This repository has been archived by the owner on Nov 27, 2024. It is now read-only.
v0.19.0 - API Redesign
API Redesign (Breaking Changes!!)
The Stable Diffusion API has been made a lot lower level, dependency injection has been pretty much removed and a lot of helper methods added to make it similar in use to diffusers (well sort of).
Running SD can now be as simple as:
var pipeline = StableDiffusionPipeline.CreatePipeline("D:\\Repositories\\stable-diffusion-v1-5");
var promptOptions = new PromptOptions { Prompt = "Photo of a cute dog." };
var result = await pipeline.RunAsync(promptOptions);
var image = result.ToImage();
await image.SaveAsPngAsync("D:\\Results\\Image.png");
What's Changed
- VideoToVideo Example Runners And StableDiffusionService GenerateInputVideoFrames Bug Fix by @TheyCallMeHex in #103
- chore: Examples add schedulerOptions.Width and Height ; Fix Batchinde… by @kimi0230 in #106
- FeatureExtractor Project by @saddam213 in #104
- Major: API Refactor by @saddam213 in #107
- VideoToVideo example by @saddam213 in #108
New Contributors
Nuget Packages Updated
https://www.nuget.org/packages/OnnxStack.Core/0.19.0
https://www.nuget.org/packages/OnnxStack.ImageUpscaler/0.19.0
https://www.nuget.org/packages/OnnxStack.StableDiffusion/0.19.0
https://www.nuget.org/packages/OnnxStack.FeatureExtractor/0.19.0
Full Changelog: v0.17.0...v0.19.0
Example UI Builds
Name | ONNX EP | Device | Requirements | Status |
---|---|---|---|---|
OnnxStack_UI.zip | DirectML | CPU, AMD GPU, Nvidia GPU | At least Windows10 | Stable |
OnnxStack_UI_Cuda.zip | CUDA | CPU, Nvidia GPU | CUDA 11 and cuDNN toolkit |
Experimental |
OnnxStack_UI_TensorRT.zip | TensorRT | CPU, Nvidia GPU | CUDA 11 , cuDNN and TensorRT libraries |
Experimental |