Skip to content

Conversation

@stevhliu
Copy link
Member

@stevhliu stevhliu commented Oct 24, 2023

Provides a brief walkthrough of the training scripts so users know what they can modify for their use case and are familiar with what's happening where inside the script. It also updates the training overview to show supported training tasks and clarifies that SDPA is used by default during training (no need to enable anything).

I'm still experimenting with the format of the training script walkthrough, so I'm open to all suggestions/feedback!

Remaining scripts to update:

  • text-to-image
  • textual inversion
  • DreamBooth
  • LoRA
  • ControlNet
  • InstructPix2Pix
  • Custom Diffusion
  • T2I-Adapters
  • SDXL
  • Kandinsky
  • Wuerstchen

@HuggingFaceDocBuilderDev
Copy link

HuggingFaceDocBuilderDev commented Oct 25, 2023

The documentation is not available anymore as the PR was closed or merged.

@stevhliu
Copy link
Member Author

Would you love thoughts/feedback on the initial version of the training docs (unconditional image generation/text-to-image) refresh @sayakpaul @patrickvonplaten before I start working on the rest of them!

@sayakpaul sayakpaul requested a review from pcuenca October 27, 2023 04:05
Copy link
Member

@sayakpaul sayakpaul left a comment

Choose a reason for hiding this comment

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

Love it!

Copy link
Contributor

@patrickvonplaten patrickvonplaten left a comment

Choose a reason for hiding this comment

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

Very nice refactor!

Copy link
Member

@pcuenca pcuenca left a comment

Choose a reason for hiding this comment

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

Very nice!

@stevhliu
Copy link
Member Author

stevhliu commented Nov 1, 2023

Thanks for validating y'all, I'll start working on bringing similar changes to the rest of the training docs! 😄

@stevhliu
Copy link
Member Author

stevhliu commented Nov 9, 2023

Hey y'all! Some updates:

  • split the training section into "models" and "methods" because these are different things, and I think it'll help avoid confusion
  • adds SDXL, Kandinsky, and Wuerstchen to the "models" subsection

There are still a few minor things left to polish, but most of the content and refactoring is complete!

@stevhliu stevhliu marked this pull request as ready for review November 9, 2023 00:33
@stevhliu stevhliu merged commit bae14c8 into huggingface:main Nov 14, 2023
@stevhliu stevhliu deleted the training-docs branch November 14, 2023 18:30
AmericanPresidentJimmyCarter pushed a commit to AmericanPresidentJimmyCarter/diffusers that referenced this pull request Apr 26, 2024
* first draft

* try hfoption syntax

* fix hfoption id

* add text2image

* fix tag

* feedback

* feedbacks

* add textual inversion

* DreamBooth

* lora

* controlnet

* instructpix2pix

* custom diffusion

* t2i

* separate training methods and models

* sdxl

* kandinsky

* wuerstchen

* light edits
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.

5 participants