frozenleaves/vllm
1
0
Fork 0
mirror of https://github.com/vllm-project/vllm.git synced 2025-10-20 23:03:52 +08:00
Code Issues Packages Projects Releases Wiki Activity

[BugFix] gemma loading after quantization or LoRA. (#3553)

Browse Source
This commit is contained in:
Taemin Lee
2024-03-22 05:16:57 +09:00
committed by GitHub
parent c188ecb080
commit b7050ca7df
1 changed files with 4 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
vllm/model_executor/models/gemma.py
View File

@ -340,6 +340,10 @@ class GemmaForCausalLM(nn.Module):
weight_loader(param, loaded_weight, shard_id) weight_loader(param, loaded_weight, shard_id)
break break
else: else:
# lm_head is not used in vllm as it is tied with embed_token.
# To prevent errors, skip loading lm_head.weight.
if "lm_head.weight" in name:
continue
# Skip loading extra bias for GPTQ models. # Skip loading extra bias for GPTQ models.
if name.endswith(".bias") and name not in params_dict: if name.endswith(".bias") and name not in params_dict:
continue continue