Instructions to use MATLOWAI/MiniMax-H3-Motion-Adapter with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use MATLOWAI/MiniMax-H3-Motion-Adapter 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("MiniMaxAI/MiniMax-H3", dtype=torch.bfloat16, device_map="cuda") pipe.load_lora_weights("MATLOWAI/MiniMax-H3-Motion-Adapter") prompt = "A man with short gray hair plays a red electric guitar." input_image = load_image("https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/guitar-man.png") output = pipe(image=input_image, prompt=prompt).frames[0] export_to_video(output, "output.mp4") - Inference
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- Draw Things
Commit History
22 days ago
-
Add recovered fight source and reproducible four-tile canvas examples
Verified signatureAdd source and verified winner MP4s, beginner-friendly canvas workflows, historical API graphs, a runner, verification records and download bundle. Preserve the original benchmark exhibits and add reproduction links to the current model card.42cbd13
-
model card: the temporal-expansion ladder section (measured curve, the eye-vs-meter honesty, usage, provenance)
Verified signature0bfe4d6 -
temporal-expansion 2x3 comparison tile + poster
Verified signature02e532d -
temporal-expansion ladder: warm100 + every 25th step of the 375-step dense-target run (real GOPRO_Large intermediate frames, truthful world clock)
Verified signature5fdabd8
-
card: lead with two worked exhibits, name the panels, say it needs the de-rope pass
Verified signature724293a -
card media: two comparison quads and their posters
Verified signaturebc18cd4
-
card: note the future direction (an adapter that assigns compute on its own)
Verified signature2e2d6c7 -
card: state sample sizes (Ref2VA two clips, anime one clip), add the 0.75 fewer-objects trade, add the rooftop smoke test with its caveat
Verified signature5e2d275 -
card: say plainly that the LoRA adds no compute; the nodes stretch the clock, the LoRA changes behaviour on it
Verified signatured9cc2db -
add the MiniMax H3 Community License alongside MIT; credit MiniMax and mark the adapter relation
Verified signature35b095e -
model card: colour muting is about strong colour and particle effects, seen on both checkpoints
Verified signatureed87466 -
model card: link the demo page anchor and the node pack
Verified signaturef3e9df0 -
Motion adapter pilot r16: weights, model card, ComfyUI graph
Verified signature7d8c909 -
initial commit
Verified signaturede9fbde