSED(op);
+}
+
+static struct ggml_backend_i metal_backend_i = {
+    /* .get_name            = */ ggml_backend_metal_name,
+    /* .free                = */ ggml_backend_metal_free,
+    /* .alloc_buffer        = */ ggml_backend_metal_alloc_buffer,
+    /* .get_alignment       = */ ggml_backend_metal_get_alignment,
+    /* .set_tensor_async    = */ ggml_backend_metal_set_tensor_async,
+    /* .get_tensor_async    = */ ggml_backend_metal_get_tensor_async,
+    /* .synchronize         = */ ggml_backend_metal_synchronize,
+    /* .cpy_tensor_from     = */ ggml_backend_metal_cpy_tensor_from,
+    /* .cpy_tensor_to       = */ ggml_backend_metal_cpy_tensor_to,
+    /* .graph_plan_create   = */ NULL, // the metal implementation does not require creating graph plans atm
+    /* .graph_plan_free     = */ NULL,
+    /* .graph_plan_compute  = */ NULL,
+    /* .graph_compute       = */ ggml_backend_metal_graph_compute,
+    /* .supports_op         = */ ggml_backend_metal_supports_op,
+};
+
+ggml_backend_t ggml_backend_metal_init(void) {
+    struct ggml_metal_context * ctx = malloc(sizeof(struct ggml_metal_context));
+
+    ctx = ggml_metal_init(GGML_DEFAULT_N_THREADS);
+
+    ggml_backend_t metal_backend = malloc(sizeof(struct ggml_backend));
+
+    *metal_backend = (struct ggml_backend) {
+        /* .interface = */ metal_backend_i,
+        /* .context   = */ ctx,
+    };
+
+    return metal_backend;
+}
+
+bool ggml_backend_is_metal(ggml_backend_t backend) {
+    return backend->iface.get_name == ggml_backend_metal_name;
+}
+
+void ggml_backend_metal_set_n_cb(ggml_backend_t backend, int n_cb) {
+    struct ggml_metal_context * ctx = (struct ggml_metal_context *)backend->context;
+
+    ggml_metal_set_n_cb(ctx, n_cb);
+}

ggml-metal.metal

@@ -347,6 +347,7 @@ kernel void kernel_rms_norm(
         uint   ntg[[threads_per_threadgroup]]) {
     device const float4 * x        = (device const float4 *) ((device const char *) src0 + tgpig*nb01);
     device const float  * x_scalar = (device const float  *) x;
+
     float4 sumf = 0;
     float all_sum = 0;
 
@@ -361,6 +362,7 @@ kernel void kernel_rms_norm(
     }
 
     threadgroup_barrier(mem_flags::mem_threadgroup);
+
     // broadcast, simd group number is ntg / 32
     for (uint i = ntg / 32 / 2; i > 0; i /= 2) {
        if (tpitg < i) {
@@ -368,7 +370,9 @@ kernel void kernel_rms_norm(
        }
     }
     if (tpitg == 0) {
-        for (int i = 4 * (ne00 / 4); i < ne00; i++) {sum[0] += x_scalar[i];}
+        for (int i = 4 * (ne00 / 4); i < ne00; i++) {
+            sum[0] += x_scalar[i];
+        }
         sum[0] /= ne00;
     }
 
@@ -383,7 +387,9 @@ kernel void kernel_rms_norm(
         y[i00] = x[i00] * scale;
     }
     if (tpitg == 0) {
-        for (int i00 = 4 * (ne00 / 4); i00 < ne00; i00++) {y_scalar[i00] = x_scalar[i00] * scale;}
+        for (int i00 = 4 * (ne00 / 4); i00 < ne00; i00++) {
+            y_scalar[i00] = x_scalar[i00] * scale;
+        }
     }
 }
 
@@ -435,18 +441,23 @@ void mul_vec_q_n_f32(device const void * src0, device const float * src1, device
                     int64_t ne00, int64_t ne01, int64_t ne02, int64_t ne10, int64_t ne12, int64_t ne0, int64_t ne1, uint gqa,
                     uint3 tgpig, uint tiisg, uint sgitg) {
     const int nb = ne00/QK4_0;
+
     const int r0 = tgpig.x;
     const int r1 = tgpig.y;
     const int im = tgpig.z;
+
     const int first_row = (r0 * nsg + sgitg) * nr;
+
     const uint offset0 = first_row * nb + im/gqa*(nb*ne0);
+
     device const block_q_type * x = (device const block_q_type *) src0 + offset0;
     device const float        * y = (device const float        *) src1 + r1*ne10 + im*ne00*ne1;
+
     float yl[16]; // src1 vector cache
     float sumf[nr] = {0.f};
 
-    const int ix = tiisg/2;
-    const int il = 8*(tiisg%2);
+    const int ix = (tiisg/2);
+    const int il = (tiisg%2)*8;
 
     device const float * yb = y + ix * QK4_0 + il;
 
@@ -457,6 +468,7 @@ void mul_vec_q_n_f32(device const void * src0, device const float * src1, device
             sumy += yb[i] + yb[i+1];
             yl[i+0] = yb[i+ 0];
             yl[i+1] = yb[i+ 1]/256.f;
+
             sumy += yb[i+16] + yb[i+17];
             yl[i+8] = yb[i+16]/16.f;
             yl[i+9] = yb[i+17]/4096.f;
@@ -472,12 +484,12 @@ void mul_vec_q_n_f32(device const void * src0, device const float * src1, device
     for (int row = 0; row < nr; ++row) {
         const float tot = simd_sum(sumf[row]);
         if (tiisg == 0 && first_row + row < ne01) {
-            dst[r1*ne0 + im*ne0*ne1 + first_row + row] = tot;
+            dst[im*ne0*ne1 + r1*ne0 + first_row + row] = tot;
         }
     }
 }
 
-kernel void kernel_mul_mat_q4_0_f32(
+kernel void kernel_mul_mv_q4_0_f32(
         device const  void * src0,
         device const float * src1,
         device       float * dst,
@@ -495,7 +507,7 @@ kernel void kernel_mul_mat_q4_0_f32(
     mul_vec_q_n_f32<block_q4_0, N_DST, N_SIMDGROUP, N_SIMDWIDTH>(src0,src1,dst,ne00,ne01,ne02,ne10,ne12,ne0,ne1,gqa,tgpig,tiisg,sgitg);
 }
 
-kernel void kernel_mul_mat_q4_1_f32(
+kernel void kernel_mul_mv_q4_1_f32(
         device const  void * src0,
         device const float * src1,
         device       float * dst,
@@ -515,7 +527,7 @@ kernel void kernel_mul_mat_q4_1_f32(
 
