Instructions to use Comfy-Org/MiniMax-H3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusion Single File
How to use Comfy-Org/MiniMax-H3 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
RTX 4090 + 16GB VRAM CUDA unknown error using nvfp4 text encoder
#68
by Chongwei - opened
Anyone facing similar issue that CUDA unknown errors happen after encoders and before samplers using nvfp4 text_encoder on RTX 4090 + 16GB VRAM?
This happens to me quite often at first and was mitigated by disabling pinned memory + add vram_headroom. However, this seems inevitable again when I'm using ref video in addition to ref images in the ref2va workflows.
Only the RTX 50 series GPUs support NVFP4.
Only the RTX 50 series GPUs support NVFP4.
https://huggingface.co/Comfy-Org/MiniMax-H3/discussions/16#6a70a3aa7bfcdfbddb7f4062