Add float16 support to load op#10541
Merged
kexinzhao merged 4 commits intoPaddlePaddle:developfrom May 10, 2018
Merged
Conversation
sidgoyal78
previously approved these changes
May 9, 2018
Contributor
sidgoyal78
left a comment
sidgoyal78
left a comment
There was a problem hiding this comment.
Thanks for the PR, this looks good.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Now save op support saving float32 weights as float16 on disk.
This can be easily extended to load op so that it can load float32 weights from disk to float16 weights in the scope.
@sidgoyal78 can work on add float16 option to load_combined_op after this pr is approved.