 #define NB_Q8_0 8
 
-kernel void kernel_mul_mat_q8_0_f32(
+kernel void kernel_mul_mv_q8_0_f32(
         device const  void * src0,
         device const float * src1,
         device       float * dst,
@@ -579,7 +591,7 @@ kernel void kernel_mul_mat_q8_0_f32(
 
 #define N_F32_F32 4
 
-kernel void kernel_mul_mat_f32_f32(
+kernel void kernel_mul_mv_f32_f32(
         device const  char * src0,
         device const  char * src1,
         device       float * dst,
@@ -650,7 +662,7 @@ kernel void kernel_mul_mat_f32_f32(
     }
 }
 
-kernel void kernel_mul_mat_f16_f32_1row(
+kernel void kernel_mul_mv_f16_f32_1row(
         device const  char * src0,
         device const  char * src1,
         device       float * dst,
@@ -704,7 +716,7 @@ kernel void kernel_mul_mat_f16_f32_1row(
 
 #define N_F16_F32 4
 
-kernel void kernel_mul_mat_f16_f32(
+kernel void kernel_mul_mv_f16_f32(
         device const  char * src0,
         device const  char * src1,
         device       float * dst,
@@ -776,7 +788,7 @@ kernel void kernel_mul_mat_f16_f32(
 }
 
 // Assumes row size (ne00) is a multiple of 4
-kernel void kernel_mul_mat_f16_f32_l4(
+kernel void kernel_mul_mv_f16_f32_l4(
         device const  char * src0,
         device const  char * src1,
         device       float * dst,
@@ -1300,7 +1312,7 @@ static inline uchar4 get_scale_min_k4(int j, device const uint8_t * q) {
 
 //====================================== dot products =========================
 
-kernel void kernel_mul_mat_q2_K_f32(
+kernel void kernel_mul_mv_q2_K_f32(
         device const  void * src0,
         device const float * src1,
         device       float * dst,
@@ -1444,7 +1456,7 @@ kernel void kernel_mul_mat_q2_K_f32(
 }
 
 #if QK_K == 256
-kernel void kernel_mul_mat_q3_K_f32(
+kernel void kernel_mul_mv_q3_K_f32(
         device const  void * src0,
         device const float * src1,
         device       float * dst,
@@ -1596,7 +1608,7 @@ kernel void kernel_mul_mat_q3_K_f32(
     }
 }
 #else
-kernel void kernel_mul_mat_q3_K_f32(
+kernel void kernel_mul_mv_q3_K_f32(
         device const  void * src0,
         device const float * src1,
         device       float * dst,
@@ -1667,7 +1679,7 @@ kernel void kernel_mul_mat_q3_K_f32(
 #endif
 
 #if QK_K == 256
-kernel void kernel_mul_mat_q4_K_f32(
+kernel void kernel_mul_mv_q4_K_f32(
         device const  void * src0,
         device const float * src1,
         device       float * dst,
@@ -1773,7 +1785,7 @@ kernel void kernel_mul_mat_q4_K_f32(
     }
 }
 #else
-kernel void kernel_mul_mat_q4_K_f32(
+kernel void kernel_mul_mv_q4_K_f32(
         device const  void * src0,
         device const float * src1,
         device       float * dst,
@@ -1862,7 +1874,7 @@ kernel void kernel_mul_mat_q4_K_f32(
 }
 #endif
 
-kernel void kernel_mul_mat_q5_K_f32(
+kernel void kernel_mul_mv_q5_K_f32(
         device const  void * src0,
         device const float * src1,
         device       float * dst,
@@ -2035,7 +2047,7 @@ kernel void kernel_mul_mat_q5_K_f32(
 
 }
 
-kernel void kernel_mul_mat_q6_K_f32(
+kernel void kernel_mul_mv_q6_K_f32(
         device const  void * src0,
         device const float * src1,
         device       float * dst,
@@ -2373,7 +2385,7 @@ kernel void kernel_get_rows(
 }
 
 #define BLOCK_SIZE_M 64 // 8 simdgroup matrices from matrix A
-#define BLOCK_SIZE_N 32 // 4 simdgroup matrices from matrix A
+#define BLOCK_SIZE_N 32 // 4 simdgroup matrices from matrix B
 #define BLOCK_SIZE_K 32
 #define THREAD_MAT_M 4 // each thread take 4 simdgroup matrices from matrix A
 #define THREAD_MAT_N 2 // each thread take 2 simdgroup matrices from matrix B
@@ -2410,9 +2422,11 @@ kernel void kernel_mul_mm(device const  uchar * src0,
     const uint r0 = tgpig.y;
     const uint r1 = tgpig.x;
     const uint im = tgpig.z;
+
     // if this block is of 64x32 shape or smaller
     short n_rows = (ne0 - r0 * BLOCK_SIZE_M < BLOCK_SIZE_M) ? (ne0 - r0 * BLOCK_SIZE_M) : BLOCK_SIZE_M;
     short n_cols = (ne1 - r1 * BLOCK_SIZE_N < BLOCK_SIZE_N) ? (ne1 - r1 * BLOCK_SIZE_N) : BLOCK_SIZE_N;
+
     // a thread shouldn't load data outside of the matrix
     short thread_row = ((short)tiitg/THREAD_PER_ROW) < n_rows ? ((short)tiitg/THREAD_PER_ROW) : n_rows - 1;
     short thread_col = ((short)tiitg/THREAD_PER_COL) < n_cols ? ((short)tiitg/THREAD_PER_COL) : n_cols - 1;
@@ -2440,22 +2454,26 @@ kernel void kernel_mul_mm(device const  uchar * src0,
         half4x4 temp_a;
         dequantize_func(x, il, temp_a);
         threadgroup_barrier(mem_flags::mem_threadgroup);
+
         #pragma unroll(16)
         for (int i = 0; i < 16; i++) {
             *(sa + SG_MAT_SIZE * ((tiitg / THREAD_PER_ROW / 8) \
-            + 16 * (tiitg % THREAD_PER_ROW) + 8 * (i / 8)) \
+            +                     (tiitg % THREAD_PER_ROW) * 16 + (i / 8) * 8) \
             +                     (tiitg / THREAD_PER_ROW) % 8  + (i & 7) * 8) = temp_a[i/4][i%4];
         }
-        *(threadgroup float2x4 *)(sb + (tiitg % THREAD_PER_COL) * 8 * 32 + 8 * (tiitg / THREAD_PER_COL)) \
-                = *((device float2x4 *)y);
+
+        *(threadgroup float2x4 *)(sb + (tiitg % THREAD_PER_COL) * 8 * 32 + 8 * (tiitg / THREAD_PER_COL)) = *((device float2x4 *)y);
+
         il = (il + 2 < nl) ? il + 2 : il % 2;
         x  = (il < 2) ? x + (2+nl-1)/nl : x;
         y += BLOCK_SIZE_K;
 
