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

llama : MiniCPM support tied embeddings (#7664)

Browse source 
* support lm_head

* remove the code block

---------

Co-authored-by: zhangkaihuo <zhangkaihuo@modelbest.cn>
This commit is contained in:
zhangkaihuo 2024-06-03 15:49:30 +08:00 • committed by GitHub
parent 549279d804
commit 6f28a333c1
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
2 changed files with 6 additions and 7 deletions
Download patch file Download diff file Expand all files Collapse all files

1
gguf-py/gguf/constants.py
View file

@ -645,6 +645,7 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = {
],
MODEL_ARCH.MINICPM: [
MODEL_TENSOR.TOKEN_EMBD,
MODEL_TENSOR.OUTPUT,
MODEL_TENSOR.OUTPUT_NORM,
MODEL_TENSOR.ROPE_FREQS,
MODEL_TENSOR.ATTN_NORM,