ros/LLaMA-Factory
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

generalized packing & fix #6343

Browse Source 
Former-commit-id: 3b1e4194616cacd5c24f08b328e31a008bddcf29
This commit is contained in:
hiyouga
2024-12-17 10:26:19 +00:00
parent 6522467ddb
commit d670d62a66
4 changed files with 10 additions and 57 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
src/llamafactory/model/patcher.py
View File

@@ -96,7 +96,7 @@ def patch_config(
configure_quantization(config, tokenizer, model_args, init_kwargs)
configure_moe(config, model_args, is_trainable)
configure_visual_model(config)
configure_packing(config, model_args, is_trainable)
configure_packing(model_args, is_trainable)
if model_args.use_cache and not is_trainable:
setattr(config, "use_cache", True)