README.md

metadata

title: FlowWAM
emoji: 🌊
colorFrom: pink
colorTo: indigo
sdk: gradio
sdk_version: 6.20.0
app_file: app.py
short_description: Predict future RGB video + optical flow from an image
python_version: '3.10'
startup_duration_timeout: 1h

FlowWAM

Interactive demo of FlowWAM: Optical Flow as a Unified Action Representation for World Action Models.

A single dual-stream video diffusion model (built on Wan2.2-TI2V-5B) jointly predicts a future RGB video and its optical-flow field from one input image and a short text instruction.

• Paper: https://huggingface.co/papers/2607.13017
• Code: https://github.com/YixiangChen515/FlowWAM
• Weights: https://huggingface.co/YixiangChen/FlowWAM