         threadgroup_barrier(mem_flags::mem_threadgroup);
+
         // load matrices from threadgroup memory and conduct outer products
         threadgroup half  * lsma = (sa + THREAD_MAT_M * SG_MAT_SIZE * (sgitg % 2));
         threadgroup float * lsmb = (sb + THREAD_MAT_N * SG_MAT_SIZE * (sgitg / 2));
+
         #pragma unroll(4)
         for (int ik = 0; ik < BLOCK_SIZE_K / 8; ik++) {
             #pragma unroll(4)
@@ -2470,6 +2488,7 @@ kernel void kernel_mul_mm(device const  uchar * src0,
 
             lsma += BLOCK_SIZE_M / SG_MAT_ROW * SG_MAT_SIZE;
             lsmb += BLOCK_SIZE_N / SG_MAT_ROW * SG_MAT_SIZE;
+
             #pragma unroll(8)
             for (int i = 0; i < 8; i++){
                 simdgroup_multiply_accumulate(c_res[i], mb[i/4], ma[i%4], c_res[i]);
@@ -2478,7 +2497,7 @@ kernel void kernel_mul_mm(device const  uchar * src0,
     }
 
     if ((r0 + 1) * BLOCK_SIZE_M <= ne0 && (r1 + 1) * BLOCK_SIZE_N <= ne1) {
-        device float *C = dst + BLOCK_SIZE_M * r0 + 32 * (sgitg&1) \
+        device float * C = dst + (BLOCK_SIZE_M * r0 + 32 * (sgitg &  1)) \
                                + (BLOCK_SIZE_N * r1 + 16 * (sgitg >> 1)) * ne0 + im*ne1*ne0;
         for (int i = 0; i < 8; i++) {
             simdgroup_store(c_res[i], C + 8 * (i%4) + 8 * ne0 * (i/4), ne0);
@@ -2493,7 +2512,8 @@ kernel void kernel_mul_mm(device const  uchar * src0,
         }
 
         threadgroup_barrier(mem_flags::mem_threadgroup);
-        device float *C = dst + BLOCK_SIZE_M * r0 + (BLOCK_SIZE_N * r1) * ne0 + im*ne1*ne0;
+
+        device float * C = dst + (BLOCK_SIZE_M * r0) + (BLOCK_SIZE_N * r1) * ne0 + im*ne1*ne0;
         if (sgitg == 0) {
             for (int i = 0; i < n_rows; i++) {
                 for (int j = tiitg; j < n_cols; j += BLOCK_SIZE_N) {

ggml-opencl.cpp

@@ -19,7 +19,7 @@
 #pragma warning(disable: 4244 4267) // possible loss of data
 #endif
 
-#define CL_DMMV_BLOCK_SIZE 32
+#define CL_DMMV_LOCAL_SIZE 32
 
 #ifndef K_QUANTS_PER_ITERATION
 #define K_QUANTS_PER_ITERATION 1
@@ -338,7 +338,7 @@ __kernel void dequantize_mul_mat_vec_q2_K(__global const struct block_q2_K * xx,
     const int row = get_group_id(0);
 
     const int num_blocks_per_row = ncols / QK_K;
-    const int ib0 = row*num_blocks_per_row;
+    const int ib0 = row*num_blocks_per_row + get_global_offset(0);
 
     __global const struct block_q2_K * x = xx + ib0;
 
@@ -413,7 +413,7 @@ __kernel void dequantize_mul_mat_vec_q3_K(__global const struct block_q3_K * xx,
     const int row = get_group_id(0);
 
     const int num_blocks_per_row = ncols / QK_K;
-    const int ib0 = row*num_blocks_per_row;
+    const int ib0 = row*num_blocks_per_row + get_global_offset(0);
 
     __global const struct block_q3_K * x = xx + ib0;
 
@@ -489,7 +489,7 @@ __kernel void dequantize_mul_mat_vec_q4_K(__global const struct block_q4_K * xx,
 
     const int row = get_group_id(0);
     const int num_blocks_per_row = ncols / QK_K;
-    const int ib0 = row*num_blocks_per_row;
+    const int ib0 = row*num_blocks_per_row + get_global_offset(0);
 
     const int tid = get_local_id(0)/K_QUANTS_PER_ITERATION;  // 0...15
     const int ix  = get_local_id(0)%K_QUANTS_PER_ITERATION;
@@ -562,7 +562,7 @@ __kernel void dequantize_mul_mat_vec_q5_K(__global const struct block_q5_K * xx,
 
     const int row = get_group_id(0);
     const int num_blocks_per_row = ncols / QK_K;
-    const int ib0 = row*num_blocks_per_row;
+    const int ib0 = row*num_blocks_per_row + get_global_offset(0);
 
     const int tid = get_local_id(0)/2;  // 0...15
     const int ix  = get_local_id(0)%2;
@@ -641,7 +641,7 @@ __kernel void dequantize_mul_mat_vec_q6_K(__global const struct block_q6_K * xx,
     const int row = get_group_id(0);
 
     const int num_blocks_per_row = ncols / QK_K;
-    const int ib0 = row*num_blocks_per_row;
+    const int ib0 = row*num_blocks_per_row + get_global_offset(0);
 
     __global const struct block_q6_K * x = xx + ib0;
 
@@ -745,19 +745,21 @@ __kernel void KERNEL_NAME(__global X_TYPE* x, __global float* y) {
 
 std::string dequant_mul_mat_vec_template = MULTILINE_QUOTE(
 __kernel void KERNEL_NAME(__global X_TYPE* x, __local float* tmp, __global float* y, __global float* dst, const int ncols) {
-    const int block_size = get_local_size(0);
+    const int local_size = get_local_size(0);
     const int row = get_group_id(0);
     const int tid = get_local_id(0);
 
     const uint qk = QUANT_K;
     const uint qr = QUANT_R;
 
+    const int col_step = local_size * 2;
     const int y_offset = qr == 1 ? 1 : qk/2;
 
+    x += get_global_offset(0);
+
     tmp[tid] = 0;
 
-    for (int i = 0; i < ncols/block_size; i += 2) {
-        const int col = i*block_size + 2*tid;
+    for (int col = tid*2; col < ncols; col += col_step) {
         const int ib = (row*ncols + col)/qk; // block index
         const int iqs = (col%qk)/qr; // quant index
         const int iybs = col - col%qk; // y block start index
@@ -773,7 +775,7 @@ __kernel void KERNEL_NAME(__global X_TYPE* x, __local float* tmp, __global float
 
