diffusers

diffusers copied to clipboard

[Tests] help us speed up the fast pipeline tests

6

Currently, the pipeline tests (located under [`tests/pipelines`](https://github.com/huggingface/diffusers/tree/main/tests/pipelines/)) take a bit of time to run. We recently performed some optimizations to speed them up but, we can do better. This is...

sayakpaul

[ip-adapter] fix ip-adapter for StableDiffusionInstructPix2PixPipeline

1

fix https://github.com/huggingface/diffusers/issues/7799 this works now ```python import requests from io import BytesIO from PIL import Image import torch from diffusers import StableDiffusionInstructPix2PixPipeline pipe = StableDiffusionInstructPix2PixPipeline.from_pretrained( "timbrooks/instruct-pix2pix", torch_dtype=torch.float16, variant="fp16" ) pipe...

yiyixuxu

[scheduler] support custom `timesteps` and `sigmas`

16

# custom `timesteps` and `sigmas` for schedulers The general logic is: when a custom `timesteps` is passed, all the scheduler configuration that is used to generate timesteps (e.g. `timestep_spacing`) will...

yiyixuxu

I feel confused about this TODO issue. how to pass timesteps as tensors?

1

https://github.com/huggingface/diffusers/blob/235d34cf567e78bf958344d3132bb018a8580295/src/diffusers/models/unets/unet_2d_condition.py#L918

RosieApril

[docs] Community pipelines

1

This PR combines the "Community pipelines", "Load community pipelines and components", and "Contribute a community pipeline" docs together to make it easier to find everything about community pipelines on one...

stevhliu

Multi-controlnet formatting issue

1

### Describe the bug Hi. There is an inconsistency between `from_pretrained` and `save_pretrained` within the Multicontrolnet class. The from_pretrained function returns a directory structure like this: controlnet, controlnet_1, controlnet_2, whereas...

rebel-kblee

# What does this PR do? This PR introduces tests for the standalone Transformer2DModel class. It's quite heavily used and testing it in isolation is necessary to ensure we have...

sayakpaul

Inquiry About Using Non-Square Images for ControlNet Training

1

Subject: Inquiry About Using Non-Square Images for ControlNet Training Dear [Team/Developer], First of all, thank you for providing the training code for ControlNet. I have recently been utilizing this code...

starinskycc

Auth token is not passable to load_textual_inversion due to named argument conflicting with accepted kwargs

2

### Describe the bug You cannot provide an auth token value to the **kwargs of [``load_textual_inversion``](https://github.com/huggingface/diffusers/blob/cf6e0407e051467b480830d3ed97d2873b5019d3/src/diffusers/loaders/textual_inversion.py#L267) in ``src/diffusers/loaders/textual_inversion.py`` due to it being a named argument used for another purpose. It...

Teriks

Make the sampling loop of pipelines modular, with designated inputs that can be handled iteratively with functions and a default sampler step function

2

### Model/Pipeline/Scheduler description Related to #7761 . This is an effective replacement for the existing sampling loop function and the many, many kwargs that were made to allow the user...

AmericanPresidentJimmyCarter