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
About the quality of ref2v
#49
by julia527 - opened
When will a higher-quality distilled ref2v model be released? We sincerely need it
I use the FL2VID model for REF2VID for two weeks already. As long as you dont need video input the FL2VID Model works as good as the REF2VID but gives higher output quality. No matter if you use 2 or 8 Input images, FL2vid model does not care and perfectly fine.
Yes, same, i almost only use the FL2V model for all my R2V videos.
The FL2V model give a lot better quality.
For the same video/audio quality i have with FL2V 4 steps. To get the same video/audio quality with the R2V model i need at least 8 or 10 seps.
The quality difference is enormous.
The R2V model need to be retrained, because there are some issues.