CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-14 11:07:10 +00:00
Code Issues Packages Projects Releases Wiki Activity

support convert starcoder weights to gguf

Browse Source
This commit is contained in:
Meng Zhang
2023-09-15 11:24:24 +08:00
parent 0c5d4d87b0
commit eb7f0eba3e
2 changed files with 262 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
gguf-py/gguf/gguf.py
View File

@@ -174,6 +174,16 @@ MODEL_TENSOR_NAMES: dict[MODEL_ARCH, dict[MODEL_TENSOR, str]] = {
MODEL_TENSOR.FFN_UP: "blk.{bid}.ffn_up",
},
MODEL_ARCH.STARCODER: {
MODEL_TENSOR.TOKEN_EMBD: "token_embd",
MODEL_TENSOR.POS_EMBD: "position_embd",
MODEL_TENSOR.OUTPUT_NORM: "output_norm",
MODEL_TENSOR.OUTPUT: "output",
MODEL_TENSOR.ATTN_NORM: "blk.{bid}.attn_norm",
MODEL_TENSOR.ATTN_QKV: "blk.{bid}.attn_qkv",
MODEL_TENSOR.ATTN_OUT: "blk.{bid}.attn_output",
MODEL_TENSOR.FFN_NORM: "blk.{bid}.ffn_norm",
MODEL_TENSOR.FFN_UP: "blk.{bid}.ffn_up",
MODEL_TENSOR.FFN_DOWN: "blk.{bid}.ffn_up",
},
MODEL_ARCH.GPT2: {
# TODO