[zero3] release tmp memory when consolidating fp16 weights #1220
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Outdated suggestions cannot be applied.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.
Currently
GatheredParameters
withmodifier_rank=None
doesn't release memory until the param is used next time some time in the future. As a result_zero3_consolidated_fp16_state_dict
was leaking memory. This PR uses the read/write mode which re-partitions the memory right away and releases the temporarily gathered params.@tjruwase