Skip to content

Commit 04c5a48

Browse files
committed
chore: update readme, news about iros challenge, internmanip and interndata
1 parent ba05a80 commit 04c5a48

File tree

3 files changed

+22
-0
lines changed

3 files changed

+22
-0
lines changed

README.md

Lines changed: 22 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -22,6 +22,28 @@ Built on [NVIDIA Isaac Sim](https://developer.nvidia.com/isaac-sim), GenManip of
2222
* 🧱 A scalable asset pool with **10,000+ rigid** and **100+ articulated** objects with vision-language labels
2323
* 🧭 Evaluation of **spatial**, **appearance**, **commonsense**, and **long-horizon** reasoning capabilities
2424

25+
## News!!!
26+
27+
We are thrilled to announce the release of **10 hand-crafted post-training tasks** and the synthesis of **over 55K generalizable pick-and-place tasks across \~14K objects on the ALOHA platform** for the **IROS 2025 Challenge: Vision-Language Manipulation in Open Tabletop Environments**.
28+
29+
The pre-training dataset from the GenManip platform features **randomized backgrounds**, **randomized objects**, and **diverse instructions**, enabling strong generalization across a wide range of scenarios.
30+
31+
📌 **Register for the IROS Challenge**: [https://eval.ai/web/challenges/challenge-page/2626/overview](https://eval.ai/web/challenges/challenge-page/2626/overview)
32+
33+
![](readme_assets/iros-teaser.jpg)
34+
35+
**📂 Data Access**
36+
37+
* **Pre-training Data (Dual-arm Generalizable Pick-and-Place)**: [Hugging Face Link](https://huggingface.co/datasets/InternRobotics/IROS-2025-Challenge-Manip/tree/main)
38+
* **Post-training Data (Dual-arm Manipulation, 10 Tasks)**: [Hugging Face Link](https://huggingface.co/datasets/InternRobotics/IROS-2025-Challenge-Manip/tree/main)
39+
40+
<video src="readme_assets/scaling_data.mp4" controls width="600"></video>
41+
42+
In addition, the **GenManip Benchmark** will soon merge high-quality tasks into [InternManip](https://github.com/InternRobotics/InternManip). We also provide data generated by GenManip through [InternData-M1](https://huggingface.co/datasets/InternRobotics/InternData-M1) — a large-scale embodied robotics dataset containing \~250,000 simulation demonstrations with rich frame-level annotations, including 2D/3D bounding boxes, trajectories, grasp points, and semantic masks.
43+
44+
Due to its massive size, converting from the GenManip format to the LeRobot format has taken some time, and uploads are still in progress. Rest assured, all data has already been generated and will be available soon. Even more **scaling data for long-horizon tasks** is on the way! 🚀
45+
46+
2547
## ✨ Key Features
2648

2749
| Feature | Description |

readme_assets/iros-teaser.jpg

113 KB
Loading

readme_assets/scaling-data.mp4

12.1 MB
Binary file not shown.

0 commit comments

Comments
 (0)