[textual inversion] add gradient checkpointing and small fixes. #1848
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
patil-suraj
requested review from
patrickvonplaten,
pcuenca and
williamberman
patil-suraj
commented
Dec 28, 2022
| vae.to(accelerator.device, dtype=weight_dtype) | ||
|
|
||
| # Keep vae and unet in eval model as we don't train these | ||
| vae.eval() |
Contributor
Author
There was a problem hiding this comment.
vae is already in eval mode by default, when loaded using from_pretrained.
|
The documentation is not available anymore as the PR was closed or merged. |
Contributor
Author
|
Did a trial run, and verified that keeping the |
Contributor
|
Thanks! |
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Outdated suggestions cannot be applied.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.
This PR is a continuation of #687 by @Ttl . This PR
--gradient_checkpointingto enable gradient checkpointing. The gradient checkpointing is enabled for thetext_encoderandunet. For this, theunetis kept in train mode and as the dropout is 0 it should not affect the results.freeze_paramsand instead userequires_grad_to disable grads.accelerator.wait_for_everyone()after every epoch, it's only needed before saving the pipeline.