     // sum up partial sums and write back result
     barrier(CLK_LOCAL_MEM_FENCE);
-    for (int s=block_size/2; s>0; s>>=1) {
+    for (int s=local_size/2; s>0; s>>=1) {
         if (tid < s) {
             tmp[tid] += tmp[tid + s];
         }
@@ -1704,7 +1706,7 @@ static void ggml_cl_mul_mat_q_f32(const ggml_tensor * src0, const ggml_tensor *
     const int nb2  = dst->nb[2];
     const int nb3  = dst->nb[3];
     const ggml_type type = src0->type;
-    const bool mul_mat_vec = ne11 == 1;
+    const bool mul_mat_vec = ne11 == 1 && ne00%2 == 0;
 
     const int64_t r2 = ne12 / ne02;
     const int64_t r3 = ne13 / ne03;
@@ -1737,7 +1739,7 @@ static void ggml_cl_mul_mat_q_f32(const ggml_tensor * src0, const ggml_tensor *
     GGML_ASSERT(to_fp32_cl != nullptr);
 
     const size_t global_denom = ggml_cl_global_denom(type);
-    const size_t local = ggml_cl_local_size(type);
+    const size_t local = mul_mat_vec ? CL_DMMV_LOCAL_SIZE : ggml_cl_local_size(type);
 
     size_t ev_idx = 0;
     std::vector<cl_event> events;
@@ -1770,8 +1772,8 @@ static void ggml_cl_mul_mat_q_f32(const ggml_tensor * src0, const ggml_tensor *
                 CL_CHECK(ggml_cl_h2d_tensor_2d(queue, d_Y, 0, src1, i13, i12, events.data() + ev_idx++));
 
                 // compute
-                const size_t global = ne01 * CL_DMMV_BLOCK_SIZE;
-                const size_t local = CL_DMMV_BLOCK_SIZE;
+                const size_t global = ne01 * local;
+                const size_t offset = src0->backend == GGML_BACKEND_GPU ? (i03 * ne02 + i02) * x_bps : 0;
                 const cl_int ncols = ne00;
                 events.emplace_back();
                 CL_CHECK(clSetKernelArg(*dmmv, 0, sizeof(cl_mem), &d_Q));
@@ -1779,7 +1781,7 @@ static void ggml_cl_mul_mat_q_f32(const ggml_tensor * src0, const ggml_tensor *
                 CL_CHECK(clSetKernelArg(*dmmv, 2, sizeof(cl_mem), &d_Y));
                 CL_CHECK(clSetKernelArg(*dmmv, 3, sizeof(cl_mem), &d_D));
                 CL_CHECK(clSetKernelArg(*dmmv, 4, sizeof(cl_int), &ncols));
-                CL_CHECK(clEnqueueNDRangeKernel(queue, *dmmv, 1, NULL, &global, &local, events.size() - 1, events.data(), events.data() + ev_idx++));
+                CL_CHECK(clEnqueueNDRangeKernel(queue, *dmmv, 1, &offset, &global, &local, events.size() - 1, events.data(), events.data() + ev_idx++));
             } else { // general dequantization kernel + CLBlast matrix matrix multiplication
                 // convert src0 to fp32 on device
                 const size_t global = x_ne / global_denom;

ggml.c

@@ -163,40 +163,16 @@ typedef void * thread_ret_t;
 
 #define GGML_PRINT(...) printf(__VA_ARGS__)
 
+//
+// end of logging block
+//
+
 #ifdef GGML_USE_ACCELERATE
 // uncomment to use vDSP for soft max computation
 // note: not sure if it is actually faster
 //#define GGML_SOFT_MAX_ACCELERATE
 #endif
 
-//
-// logging
-//
-
-#if (GGML_DEBUG >= 1)
-#define GGML_PRINT_DEBUG(...) printf(__VA_ARGS__)
-#else
-#define GGML_PRINT_DEBUG(...)
-#endif
-
-#if (GGML_DEBUG >= 5)
-#define GGML_PRINT_DEBUG_5(...) printf(__VA_ARGS__)
-#else
-#define GGML_PRINT_DEBUG_5(...)
-#endif
-
-#if (GGML_DEBUG >= 10)
-#define GGML_PRINT_DEBUG_10(...) printf(__VA_ARGS__)
-#else
-#define GGML_PRINT_DEBUG_10(...)
-#endif
-
-#define GGML_PRINT(...) printf(__VA_ARGS__)
-
-//
-// end of logging block
-//
-
 #if defined(_MSC_VER) || defined(__MINGW32__)
 #define GGML_ALIGNED_MALLOC(size) _aligned_malloc(size, GGML_MEM_ALIGN)
 #define GGML_ALIGNED_FREE(ptr)    _aligned_free(ptr)
@@ -4952,6 +4928,7 @@ static struct ggml_tensor * ggml_new_tensor_impl(
     *result = (struct ggml_tensor) {
         /*.type         =*/ type,
         /*.backend      =*/ GGML_BACKEND_CPU,
+        /*.buffer       =*/ NULL,
         /*.n_dims       =*/ n_dims,
         /*.ne           =*/ { 1, 1, 1, 1 },
         /*.nb           =*/ { 0, 0, 0, 0 },
@@ -5518,6 +5495,39 @@ struct ggml_tensor * ggml_view_tensor(
     return result;
 }
 
+struct ggml_tensor * ggml_get_first_tensor(struct ggml_context * ctx) {
+    struct ggml_object * obj = ctx->objects_begin;
+
+    char * const mem_buffer = ctx->mem_buffer;
+
+    while (obj != NULL) {
+        if (obj->type == GGML_OBJECT_TENSOR) {
+            return (struct ggml_tensor *)(mem_buffer + obj->offs);
+        }
+
+        obj = obj->next;
+    }
+
+    return NULL;
+}
+
+struct ggml_tensor * ggml_get_next_tensor(struct ggml_context * ctx, struct ggml_tensor * tensor) {
+    struct ggml_object * obj = (struct ggml_object *) ((char *)tensor - GGML_OBJECT_SIZE);
+    obj = obj->next;
+
+    char * const mem_buffer = ctx->mem_buffer;
+
+    while (obj != NULL) {
+        if (obj->type == GGML_OBJECT_TENSOR) {
+            return (struct ggml_tensor *)(mem_buffer + obj->offs);
+        }
+
+        obj = obj->next;
+    }
+
+    return NULL;
+}
+
 struct ggml_tensor * ggml_get_tensor(struct ggml_context * ctx, const char * name) {
     struct ggml_object * obj = ctx->objects_begin;
 
@@ -8700,6 +8710,7 @@ void ggml_set_param(
 
     GGML_ASSERT(tensor->grad == NULL);
     tensor->grad = ggml_dup_tensor(ctx, tensor);
+    ggml_format_name(tensor->grad, "%s (grad)", tensor->name);
 }
 
 // ggml_compute_forward_dup
@@ -13112,24 +13123,22 @@ static void ggml_compute_forward_alibi_f32(
         return;
     }
 
