license: other
library_name: pytorch
pipeline_tag: robotics
tags:
- long-wam
- robotwin2.0
- idm
Long-WAM RoboTwin2.0-IDM
Long-WAM policy weights for RoboTwin2.0. IDM predicts actions after generating future-video latents.
Download
hf download Efficient-Large-Model/Long-WAM-RoboTwin2.0-IDM --local-dir ./weights/Long-WAM-RoboTwin2.0-IDM
Inference
Load model.pt with the matching Long-WAM runtime, using config.yaml and
the supplied dataset_stats.json for preprocessing and normalization.
P48 context uses 48 past control steps plus the current observation, encoded into four clean latent frames.
Generate two future-video latents using 4 steps (imagination sigma 0.9), then predict a 32-step action chunk using 10 action steps. Replan every 24 executed actions. Reset history at the start of each episode.
The matching inference runtime and base-model VAE/text assets are required separately. This release contains weights and inference settings only; optimizer states and training artifacts are not included.