Refactored DETR implementation on optax API.
#1062
Open
Conversation
This file contains 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
* Updated `detr_config.py`. * `get_detr_optimizer` function now returns optax transformation.
|
Hi @MasterSkepticista, I am experimenting with your modified detr, I have encountered many environment conflicts and tried to resolve them, but there are still some difficult bugs, can you provide your environment as a reference. Thank you very much. |
|
@durianer-D I created a new repo with a minimal implementation. It is also significantly faster to train than this PR. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Many functions and calls within existing
detrimplementation are deprecated in Flax as well as OTT.This PR includes two major changes (all changes only under
projects/baselines/detr):flax.optimtooptax.sinkhornsolver now uses updatedott-jaxcalls.I have tested both configs for train/eval/checkpointing.