-    const int n_past = ((int32_t *) dst->op_params)[0]; UNUSED(n_past);
+    //const int n_past = ((int32_t *) dst->op_params)[0];
     const int n_head = ((int32_t *) dst->op_params)[1];
     float max_bias;
     memcpy(&max_bias, (int32_t *) dst->op_params + 2, sizeof(float));
 
-    assert(n_past >= 0);
+    const int64_t ne0 = src0->ne[0]; // all_seq_len = n_past + ne1
+    const int64_t ne1 = src0->ne[1]; // seq_len_without_past
+    const int64_t ne2 = src0->ne[2]; // n_head -> this is k
+    //const int64_t ne3 = src0->ne[3]; // 1 -> bsz
 
-    const int ne0 = src0->ne[0]; // all_seq_len = n_past + ne1
-    const int ne1 = src0->ne[1]; // seq_len_without_past
-    const int ne2 = src0->ne[2]; // n_head -> this is k
-    //const int ne3 = src0->ne[3]; // 1 -> bsz
+    const int64_t n  = ggml_nrows(src0);
+    const int64_t ne2_ne3 = n/ne1; // ne2*ne3
 
-    const int n  = ggml_nrows(src0);
-    const int ne2_ne3 = n/ne1; // ne2*ne3
-
-    const int nb0 = src0->nb[0];
-    const int nb1 = src0->nb[1];
-    const int nb2 = src0->nb[2];
+    const size_t nb0 = src0->nb[0];
+    const size_t nb1 = src0->nb[1];
+    const size_t nb2 = src0->nb[2];
     //const int nb3 = src0->nb[3];
 
     GGML_ASSERT(nb0 == sizeof(float));
@@ -13141,9 +13150,9 @@ static void ggml_compute_forward_alibi_f32(
     const float m0 = powf(2.0f, -(max_bias) / n_heads_log2_floor);
     const float m1 = powf(2.0f, -(max_bias / 2.0f) / n_heads_log2_floor);
 
-    for (int i = 0; i < ne0; i++) {
-        for (int j = 0; j < ne1; j++) {
-            for (int k = 0; k < ne2_ne3; k++) {
+    for (int64_t i = 0; i < ne0; i++) {
+        for (int64_t j = 0; j < ne1; j++) {
+            for (int64_t k = 0; k < ne2_ne3; k++) {
                 float * const src = (float *)((char *) src0->data + i*nb0 + j*nb1 + k*nb2);
                 float *      pdst = (float *)((char *)  dst->data + i*nb0 + j*nb1 + k*nb2);
 
@@ -13158,7 +13167,6 @@ static void ggml_compute_forward_alibi_f32(
                 }
 
                 pdst[0] = i * m_k + src[0];
-
             }
         }
     }
@@ -14545,7 +14553,7 @@ static void ggml_compute_forward_conv_2d_f16_f32(
     int64_t t0 = ggml_perf_time_us();
     UNUSED(t0);
 
-    GGML_TENSOR_BINARY_OP_LOCALS
+    GGML_TENSOR_BINARY_OP_LOCALS;
 
     const int ith = params->ith;
     const int nth = params->nth;
@@ -20299,6 +20307,10 @@ static enum ggml_opt_result ggml_opt_lbfgs(
         ggml_vec_cpy_f32(nx, xp, x);
         ggml_vec_cpy_f32(nx, gp, g);
 
+        // TODO: instead of passing &cancel here, use the return code of the linesearch
+        //       to determine if the optimization should be cancelled
+        //       this is a simple change, but not doing this atm, since I don't have a nice
+        //       way to test and don't want to break something with so many changes lined up
         ls = linesearch_backtracking(&params, nx, x, &fx, g, d, step, xp, f, gb, &cplan, np, ps, &cancel, callback, callback_data);
         if (cancel) {
             return GGML_OPT_CANCEL;

ggml.h

@@ -326,7 +326,7 @@ extern "C" {
         GGML_TYPE_COUNT,
     };
 
-    enum ggml_backend {
+    enum ggml_backend_type {
         GGML_BACKEND_CPU = 0,
         GGML_BACKEND_GPU = 10,
         GGML_BACKEND_GPU_SPLIT = 20,
@@ -480,7 +480,9 @@ extern "C" {
     // n-dimensional tensor
     struct ggml_tensor {
         enum ggml_type         type;
-        enum ggml_backend backend;
+        enum ggml_backend_type backend;
+
+        struct ggml_backend_buffer * buffer;
 
         int     n_dims;
         int64_t ne[GGML_MAX_DIMS]; // number of elements
@@ -514,7 +516,7 @@ extern "C" {
 
         void * extra; // extra things e.g. for ggml-cuda.cu
 
-        char padding[4];
+        char padding[12];
     };
 
     static const size_t GGML_TENSOR_SIZE = sizeof(struct ggml_tensor);
@@ -702,6 +704,9 @@ extern "C" {
     GGML_API struct ggml_tensor * ggml_dup_tensor (struct ggml_context * ctx, const struct ggml_tensor * src);
     GGML_API struct ggml_tensor * ggml_view_tensor(struct ggml_context * ctx, struct ggml_tensor * src);
 
+    // Context tensor enumeration and lookup
+    GGML_API struct ggml_tensor * ggml_get_first_tensor(struct ggml_context * ctx);
+    GGML_API struct ggml_tensor * ggml_get_next_tensor (struct ggml_context * ctx, struct ggml_tensor * tensor);
     GGML_API struct ggml_tensor * ggml_get_tensor(struct ggml_context * ctx, const char * name);
 
     GGML_API struct ggml_tensor * ggml_set_zero(struct ggml_tensor * tensor);
@@ -1372,7 +1377,7 @@ extern "C" {
 
     // alibi position embedding
     // in-place, returns view(a)
-    struct ggml_tensor * ggml_alibi(
+    GGML_API struct ggml_tensor * ggml_alibi(
             struct ggml_context * ctx,
             struct ggml_tensor  * a,
             int                   n_past,
@@ -1381,7 +1386,7 @@ extern "C" {
 
     // clamp
     // in-place, returns view(a)
-    struct ggml_tensor * ggml_clamp(
+    GGML_API struct ggml_tensor * ggml_clamp(
             struct ggml_context * ctx,
             struct ggml_tensor  * a,
             float                 min,
@@ -2116,7 +2121,7 @@ extern "C" {
         enum ggml_type    vec_dot_type;
     } ggml_type_traits_t;
 
-    ggml_type_traits_t ggml_internal_get_type_traits(enum ggml_type type);
+    GGML_API ggml_type_traits_t ggml_internal_get_type_traits(enum ggml_type type);
 
 #ifdef  __cplusplus
 }

gguf-py/README.md

@@ -69,4 +69,3 @@ python -m twine upload dist/*
 ## TODO
 - [ ] Add tests
 - [ ] Include conversion scripts as command line entry points in this package.
-- Add CI workflow for releasing the package.

