Skip to content

Conversation

@lawrence-cj
Copy link
Contributor

What does this PR do?

Fix the 4K inference bug in SanaPAGPipeline

Cc: @yiyixuxu , @a-r-r-o-w

@lawrence-cj lawrence-cj changed the title Sana 4 k [Sana-4K] Jan 12, 2025
image = self.vae.decode(latents / self.vae.config.scaling_factor, return_dict=False)[0]
try:
image = self.vae.decode(latents / self.vae.config.scaling_factor, return_dict=False)[0]
except torch.cuda.OutOfMemoryError as e:
Copy link
Collaborator

Choose a reason for hiding this comment

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

ok here, if you feel strongly about this, we prefer to throw out a warning with recommendations and user can explicitly enable the optimization

@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.

@yiyixuxu yiyixuxu merged commit 3d70777 into huggingface:main Jan 14, 2025
12 checks passed
@lawrence-cj
Copy link
Contributor Author

Thanks! @yiyixuxu

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants