Instructions to use AX1Y2JP/MiniMax-H3-W4A8-ConvRot with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusion Single File
How to use AX1Y2JP/MiniMax-H3-W4A8-ConvRot 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
w4a8 without dynamic vram support
#7
by inykokso - opened
so i tested this on 5060ti 16gb , works ok if it fit on gpu (model+latent) , if it doesnt then oom...
max i was able to do 864x480 10s , 13s if i close all other app
int 8 is way slower as it not fit to 16gb , but creating longer or/and higher res is posible
solved , updated cuda from cu128 to cu130 , and disabled 2nd gpu