423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-10-14 15:52:49 +08:00
Code Issues Packages Projects Releases Wiki Activity

Update patcher.py

Browse Source 
Former-commit-id: 4c31a21f2106adcdad100119bad83ecaef0be3f3
This commit is contained in:
hoshi-hiyouga 2024-05-15 15:37:07 +08:00 committed by GitHub
parent 8e518d6c62
commit c4f5e49d0d
1 changed files with 2 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
src/llmtuner/model/patcher.py
View File

@ -17,7 +17,7 @@ from .utils.moe import add_z3_leaf_module, configure_moe
from .utils.quantization import configure_quantization
from .utils.rope import configure_rope
from .utils.valuehead import prepare_valuehead_model
from .utils.visual import autocast_projector_dtype, configure_hidden_size, configure_visual
from .utils.visual import autocast_projector_dtype, configure_visual_model
if TYPE_CHECKING:
@ -54,8 +54,7 @@ def patch_config(
configure_longlora(config, model_args, is_trainable)
configure_quantization(config, tokenizer, model_args, init_kwargs)
configure_moe(config, model_args, is_trainable)
configure_hidden_size(config)
configure_visual(config, model_args)
configure_visual_model(config)
if model_args.use_cache and not is_trainable:
setattr(config, "use_cache", True)