Instructions to use lightx2v/Minimax-h3-Turbo with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use lightx2v/Minimax-h3-Turbo with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline from diffusers.utils import load_image, export_to_video # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("lightx2v/Minimax-h3-Turbo", dtype=torch.bfloat16, device_map="cuda") pipe.to("cuda") prompt = "A man with short gray hair plays a red electric guitar." image = load_image( "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/guitar-man.png" ) output = pipe(image=image, prompt=prompt).frames[0] export_to_video(output, "output.mp4") - Notebooks
- Google Colab
- Kaggle
Thanks and ComfyUI conversion
#3
by Kijai - opened
Hey, thank you once again for your continued distillation work, the community has been anxious for this one!
Here's quick ComfyUI compatible version: https://huggingface.co/Kijai/MiniMax-H3_comfy/
Things needed for conversion:
- fused QKV
- key renaming
- add alpha to match peft strength
We have observed some differences between the current Diffusers and ComfyUI inference results. In particular, the audio generated with Diffusers appears to be more natural. Please follow https://github.com/ModelTC/Minimax-H3-Turbo to reproduce the results.
The audio bug has been solved, please update Comfy to a nightly version.
Please see Kijai's comment in the original post