vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

llama : add phi-2 + fix NeoX rope + ggml_mul_mat_set_prec (#4490)

Browse source 
* phi2 implementation

* fix breaking change

* phi-2 : various fixes

* phi-2 : use layer norm eps

* py : whitespaces

* llama : fix meta KV override bug

* convert : phi don't add BOS token

* convert : revert "added_tokens_decoder" change

* phi-2 : scale Q instead of KQ for better precision

* ggml : fix NeoX rope to rotate just first n_dims

* cuda : less diff in the rope_neox kernel

* ggml : add ggml_mul_mat_set_prec

ggml-ci

* Update ggml-cuda.cu

Co-authored-by: slaren <slarengh@gmail.com>

* Update ggml-cuda.cu

Co-authored-by: slaren <slarengh@gmail.com>

* cuda : ggml_cuda_op_mul_mat_cublas support F32 precision

* cuda : remove oboslete comment

---------

Co-authored-by: Ebey Abraham <ebeyabraham@microsoft.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: slaren <slarengh@gmail.com>
This commit is contained in:
Ebey Abraham 2023-12-18 17:27:47 +00:00 • committed by GitHub
parent 3c04bf6da8
commit b9e74f9bca
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
9 changed files with 463 additions and 76 deletions
Download patch file Download diff file Expand all files Collapse all files

13
gguf-py/gguf/constants.py
View file

@ -95,6 +95,7 @@ class MODEL_ARCH(IntEnum):
BLOOM = auto()
STABLELM = auto()
QWEN = auto()
PHI2 = auto()
class MODEL_TENSOR(IntEnum):
@ -140,6 +141,7 @@ MODEL_ARCH_NAMES: dict[MODEL_ARCH, str] = {
MODEL_ARCH.BLOOM: "bloom",
MODEL_ARCH.STABLELM: "stablelm",
MODEL_ARCH.QWEN: "qwen",
MODEL_ARCH.PHI2: "phi2",
}
TENSOR_NAMES: dict[MODEL_TENSOR, str] = {
@ -350,6 +352,17 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = {
MODEL_ARCH.GPT2: [
# TODO
],
MODEL_ARCH.PHI2: [
MODEL_TENSOR.TOKEN_EMBD,
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,
]
# TODO
}