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

Add support for starcoder2

Browse source
This commit is contained in:
Sourab Mangrulkar 2024-02-29 17:31:25 +05:30
parent d5ab29757e
commit ab4eab3a82
4 changed files with 211 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

17
gguf-py/gguf/constants.py
View file

@ -112,6 +112,7 @@ class MODEL_ARCH(IntEnum):
INTERNLM2 = auto()
MINICPM = auto()
GEMMA = auto()
STARCODER2 = auto()
class MODEL_TENSOR(IntEnum):
@ -169,6 +170,7 @@ MODEL_ARCH_NAMES: dict[MODEL_ARCH, str] = {
MODEL_ARCH.INTERNLM2: "internlm2",
MODEL_ARCH.MINICPM: "minicpm",
MODEL_ARCH.GEMMA: "gemma",
MODEL_ARCH.STARCODER2: "starcoder2",
}
TENSOR_NAMES: dict[MODEL_TENSOR, str] = {
@ -273,6 +275,21 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = {
MODEL_TENSOR.FFN_DOWN,
MODEL_TENSOR.FFN_UP,
],
MODEL_ARCH.STARCODER2: [
MODEL_TENSOR.TOKEN_EMBD,
MODEL_TENSOR.OUTPUT_NORM,
MODEL_TENSOR.OUTPUT,
MODEL_TENSOR.ROPE_FREQS,
MODEL_TENSOR.ATTN_NORM,
MODEL_TENSOR.ATTN_Q,
MODEL_TENSOR.ATTN_K,
MODEL_TENSOR.ATTN_V,
MODEL_TENSOR.ATTN_OUT,
MODEL_TENSOR.ATTN_ROT_EMBD,
MODEL_TENSOR.FFN_NORM,
MODEL_TENSOR.FFN_DOWN,
MODEL_TENSOR.FFN_UP,
],
MODEL_ARCH.BERT: [
MODEL_TENSOR.TOKEN_EMBD,
MODEL_TENSOR.TOKEN_EMBD_NORM,