Instructions to use FrancisRing/Prism with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use FrancisRing/Prism 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("FrancisRing/Prism", 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
This model will be a strong competitor to the Minimax H3!
This model will be a strong competitor to the Minimax H3! I hope they soon release pruned/turbo versions and a custom workflow for ComfyUI! Just by releasing a pruned/turbo version for ComfyUI, they’ve already garnered a lot of attention and prominence within the open-source community!
The 'Prism' team needs to create a turbo/pruned 4/8-step INT8 version, formalize a workflow integration with ComfyUI, and ramp up their marketing—much like Minimax H3 does—by posting official model teasers and trailers, comparison videos, and updates across all social media platforms (X/Twitter, Reddit, Facebook, Discord, etc.). If they do this, they will gain significant traction in the community and go head-to-head with the current leader, Minimax H3. This model has incredible potential, but it needs to take these steps!
One other thing I forgot to mention... Create an official page or account named 'Prism' on X/Twitter, just like LTX, WAN, Minimax, and Flux did. That account would be the place where we can all follow the latest news, daily or weekly updates, and reposts of videos created by Prism users. All of this fosters an ecosystem of excitement, longevity, and loyalty between us users and the model/team.