github
/
minimind-v
mirror of https://github.com/jingyaogong/minimind-v.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

[fix] reduce aux_loss_alpha

pull/107/head
jingyaogong 2025-12-05 23:07:55 +08:00
parent 13c3a2c093
commit b7cb98c136
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
model/model_minimind.py
View File

@ -34,7 +34,7 @@ class MiniMindConfig(PretrainedConfig):
n_routed_experts: int = 4, n_routed_experts: int = 4,
n_shared_experts: int = 1, n_shared_experts: int = 1,
scoring_func: str = 'softmax', scoring_func: str = 'softmax',
aux_loss_alpha: float = 0.1, aux_loss_alpha: float = 0.01,
seq_aux: bool = True, seq_aux: bool = True,
norm_topk_prob: bool = True, norm_topk_prob: bool = True,
**kwargs **kwargs