423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-12-15 03:10:35 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix qwen2 moe (#6684)

Browse Source 
Former-commit-id: 7bf09abf1c
This commit is contained in:
hoshi-hiyouga
2025-01-17 13:46:09 +08:00
committed by GitHub
parent 555f17c1ee
commit 788accb601
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
src/llamafactory/model/model_utils/moe.py
View File

@@ -61,7 +61,7 @@ def add_z3_leaf_module(model: "PreTrainedModel") -> None:
_set_z3_leaf_modules(model, [MixtralSparseMoeBlock]) _set_z3_leaf_modules(model, [MixtralSparseMoeBlock])
if model_type == "qwen2moe": if model_type == "qwen2_moe":
from transformers.models.qwen2_moe.modeling_qwen2_moe import Qwen2MoeSparseMoeBlock from transformers.models.qwen2_moe.modeling_qwen2_moe import Qwen2MoeSparseMoeBlock
_set_z3_leaf_modules(model, [Qwen2MoeSparseMoeBlock]) _set_z3_leaf_modules(model, [Qwen2MoeSparseMoeBlock])