BF16 optimizer sync parameters with module when load_module_only=True #56
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR lets a BF16 optimizer synchronize parameters with module in when
load_module_only=True
.With the previous conditional statement
if self.optimizer is not None and self.fp16_enabled():
, an engine with a BF16 optimizer will not executeself.optimizer.refresh_fp32_params()
and thus the optimizer will maintain random parameters, and after the first update step, model parameters will be updated to these random parameters.After changing the conditional statement to
if self.optimizer is not None:
, a BFf16 optimizer can correctly sync its parameters with the loaded module parameters.Following are logs of finetuning a pre-trained model with the argument
finetune=True