DeepSpeed
Fix zero stage2 cpu_offload when some model trainable parameters skipped in training
#861

Merged

Commits

Merge pull request #1 from microsoft/master

ghosthamlet committed 5 years ago
Fix zero stage2 cpu_offload when some model trainable parameters skipped in training, as in https://github.com/microsoft/DeepSpeed/issues/707

ghosthamlet committed 5 years ago
Trim space

ghosthamlet committed 5 years ago
Trim space

ghosthamlet committed 5 years ago
Merge branch 'master' into ghosthamlet-stage2-bug

tjruwase committed 5 years ago