diffusers

diffusers copied to clipboard

update the logic of `is_sequential_cpu_offload`

1

follow up on this https://github.com/huggingface/accelerate/issues/2701 when the sequential CPU offloading method is enabled for the pipeline, accelerate will try to install an `AlignDevicesHook` to each model component; if the model...

yiyixuxu

Implement: Align Your Steps: Optimizing Sampling Schedules in Diffusion Models

12

Align Your Steps: Optimizing Sampling Schedules in Diffusion Models is a general and principled approach to optimizing the sampling schedules of DMs for high-quality outputs. This work is presented by...

joe-aivatarz

[Core] introduce videoprocessor.

5

# What does this PR do? Introduces a `VideoProcessor` akin to `VaeImageProcessor` to encapsulate the logic of dealing with videos. ## TODOs - [x] Add tests - [x] Docs

sayakpaul

[mps] training / inference dtype issues

32

when training on Diffusers without attention slicing, we see: ``` /AppleInternal/Library/BuildRoots/ce725a5f-c761-11ee-a4ec-b6ef2fd8d87b/Library/Caches/com.apple.xbs/Sources/MetalPerformanceShaders/MPSCore/Types/MPSNDArray.mm:788: failed assertion `[MPSNDArray initWithDevice:descriptor:] Error: total bytes of NDArray > 2**32' ``` but with attention slicing, this error disappears....

bghira

Combined loss term for VQ-VAE (`diffusers.VQModel`)

2

For training the VQ-VAE component of a latent diffusion model a la `CompVis/ldm-celebahq-256` (which uses `diffusers.VQModel`), is there a combined loss term for each of the losses as described by...

asy51

Official callbacks

38

# What does this PR do? Initial draft to support for official callbacks. This is the most basic implementation I could think of without the need of modifying the pipelines....

asomoza

confusion when train dreambooth via script train_dreambooth_lora_sdxl_advanced

19

I used the train_dreambooth_lora_sdxl_advanced script downloaded from diffusers' official exmaples to train my own images, however, both the validation images generated and the images generated by trained model are of...

joey0922

Clear `ruff` Configuration

3

### Update 1. `C` in the select list includes only complex-structure (C901), so it is not meaningful to select `C` and ignore complex-structure (C901) at the same time, thus removing...

tolgacangoz

[Pipeline] Fix error of SVD pipeline when num_videos_per_prompt > 1

3

# What does this PR do? Fixes #[Issue](https://github.com/huggingface/diffusers/issues/7773#issuecomment-2078411517) it should repeat latents before concatenate do_classifier_free_guidance latents to align with the CFG process after noise_pred. ## Before submitting - [ ]...

wuyushuwys

How to input gradio settings in Python

3

Hi. I use **realisticStockPhoto_v20** on Fooocus with **sdxl_film_photography_style** lora and I really like the results. Fooocus and other gradio implementations come with settings inputs that I want to utilize in...

levoz92