Issues

• Questions about RoboTwin reproduction and action chunk size

  thu-ml/Motus#46 · Filter by author Sonettovo opened on Aug 27, 2026

• Request: official RoboTwin inference/evaluation recipe for pi0.5_robotwin2

  thu-ml/Motus#45 · Filter by author XinrunXu opened on Aug 7, 2026

• Will you release IDM related codes and checkpoints?

  thu-ml/Motus#44 · Filter by author xiejingjacob opened on Jul 12, 2026

• How to generate videos with the stage1 pretrained Wan2.2-5B Checkpoint?

  thu-ml/Motus#42 · Filter by author Zhexin-Zhang opened on Jun 24, 2026

• RoboTwin demo_randomized eval: ~83–86% vs reported 87.02% — any missing setup details?

  thu-ml/Motus#41 · Filter by author leo038 opened on Jun 16, 2026

• LIBERO training and inference

  thu-ml/Motus#40 · Filter by author RuiWang724 opened on May 29, 2026
  1

• Request for Open-Sourcing LIBERO Training and Inference Code

  thu-ml/Motus#39 · Filter by author strive789 opened on May 18, 2026

• wechat discussion

  thu-ml/Motus#38 · Filter by author smilebetterworld opened on May 12, 2026
  1

• joint attention mask？

  thu-ml/Motus#37 · Filter by author smilebetterworld opened on May 10, 2026

• Evaluation of RoboCasa GR1

  thu-ml/Motus#36 · Filter by author qiulu66 opened on Apr 29, 2026

• RoboTwin dataset

  thu-ml/Motus#35 · Filter by author liujiahao2077 opened on Apr 28, 2026

• questions about latent action model

  thu-ml/Motus#34 · Filter by author hrz2000 opened on Apr 22, 2026
  1

• how to train motus with lerobot data?

  thu-ml/Motus#33 · Filter by author stephonjou opened on Mar 28, 2026

• deploy_policy.py: current_state_norm is computed from stat.json but never used in inference

  thu-ml/Motus#32 · Filter by author Vincia-Jun opened on Mar 19, 2026

• Iteration time

  thu-ml/Motus#30 · Filter by author Jun-CEN opened on Mar 4, 2026
  2

• real world infer script wrong

  thu-ml/Motus#27 · Filter by author quantumiracle opened on Feb 12, 2026
  1

• OOM on H100 80GB when training Motus

  thu-ml/Motus#26 · Filter by author WUyinwei-hah opened on Feb 12, 2026
  1

• 在RoboTwin环境进行测试，使用测试脚本auto_eval.sh。测试结果是 error occurs！请问是什么问题呢？

  thu-ml/Motus#25 · Filter by author LZCai opened on Feb 11, 2026
  2

• Is there any problem that the generated high cam image doesn't match wrist cam?

  thu-ml/Motus#24 · Filter by author wjydym opened on Feb 10, 2026

• Issue: Offline Inference failed caused by no robotwin.yml

  thu-ml/Motus#23 · Filter by author AlexHe99 opened on Feb 10, 2026
  1

• Question: Stage-3 Training with DeepSpeed ZeRO-3

  thu-ml/Motus#21 · Filter by author Qian0733 opened on Feb 2, 2026

• WeChat QR Code on Website Has Expired

  thu-ml/Motus#20 · Filter by author BitCalSaul opened on Jan 26, 2026

• 感谢你们的高质量工作，我对训练wan部分的代码很感兴趣，请问你们有计划开源这部分代码吗？ 希望你们不吝赐教！

  thu-ml/Motus#19 · Filter by author frost-ice0107 opened on Jan 23, 2026
  2

• Why are register tokens disabled during pre-training but enabled during fine-tuning?

  thu-ml/Motus#17 · Filter by author einnullnull opened on Jan 17, 2026
  4

• 大佬们 是否可以换成全模态模型 增加latent action token 然后增加一个action decoder解码latent action为实际动作的形式

  thu-ml/Motus#16 · Filter by author ryanwangran98 opened on Jan 15, 2026