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
Wide aspect ratio quality
Are there any plans to improve the quality of generations in wide aspects ratios (i.e. 16:9 and similar)? This problem is significant in SDXL-based models, while newer ones (FLUX, Chorma, Z-image, Qwen) don't have such a problem.
Anima has a lot of potential, but it would be disappointing if it would degrade on wide pictures, especially with the level of small details it can generate (quite good for wallpapers, actually).
I'm having good luck with it, mostly. I think it just needs more training. A good amount of the time I'll get good ones like this, and then on others with similar prompts it'll be an incoherent mess. I think there's just gaps in the training right now. The same thing happened with Chroma before it was "finished".
Yes, it's not broken, but it's...simpler. Less details and simpler color palette compared to vertical and square pictures. I hope it will be fixed with later training.