gguf-py/gguf/gguf.py

@@ -91,10 +91,12 @@ class MODEL_ARCH(IntEnum):
     PERSIMMON     : int = auto()
     REFACT        : int = auto()
     BERT          : int = auto()
+    BLOOM         : int = auto()
 
 
 class MODEL_TENSOR(IntEnum):
     TOKEN_EMBD      : int = auto()
+    TOKEN_EMBD_NORM : int = auto()
     TOKEN_TYPES     : int = auto()
     POS_EMBD        : int = auto()
     OUTPUT          : int = auto()
@@ -128,10 +130,12 @@ MODEL_ARCH_NAMES: dict[MODEL_ARCH, str] = {
     MODEL_ARCH.PERSIMMON:      "persimmon",
     MODEL_ARCH.REFACT:         "refact",
     MODEL_ARCH.BERT:           "bert",
+    MODEL_ARCH.BLOOM:          "bloom",
 }
 
 TENSOR_NAMES: dict[MODEL_TENSOR, str] = {
     MODEL_TENSOR.TOKEN_EMBD:      "token_embd",
+    MODEL_TENSOR.TOKEN_EMBD_NORM: "token_embd_norm",
     MODEL_TENSOR.TOKEN_TYPES:     "token_types",
     MODEL_TENSOR.POS_EMBD:        "position_embd",
     MODEL_TENSOR.OUTPUT_NORM:     "output_norm",
@@ -285,6 +289,18 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = {
         MODEL_TENSOR.FFN_DOWN,
         MODEL_TENSOR.FFN_UP,
     ],
+    MODEL_ARCH.BLOOM: [
+        MODEL_TENSOR.TOKEN_EMBD,
+        MODEL_TENSOR.TOKEN_EMBD_NORM,
+        MODEL_TENSOR.OUTPUT_NORM,
+        MODEL_TENSOR.OUTPUT,
+        MODEL_TENSOR.ATTN_NORM,
+        MODEL_TENSOR.ATTN_QKV,
+        MODEL_TENSOR.ATTN_OUT,
+        MODEL_TENSOR.FFN_NORM,
+        MODEL_TENSOR.FFN_DOWN,
+        MODEL_TENSOR.FFN_UP,
+    ],
     MODEL_ARCH.GPT2: [
         # TODO
     ],
@@ -314,6 +330,7 @@ class TensorNameMap:
             "gpt_neox.embed_in",                        # gptneox
             "transformer.wte",                          # gpt2 gpt-j mpt refact
             "transformer.word_embeddings",              # falcon
+            "word_embeddings",                          # bloom
             "model.embed_tokens",                       # llama-hf
             "tok_embeddings",                           # llama-pth
             "embeddings.word_embeddings",               # bert
@@ -325,6 +342,11 @@ class TensorNameMap:
             "embeddings.token_type_embeddings",  # bert
         ),
 
+        # Normalization of token embeddings
+        MODEL_TENSOR.TOKEN_EMBD_NORM: (
+            "word_embeddings_layernorm",  # bloom
+        ),
+
         # Position embeddings
         MODEL_TENSOR.POS_EMBD: (
             "transformer.wpe",                 # gpt2
@@ -335,7 +357,7 @@ class TensorNameMap:
         MODEL_TENSOR.OUTPUT: (
             "embed_out",                # gptneox
             "lm_head",                  # gpt2 mpt falcon llama-hf baichuan
-            "output",                   # llama-pth
+            "output",                   # llama-pth bloom
             "word_embeddings_for_head", # persimmon
         ),
 
@@ -347,7 +369,7 @@ class TensorNameMap:
             "norm",                                   # llama-pth
             "embeddings.LayerNorm",                   # bert
             "transformer.norm_f",                     # mpt
-            "ln_f",                                   # refact
+            "ln_f",                                   # refact bloom
             "language_model.encoder.final_layernorm", # persimmon
         ),
 
@@ -364,6 +386,7 @@ class TensorNameMap:
             "transformer.h.{bid}.ln_1",                            # gpt2 gpt-j refact
             "transformer.blocks.{bid}.norm_1",                     # mpt
             "transformer.h.{bid}.input_layernorm",                 # falcon7b
+            "h.{bid}.input_layernorm",                             # bloom
             "transformer.h.{bid}.ln_mlp",                          # falcon40b
             "model.layers.{bid}.input_layernorm",                  # llama-hf
             "layers.{bid}.attention_norm",                         # llama-pth
@@ -382,6 +405,7 @@ class TensorNameMap:
             "transformer.h.{bid}.attn.c_attn",                                    # gpt2
             "transformer.blocks.{bid}.attn.Wqkv",                                 # mpt
             "transformer.h.{bid}.self_attention.query_key_value",                 # falcon
+            "h.{bid}.self_attention.query_key_value",                             # bloom
             "language_model.encoder.layers.{bid}.self_attention.query_key_value", # persimmon
         ),
 
@@ -415,6 +439,7 @@ class TensorNameMap:
             "transformer.h.{bid}.attn.c_proj",                         # gpt2 refact
             "transformer.blocks.{bid}.attn.out_proj",                  # mpt
             "transformer.h.{bid}.self_attention.dense",                # falcon
+            "h.{bid}.self_attention.dense",                            # bloom
             "model.layers.{bid}.self_attn.o_proj",                     # llama-hf
             "layers.{bid}.attention.wo",                               # llama-pth
             "encoder.layer.{bid}.attention.output.dense",              # bert
@@ -432,6 +457,7 @@ class TensorNameMap:
         MODEL_TENSOR.FFN_NORM: (
             "gpt_neox.layers.{bid}.post_attention_layernorm",               # gptneox
             "transformer.h.{bid}.ln_2",                                     # gpt2 refact
+            "h.{bid}.post_attention_layernorm",                             # bloom
             "transformer.blocks.{bid}.norm_2",                              # mpt
             "model.layers.{bid}.post_attention_layernorm",                  # llama-hf
             "layers.{bid}.ffn_norm",                                        # llama-pth
@@ -445,6 +471,7 @@ class TensorNameMap:
             "transformer.h.{bid}.mlp.c_fc",                          # gpt2
             "transformer.blocks.{bid}.ffn.up_proj",                  # mpt
             "transformer.h.{bid}.mlp.dense_h_to_4h",                 # falcon
+            "h.{bid}.mlp.dense_h_to_4h",                             # bloom
             "model.layers.{bid}.mlp.up_proj",                        # llama-hf refact
             "layers.{bid}.feed_forward.w3",                          # llama-pth
             "encoder.layer.{bid}.intermediate.dense",                # bert
@@ -464,6 +491,7 @@ class TensorNameMap:
             "transformer.h.{bid}.mlp.c_proj",                        # gpt2 refact
             "transformer.blocks.{bid}.ffn.down_proj",                # mpt
             "transformer.h.{bid}.mlp.dense_4h_to_h",                 # falcon
+            "h.{bid}.mlp.dense_4h_to_h",                             # bloom
             "model.layers.{bid}.mlp.down_proj",                      # llama-hf
             "layers.{bid}.feed_forward.w2",                          # llama-pth
             "encoder.layer.{bid}.output.dense",                      # bert

