diffusers

diffusers copied to clipboard

Instruct-pix2pix pipeline: add ability to pass `cross_attention_kwargs` in call method

1

**Is your feature request related to a problem? Please describe.** I have a use-case where I need to use custom adapter together with instruct-pix2pix model, which requires passing custom `cross_attention_kwargs`...

AlexeyZhuravlev

[Design Discussion] allowing `from_pretrained()` to also load single file checkpoints

18

Since we were considering adding an option like `single_file_format` to `save_pretrained()` of `DiffusionPipeline`, it makes sense to have something similar in `from_pretrained()` to have better feature parity. We currently support...

sayakpaul

Allow image resolutions multiple of 8 instead of 64 in SVD pipeline

7

Hello! This is my first time making a PR to diffusers so I apologise if I have missed something! This PR implements the same behaviour from `unet_2d_condition.py` and `unet_3d_condition.py` to...

mlfarinha

[Wuerstchen] dreambooth lora

18

# What does this PR do? Lora dreambooth training scripts for Wuerstchen prior pipeline. Example: https://wandb.ai/krasul/dreambooth-lora/runs/7ia16ohd needs #6046

kashif

Lacking support for t2iadapter_style_sd14v1 model in T2IAdapter

5

**Is your feature request related to a problem? Please describe.** I was following [instructions for T2IAdapter](https://huggingface.co/docs/diffusers/api/pipelines/stable_diffusion/adapter) and trying to load the t2iadapter_style_sd14v1 adapter (Huggingface Hub [link](https://huggingface.co/TencentARC/T2I-Adapter/blob/main/models/t2iadapter_style_sd14v1.pth)). ``` from diffusers import...

Firenze11

HELP, can I do img2img on multi imgs at the same time?

5

In my case, I need to do img2img on multiple images, I saw the parameter "image" can be a list of PIL.image.iamge but when I used a list of images...

blx0102

The FID of the threetimestep_spacing （leading , linspace, trailing）using ddim timestep varies greatly with the number of total sampling steps(10st 100st)

3

### Describe the bug None of our three models using ddim timestep can reproduce the FID in the ddim paper, and the fid of the three models varies greatly with...

xinding64

[new model] DynamiCrafter: Animating Open-domain Images with Video Diffusion Priors

3

### Model/Pipeline/Scheduler description A very great model for image2video generation (also frame interpolation, storyboarding and looping videos). The architecture is based on the current text2video Unet paradigm, so it shouldn't...

kabachuha

[Pipeline] FRESCO

6

### Model/Pipeline/Scheduler description https://github.com/huggingface/diffusers/assets/72266394/a0f2a4d9-ed6e-403e-8a44-f53a9831467d ### Open source status - [X] The model implementation is available. - [X] The model weights are available (Only relevant if addition is not a scheduler)....

a-r-r-o-w

UNet2DModel - ValueError: cross_attention_dim must be specified for CrossAttnDownBlock2D

1

### Describe the bug When I load the model for my training, I just want to use the unconditional model. But I have this problem. Could anyone tell me how...

phuongtrau