fix(video_jepa): apply AMP from config in training loop - #19
Open
Delata22 wants to merge 1 commit into
Open
Conversation
The video_jepa training config declares use_amp=true with dtype=float16, but main.py never read cfg.training, so training ran in fp32 regardless of the config. This is inconsistent with image_jepa and ac_video_jepa, which both apply AMP correctly. This applies the same autocast/GradScaler pattern used in those examples. Since detection_head uses nn.BCELoss (unsafe under autocast), det_loss is computed outside the autocast block.
meta-cla
Bot
added
the
CLA Signed
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The video_jepa training config declares use_amp=true with dtype=float16, but main.py never read cfg.training, so training ran in fp32 regardless of the config. This is inconsistent with image_jepa and ac_video_jepa, which both apply AMP correctly.
This applies the same autocast/GradScaler pattern used in those examples. Since detection_head uses nn.BCELoss (unsafe under autocast), det_loss is computed outside the autocast block.