gguf-py/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "gguf"
-version = "0.4.0"
+version = "0.4.4"
 description = "Write ML models in GGUF for GGML"
 authors = ["GGML <ggml@ggml.ai>"]
 packages = [

k_quants.h

@@ -29,7 +29,7 @@
 
 // 2-bit quantization
 // weight is represented as x = a * q + b
-// 16 blocks of 16 elemenets each
+// 16 blocks of 16 elements each
 // Effectively 2.5625 bits per weight
 typedef struct {
     uint8_t scales[QK_K/16]; // scales and mins, quantized with 4 bits
@@ -41,7 +41,7 @@ static_assert(sizeof(block_q2_K) == 2*sizeof(ggml_fp16_t) + QK_K/16 + QK_K/4, "w
 
 // 3-bit quantization
 // weight is represented as x = a * q
-// 16 blocks of 16 elemenets each
+// 16 blocks of 16 elements each
 // Effectively 3.4375 bits per weight
 #ifdef GGML_QKK_64
 typedef struct {
@@ -62,7 +62,7 @@ static_assert(sizeof(block_q3_K) == sizeof(ggml_fp16_t) + QK_K / 4 + QK_K / 8 +
 #endif
 
 // 4-bit quantization
-// 16 blocks of 32 elements each
+// 8 blocks of 32 elements each
 // weight is represented as x = a * q + b
 // Effectively 4.5 bits per weight
 #ifdef GGML_QKK_64
@@ -83,7 +83,7 @@ static_assert(sizeof(block_q4_K) == 2*sizeof(ggml_fp16_t) + K_SCALE_SIZE + QK_K/
 #endif
 
 // 5-bit quantization
-// 16 blocks of 32 elements each
+// 8 blocks of 32 elements each
 // weight is represented as x = a * q + b
 // Effectively 5.5 bits per weight
 #ifdef GGML_QKK_64
@@ -107,7 +107,7 @@ static_assert(sizeof(block_q5_K) == 2*sizeof(ggml_fp16_t) + K_SCALE_SIZE + QK_K/
 
 // 6-bit quantization
 // weight is represented as x = a * q
-// 16 blocks of 16 elemenets each
+// 16 blocks of 16 elements each
 // Effectively 6.5625 bits per weight
 typedef struct {
     uint8_t ql[QK_K/2];      // quants, lower 4 bits

prompts/mnemonics.txt

@@ -0,0 +1,93 @@
+For each kanji character, write a Markdown‐formatted mnemonic that uses its keyword and the keyword of all its components.
+
+Kanji: 欠 (lack of)
+Components: 𠂊 (hook claw), 人 (person)
+Mnemonic: This **person** is a pirate. He lost his hand to a crocodile many years ago. Nowadays, the ***lack of*** a hand does not bother him too much. In fact, the **hook claw** that replaces it is the mark of a true pirate, so he is quite proud of it!
+
+Kanji: 類 (kind (of something))
+Components: 米 (rice), 大 (large), 頁 (page)
+Mnemonic: The waiter at a Chinese restaurant hands you a **large** menu. Each **page** has all ***kinds*** of **rice** on offer!
+
+Kanji: 燃 (burn)
+Components: 火 (fire), 然 (sort of thing)
+Mnemonic: ***Burning*** things up with **fire** is just my **sort of thing**. (Spoken like a true pyromaniac.)
+
+Kanji: 頂 (top of)
+Components: 丁 (street), 頁 (page)
+Mnemonic: To be at the ***top of*** your game, you need both practical knowledge (**street** smarts) and theoretical knowledge (having read many **pages**).
+
+Kanji: 険 (risky and steep)
+Components: 阝 (small village), 㑒 (consensus)
+Mnemonic: Everyone agrees (there is **consensus**) that the path to the **small village** is ***risky and steep***.
+
+Kanji: 困 (distressed)
+Components: 囗 (closed box), 木 (tree)
+Mnemonic: You would feel ***distressed*** too if you were a **tree** trapped in a **closed box**! I have no place to grow!
+
+Kanji: 頭 (head)
+Components: 豆 (bean), 頁 (page)
+Mnemonic: What do you have in that ***head*** of yours? A **bean** for a brain? Go read more **pages** and become more knowledgeable about the world!
+
+Kanji: 確 (certain)
+Components: 石 (stone), 冖 (roof without a chimney), 隹 (old bird)
+Mnemonic: An **old bird** has made a nest on your **roof**. What do you do? You call Misaka from a <cite>A ***Certain*** Scientific Railgun</cite> to get rid of it, of course! But she doesn’t really want to vaporize the poor thing, so she just throws a **stone** to scare it away. (What was the point of calling her, then‽)
+
+Kanji: 魚 (fish)
+Components: 𠂊 (hook claw), 田 (rice field), 灬 (fire sparks)
+Mnemonic: Catch ***fish*** with a **hook**, collect rice from the **rice field**, cook them with **fire**… And my meal is ready!
+
+Kanji: 警 (to police (something))
+Components: 敬 (respect), 言 (say)
+Mnemonic: ***To police something*** is to make people **respect** what the law **says**.
+
+Kanji: 筆 (writing brush)
+Components: 竹 (bamboo), 聿 (brush)
+Mnemonic: A traditional ***writing brush*** is a **brush** made of **bamboo**.
+
+Kanji: 獄 (prison)
+Components: 犭 (animal), 言 (say), 犬 (dog)
+Mnemonic: In ***prison***, like in the **animal** kingdom, only the toughest survive. You have to watch what you **say**. It’s a **dog**‐eat‐dog world.
+
+Kanji: 新 (new)
+Components: 立 (standing up), 木 (tree), 斤 (axe)
+Mnemonic: In order for a ***new*** construction to be made, an empty lot is needed. If there are any **trees** **standing up**, they must be cut down with an **axe**.
+
+Kanji: 怪 (suspicious)
+Components: 忄 (weak heart), 圣 (sacred)
+Mnemonic: That painting of the **Sacred** **Heart** of Jesus looks ***suspicious***. I think it might be a forgery.
+
+Kanji: 温 (warm (to the touch))
+Components: 氵 (water drops), 日 (sun), 皿 (dish)
+Mnemonic: If you leave **water** on a **dish** in the **sun**, it will get ***warm***.
+
+Kanji: 階 (floor (of a building))
+Components: 阝 (small village), 皆 (all)
+Mnemonic: It might be a **small village**, but, despite that, **all** of its buildings have many ***floors***. It’s a village of skyscrapers!
+
+Kanji: 多 (many)
+Components: 夕 (evening (before sunset)), 夕 (evening (before sunset))
+Mnemonic: Two **evenings** in a day would be one too ***many***.
+
+Kanji: 別 (separate)
+Components: 口 (mouth), 万 (ten thousand), 刂 (knife)
+Mnemonic: Tom Six is at it again. For his next flick, he wants to stitch together **ten thousand** people, **mouth**‐to‐anus. One of the most graphic and disturbing scenes will feature one of the victims using a **knife** to ***separate*** perself.
+
+Kanji: 並 (line up)
+Components: 䒑 (antlers on a wall), 业 (runway)
+Mnemonic: In order to land a plane you have to ***line up*** properly with the **runway**. The things that look like **antlers** at the end of the runway are the control towers; you should follow their instructions.
+
+Kanji: 姿 (figure)
+Components: 次 (next), 女 (woman)
+Mnemonic: The **next** **woman** that I date will have a perfect **figure**. Because I’m done with 3D women—it will *literally* be an anime figure!
+
+Kanji: 実 (real)
+Components: 宀 (roof with a chimney), 𡗗 (three people)
+Mnemonic: Living under a **roof with a chimney** with **three people** (a wife and two children)—a happy family life—is not something I could have ever imagined. It does not feel ***real***.
+
+Kanji: 謝 (apologize)
+Components: 言 (say), 射 (shoot)
+Mnemonic: **Shot** first, ***apologize*** (**say** you are sorry) later.
+
+Kanji: 提 (propose)
+Components: 扌 (left hand), 是 (go with)
+Mnemonic:

