This repository was archived by the owner on Apr 17, 2023. It is now read-only.
[NNCF] Enable fp16 optimization #95
Open
Conversation
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
alexsu52
suggested changes
Oct 13, 2022
alexsu52
left a comment
alexsu52
left a comment
There was a problem hiding this comment.
Should NNCF version be updated for this changes?
Author
Yes, it should |
daniil-lyakhov
force-pushed
the
dl/enable_fp16
branch
from
49cfd46 to
04ec6be
Compare
Author
|
Purpose of this PR is to check NNCF is working in FP16 precision. This is basic POC which could possibly broke other scenarios. |
Let's update it here |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
2 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Outdated suggestions cannot be applied.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.
Purpose of this PR is to check NNCF is working in FP16 precision. This is basic POC which could possibly broke other scenarios.
PR introduces
Fp16OptimizerHookusage for gradient scaling during training loop. With this code and updated version NNCF (with FP16 kernel improvements openvinotoolkit/nncf#1279) It's possible to optimize HRnet large-mod3 in FP16 precision