Bagheera

[df-if II] add additional input checks to ensure the input is divisible by 8

6

# What does this PR do? Fixes #7842 Adds logic to check_inputs for the IF SuperResolution pipeline so that the user receives a clear error when attempting to run the...

deepfloyd stage 2 crashes with tensor size mismatch when input image size is not divisible by 8

2

### Describe the bug DeepFloyd's upstream code supports 8px-aligned inputs for stage II, which I believe the Diffusers implementation is based upon. However, it seems that for certain sizes, there...

[FEATURE REQUEST] text-to-image finetuning with aspect bucketed data

2

### Feature Request I would like for autotrain to be able to train on aspect-bucketed datasets. ### Motivation When training text-to-image models, it is beneficial to have data bucketed into...

training example for instruct pix2pix doesn't zero out embeds

8

### Describe the bug When running inference on SDXL, the config specifies to zero out the embedding when the prompt is empty. ### Reproduction ```py # Get null conditioning def...

fix for #7365, prevent pipelines from overriding provided prompt embeds

4

# What does this PR do? Fixes #7365 ## Before submitting - [x] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's...

Hello, I'm a fan of the model, I have support for fine-tuning it in my project, but the gated nature of the weights on HF Hub is an obstacle to...

wandb token exposed to wandb users

13

hello, there exists a parameter to pass a wandb token into the training utility. unfortunately, this shows on the "Overview" tab of the training session, as a part of the...

SDXL Support

5

#88 and #84 and #56 asked for other support so there's little hope from me for this, but I humbly request SDXL support as well.

I was thinking about Sayak's suggestion lately that the training examples are too long, and went through looking for redundant/unnecessary code sections that can be reduced or eliminated for readability....

[SDXL] Update ControlNet training example to zero out embeds for empty captions

2

# What does this PR do? Fixes #7920 ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if...