Instructions to use yrlyrl/MoPlan with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use yrlyrl/MoPlan 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("yrlyrl/MoPlan", 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
MoPlan weights
Weights of MoPlan, a motion-planning method for image-to-video generation built on Wan2.2-I2V-A14B.
Code, installation and usage: https://github.com/yangruoliu/moplan-code
| path | content |
|---|---|
stage1/backend_params.pt, stage1/decoder.pt |
Stage-1 motion planner (LoRA, input/output layers of the Wan2.2 high-noise expert, and the slot decoder) |
stage1/resolved_config.yaml, stage1/trainer_state.json, stage1/checkpoint_complete.json |
Stage-1 configuration and checkpoint records read by the loader |
stage1/prior_latent.pt |
prior latent used by the Stage-1 direct objective |
stage2/high/{lora_params.pt,motion_adapter.pt} |
Stage-2 LoRA and motion adapter for the high-noise expert |
stage2/low/{lora_params.pt,motion_adapter.pt} |
Stage-2 LoRA and motion adapter for the low-noise expert |
The base model Wan-AI/Wan2.2-I2V-A14B-Diffusers is not included; setup/download_weights.py in the code
repository downloads it at the pinned revision together with these files and checks every sha256.
- Downloads last month
- -
Model tree for yrlyrl/MoPlan
Base model
Wan-AI/Wan2.2-I2V-A14B-Diffusers