ViperEkura/AstrAI
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
670 Commits 1 Branch 14 Tags
a317a4756b32532ebd7e9e49c9de812c020ee8b9
Commit Graph
2 Commits
Author SHA1 Message Date
ViperEkura a317a4756b refactor: stateless MoE routing with grouped dispatch 
- replace per-expert mask scan with sort+bincount grouped dispatch
- carry router stats in forward output instead of module state
- keep MoE diagnostics working under DDP/FSDP wrappers
- remove unused _load_balancing_loss helper
 2026-08-05 18:42:12 +08:00
ViperEkura 9b7e6c205f feat: add moe auxloss and metrics 2026-08-05 18:12:28 +08:00