DeepSpeed
Fix: only add parameter with grads to parameter group
#7869
Merged

Commits
  • only add parameter with grads to parameter group
    delock committed 166 days ago
  • Remove CUDA dependency
    delock committed 166 days ago
  • Use distributed test
    delock committed 165 days ago
  • Merge branch 'master' into gma/fix_muon_partial_training
    delock committed 165 days ago
  • Merge branch 'master' into gma/fix_muon_partial_training
    delock committed 163 days ago
  • Merge branch 'master' into gma/fix_muon_partial_training
    tjruwase committed 159 days ago
Loading