diffusers

diffusers copied to clipboard

Convert Flax trained model to PyTorch

19

Based on #1161. I've mostly taken and adapted Flax->PT code from [transformers](https://github.com/huggingface/transformers). My test scenario is currently this: ```python from diffusers import FlaxStableDiffusionPipeline, StableDiffusionPipeline from diffusers import AutoencoderKL, UNet2DConditionModel import...

jorahn

Refactor cross attention and allow mechanism to tweak cross attention function

11

We are thinking about how to best support methods that tweak the cross attention computation, such as hyper networks (where linear layers that map k-> k' and v-> v' are...

patrickvonplaten

[Community Pipeline] Modified Cross-Attention - Structured Diffusion Guidance for Compositional T2I synthesis

7

## Intro Community Pipelines are introduced in `diffusers==0.4.0` with the idea of allowing the community to quickly add, integrate, and share their custom pipelines on top of `diffusers`. You can...

apolinario

[Dtype] Align dtype casting behavior with Transformers and Accelerate

4

Some recent changes in `transformers`: https://github.com/huggingface/transformers/pull/20602 and `accelerate`: https://github.com/huggingface/accelerate/pull/920 that force us to also align the behavior in `diffusers`. For more information also have a look at: https://discuss.pytorch.org/t/discrepancy-between-loading-models-with-meta-tensors-and-normal-load-from-state-dict/168295

patrickvonplaten

expose polynomial:power and cosine_with_restarts:num_cycles params

2

to get_scheduler func, * add it to the train_dreambooth.py

zetyquickly

Remove unethically trained diffusion models from model hub

10

@patrickvonplaten @patil-suraj @anton-l I wonder what your thoughts are on the arguments the artists are making on Artstation, Twitter and Instagram right now. The main point is that the artists...

lukovnikov

Transformers library missing even tho its already installed?

2


Spiderfffun

Converted Model from Diffusers Dreambooth -> to CKPT dont work in stable-diffusion-webui ?

1

Hey there first thex for the amazing work :) i manged to get Dreambooth runnding on my trust 2080ti and the training works i have results in the 400 and...

janoschsimon

Images are the wrong size (256px for stable diffusion 1.5) after using ckpt to diffusers script

3

### Describe the bug Hello, I was testing [the conversion script for ckpts to diffusers](https://raw.githubusercontent.com/huggingface/diffusers/039958eae55ff0700cfb42a7e72739575ab341f1/scripts/convert_original_stable_diffusion_to_diffusers.py), when I realized all the images that generated by the converted models are only 256x256...

JohnnyRacer

[Question] Shouldn't the inputs for memory_efficient_attention in the CrossAttention have [B, sequence length, Heads, embeding size] format?

1

According to the xformer's documentation (https://facebookresearch.github.io/xformers/components/ops.html#xformers.ops.memory_efficient_attention ): > Input tensors must be in format [B, M, H, K], where B is the batch size, M the sequence length, H the...

Warvito