Normalize moe loss #50313
moe aux scaling
762b040f
up
c663d27a
Merge main into moe-aux-loss-scaling
3c054d31
Rename the MoE aux loss test and simplify its docstring
47afc16a
Test the MoE aux loss with and without num_items_in_batch, drop the p…
544f3c27
vasqu
commented
on 2026-11-04
loss moe normalization
438dd074
SunMarc
changed the title Fix Moe aux loss scaling Normalize moe loss 4 days ago
Merge main into moe-aux-loss-scaling
a0d3e2e6
Fix
7826cc76
Merge main into moe-aux-loss-scaling
ede58608
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub