diffusers-rs: A Diffusers API in Rust/Torch

A rusty robot holding a fire torch, generated by stable diffusion using Rust and libtorch.

The diffusers crate is a Rust equivalent to Huggingface's amazing diffusers Python library. It is based on the tch crate. The implementation is complete enough so as to be able to run Stable Diffusion v1.5 and v2.1.

In order to run the models, one has to get the weights from the huggingface model repo v2.1 or v1.5, move them in the data/ directory and then can run the following command.

cargo run --example stable-diffusion --features clap -- --prompt "A rusty robot holding a fire torch."

The final image is named sd_final.png by default. The only supported scheduler is the Denoising Diffusion Implicit Model scheduler (DDIM). The original paper and some code can be found in the associated repo.

FAQ

Memory Issues

This requires a GPU with more than 8GB of memory, as a fallback the CPU version can be used but is slower.

cargo run --example stable-diffusion --features clap -- --prompt "A very rusty robot holding a fire torch." --cpu all

For a GPU with 8GB, one can use the fp16 weights for the UNet and put only the UNet on the GPU.

PYTORCH_CUDA_ALLOC_CONF=garbage_collection_threshold:0.6,max_split_size_mb:128 RUST_BACKTRACE=1 CARGO_TARGET_DIR=target2 cargo run \
    --example stable-diffusion --features clap -- --cpu vae --cpu clip \
    --unet-weights data/unet-fp16.safetensors

Examples

A bunch of rusty robots holding some torches!

Image to Image Pipeline

The stable diffusion model can also be used to generate an image based on another image. The following command runs this image to image pipeline:

cargo run --example stable-diffusion-img2img --features clap -- --input-image media/in_img2img.jpg

The default prompt is "A fantasy landscape, trending on artstation.", but can be changed via the -prompt flag.

Inpainting Pipeline

Inpainting can be used to modify an existing image based on a prompt and modifying the part of the initial image specified by a mask. This requires different unet weights unet-inpaint.safetensors that could also be retrieved from this repo and should also be placed in the data/ directory.

The following command runs this image to image pipeline:

wget https://raw.githubusercontent.com/CompVis/latent-diffusion/main/data/inpainting_examples/overture-creations-5sI6fQgYIuo.png -O sd_input.png
wget https://raw.githubusercontent.com/CompVis/latent-diffusion/main/data/inpainting_examples/overture-creations-5sI6fQgYIuo_mask.png -O sd_mask.png
cargo run --example stable-diffusion-inpaint --features clap --input-image sd_input.png --mask-image sd_mask.png

The default prompt is "Face of a yellow cat, high resolution, sitting on a park bench.", but can be changed via the -prompt flag.

Converting the Original Weight Files

The weights can be retrieved as .safetensors or .ot files from huggingface v2.1. or v1.5. It is recommended to use the .safetensors versions.

It is also possible to download the weights for the original stable diffusion model and convert them to .safetensors files by following the instructions below or use the download scripts from the scripts/ directiory. These instructions are for version 1.5 but can easily be adapted for version 2.1 using this model repo instead.

First get the vocabulary file and uncompress it.

mkdir -p data && cd data
wget https://github.com/openai/CLIP/raw/main/clip/bpe_simple_vocab_16e6.txt.gz
gunzip bpe_simple_vocab_16e6.txt.gz

Clip Encoding Weights

For the clip encoding weights, start by downloading the weight file.

wget https://huggingface.co/openai/clip-vit-large-patch14/resolve/main/pytorch_model.bin

Then using Python, load the weights and save them in a .safetensors file.

import numpy as np
import torch
from safetensors.torch import save_file

model = torch.load("./pytorch_model.bin")
tensors = {k: v.clone().detach() for k, v in model.items() if 'text_model' in k}
save_file(tensors, 'pytorch_model.safetensors')

VAE and Unet Weights

The weight files can be downloaded from huggingface's hub but it first requires you to log in (and to accept the terms of use the first time). Then you can download the VAE weights and Unet weights.

After downloading the files, use Python to convert them to npz files.

import torch
from safetensors.torch import save_file
  
model = torch.load('./vae.bin')
save_file(dict(model), './vae.safetensors')
  
model = torch.load('./unet.bin')
save_file(dict(model), './unet.safetensors')