scripts/sync-ggml.sh

@@ -2,6 +2,7 @@
 
 cp -rpv ../ggml/src/ggml.c                  ./ggml.c
 cp -rpv ../ggml/src/ggml-alloc.c            ./ggml-alloc.c
+cp -rpv ../ggml/src/ggml-backend.c          ./ggml-backend.c
 cp -rpv ../ggml/src/ggml-cuda.h             ./ggml-cuda.h
 cp -rpv ../ggml/src/ggml-cuda.cu            ./ggml-cuda.cu
 cp -rpv ../ggml/src/ggml-opencl.h           ./ggml-opencl.h
@@ -11,6 +12,7 @@ cp -rpv ../ggml/src/ggml-metal.m          ./ggml-metal.m
 cp -rpv ../ggml/src/ggml-metal.metal        ./ggml-metal.metal
 cp -rpv ../ggml/include/ggml/ggml.h         ./ggml.h
 cp -rpv ../ggml/include/ggml/ggml-alloc.h   ./ggml-alloc.h
+cp -rpv ../ggml/include/ggml/ggml-backend.h ./ggml-backend.h
 
 cp -rpv ../ggml/tests/test-opt.cpp    ./tests/test-opt.cpp
 cp -rpv ../ggml/tests/test-grad0.cpp  ./tests/test-grad0.cpp

tests/test-tokenizer-0-falcon.cpp

@@ -36,6 +36,8 @@ static const std::map<std::string, std::vector<llama_token>> & k_tests() {
         { "   Hello"              , {     258,  23090, }, },
         { "    Hello"             , {     466,  23090, }, },
         { "    Hello\n    Hello"  , {     466,  23090,    742,  23090, }, },
+        { "\n ="                  , {    1212,     40, }, },
+        { "' era"                 , {      18,   4932, }, },
     };
 
     return _k_tests;
@@ -155,7 +157,7 @@ int main(int argc, char **argv) {
 
         fprintf(stderr, "%s : text size: %zu\n", __func__, text.size());
 
-        const std::vector<llama_token> res = llama_tokenize(ctx, text, true);
+        const std::vector<llama_token> res = llama_tokenize(ctx, text, false);
 
         fprintf(stderr, "%s : tokens: %zu\n", __func__, res.size());
 
@@ -169,10 +171,8 @@ int main(int argc, char **argv) {
             }
 
             for (const auto & tok : res) {
-                ofs << tok << " ";
+                ofs << tok << " '" << llama_detokenize_bpe(ctx, std::vector<int>{tok}) << "'" << std::endl;
             }
-
-            ofs << "\n";
         }
 
         fprintf(stderr, "%s : tokens written to '%s'\n", __func__, (fname_text + ".tokcpp").c_str());

tests/test-tokenizer-0-falcon.py

@@ -41,6 +41,8 @@ tests = [
         "   Hello",
         "    Hello",
         "    Hello\n    Hello",
+        "\n =",
+        "' era",
     ]
 
 for text in tests:
@@ -69,15 +71,14 @@ fname_tok = args.fname_tok
 if fname_tok:
     print('tokenizing file: ', fname_tok)
     fname_out = fname_tok + '.tok'
-    with open(fname_tok, 'r') as f:
+    with open(fname_tok, 'r', encoding='utf-8') as f:
         lines = f.readlines()
         s = ''.join(lines)
         res = tokenizer.encode(s)
         # write to file
-        with open(fname_out, 'w') as f:
+        with open(fname_out, 'w', encoding='utf-8') as f:
             for x in res:
-                f.write(str(x) + ' ')
-            f.write('\n')
+                f.write(str(x) + ' \'' + tokenizer.decode(x) + '\'\n')
         print('len(res): ', len(res))
         print('len(lines): ', len(lines))
     print('results written to: ', fname_out)

tests/test-tokenizer-0-llama.cpp

@@ -174,10 +174,8 @@ int main(int argc, char **argv) {
             }
 
             for (const auto & tok : res) {
-                ofs << tok << " ";
+                ofs << tok << " '" << llama_detokenize_spm(ctx, std::vector<int>{tok}) << "'" << std::endl;
             }
-
-            ofs << "\n";
         }
 
         fprintf(stderr, "%s : tokens written to '%s'\n", __func__, (fname_text + ".tokcpp").c_str());

tests/test-tokenizer-0-llama.py

@@ -81,15 +81,14 @@ fname_tok = args.fname_tok
 if fname_tok:
     print('tokenizing file: ', fname_tok)
     fname_out = fname_tok + '.tok'
-    with open(fname_tok, 'r') as f:
+    with open(fname_tok, 'r', encoding='utf-8') as f:
         lines = f.readlines()
         s = ''.join(lines)
         res = tokenizer.encode(s, add_bos=True)
         # write to file
-        with open(fname_out, 'w') as f:
+        with open(fname_out, 'w', encoding='utf-8') as f:
             for x in res:
-                f.write(str(x) + ' ')
-            f.write('\n')
+                f.write(str(x) + ' \'' + tokenizer.decode(x) + '\'\n')
         print('len(res): ', len(res))
         print('len(lines): ', len(lines))
     print('results written to: ', fname_out)