Skip to content

Conversation

@sayakpaul
Copy link
Member

What does this PR do?

Fixes #11225

@sayakpaul sayakpaul requested a review from hlky April 9, 2025 05:44
@HuggingFaceDocBuilderDev

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

Copy link
Contributor

@hlky hlky left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Left a few comments. Thanks @sayakpaul

Comment on lines +115 to +116
prompt_2=None,
prompt_3=None,
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Not a blocker for this PR but looks like prompt_2 and prompt_3 should be made Optional in the pipeline.

dataset = load_dataset(
args.train_data_dir,
cache_dir=args.cache_dir,
trust_remote_code=True,
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Nice, could be an argument but more convenient like this especially as the example dataset requires it. Can be replicated across training scripts.
For future reference we should look at the num_proc option which should help speed up processing.

@sayakpaul sayakpaul requested a review from hlky April 9, 2025 06:18
@sayakpaul sayakpaul merged commit fd02aad into main Apr 9, 2025
11 of 12 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

SD3 Controlnet Train Example, run out of memory on validation step

4 participants