[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]

![TheHumanoidHub Avatar](https://lunarcrush.com/gi/w:24/cr:twitter::1682127524963426304.png) The Humanoid Hub [@TheHumanoidHub](/creator/twitter/TheHumanoidHub) on x 59.8K followers
Created: 2025-07-16 19:43:04 UTC

A humanoid robot policy trained solely on synthetic data generated by a world model.

Research Scientist Joel Jang presents NVIDIA's DreamGen pipeline:

⦿ Post-train the world model Cosmos-Predict2 with a small set of real teleoperation demos.
⦿ Prompt the world model to generate synthetic video data with verbs and scenarios not used in the world model’s post-training.
⦿ Auto-label synthetic video data with action sequences.
⦿ Train robot policies using only synthetic data.

That's it. Deploy zero-shot to a real humanoid robot.

![](https://pbs.twimg.com/amplify_video_thumb/1945565804549812224/img/beXL3ZXLno4wtPvL.jpg)

XXXXXX engagements

![Engagements Line Chart](https://lunarcrush.com/gi/w:600/p:tweet::1945569893048619201/c:line.svg)

**Related Topics**
[generated](/topic/generated)
[robot](/topic/robot)
[humanoid](/topic/humanoid)

[Post Link](https://x.com/TheHumanoidHub/status/1945569893048619201)