Instructions to use circlestone-labs/Anima with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusion Single File
How to use circlestone-labs/Anima with Diffusion Single File:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
Had a question about Hires.fix
Is there any settings for hires that would work with the anima base model? If so what are the suggested settings, upscalers, denoise etc
You can use it directly just like SDXL, whether in ComfyUI or WebUI. The only thing to note is that if you are not using tiled upscale, it is recommended not to exceed a 1.5x upscale ratio, otherwise there is a chance of burning the image.
You can use it directly just like SDXL, whether in ComfyUI or WebUI. The only thing to note is that if you are not using tiled upscale, it is recommended not to exceed a 1.5x upscale ratio, otherwise there is a chance of burning the image.
I have a question, whenever I use Anima I have to set the denoise to 0, if I don't I get a blurry pixelated mess or a bunch of white pixels. I tried setting Hires.fix to 1.5, I usually use 2, but I still get the same result. I've tried with both Remacri and 4x Ultrasharp. But by setting it to 0, it's no different than just upscaling it, so it doesn't actually get any more details.
I leave tiled on unless I'm trying to get fine details, I've noticed that tiled can sometimes change the way the final output looks, in the end Hires.fix it still looks like hot garbage. I'm going to try some different upscalers, but I would really appreciate any tips/tricks you could share

