ayo/ipex-llm
1
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 1

[NPU] Further fix saving of generation config (#12657)

Browse source 
* Further fix saving of generation config

* Fix based on comments

* Small fix
This commit is contained in:
Yuwen Hu 2025-01-07 13:53:54 +08:00 committed by GitHub
parent 381d448ee2
commit ebdf19fa7e
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 5 additions and 8 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

13
python/llm/src/ipex_llm/transformers/npu_pipeline_model/convert_pipeline.py
View file

@ -473,10 +473,6 @@ def convert_llm_for_deploy(model: torch.nn.Module,
"n_splits_linear": n_splits_linear,
"n_splits_down_proj": n_splits_down_proj,
"lm_head_low_bit": lm_head_low_bit}
model.config.update(update_dict)
model.config.save_pretrained(save_directory)
if model.can_generate():
model.generation_config.save_pretrained(save_directory)
from .qwen import convert_qwen_layer, convert_fused_qwen_layer
from .qwen import convert_lm_head_and_embedding
@ -537,8 +533,6 @@ def convert_llm_for_deploy(model: torch.nn.Module,
"n_splits_linear": n_splits_linear,
"n_splits_down_proj": n_splits_down_proj,
"lm_head_low_bit": lm_head_low_bit}
model.config.update(update_dict)
model.config.save_pretrained(save_directory)
from .llama import convert_llama_layer, convert_fused_llama_layer
from .llama import convert_lm_head_and_embedding
@ -577,8 +571,6 @@ def convert_llm_for_deploy(model: torch.nn.Module,
"n_splits_linear": n_splits_linear,
"n_splits_down_proj": n_splits_down_proj,
"lm_head_low_bit": lm_head_low_bit}
model.config.update(update_dict)
model.config.save_pretrained(save_directory)
from .minicpm import convert_minicpm_layer, convert_fused_minicpm_layer
from .minicpm import convert_lm_head_and_embedding
@ -595,3 +587,8 @@ def convert_llm_for_deploy(model: torch.nn.Module,
save_directory, weight_dir,
convert_model=True,
max_prompt_len=max_prompt_len)
model.config.update(update_dict)
model.config.save_pretrained(save_directory)
if model.can_generate():
model.generation_config.save_pretrained(save_directory)