Hugging Face's logo

InternRobotics
/
InternVLA-N1-wo-dagger

Robotics
Transformers
Safetensors
qwen2_5_vl
text-generation
vision-language-model
video-language-model
navigation
text-generation-inference
Model card Files
xet
 Community
1

Update model card: Add main paper link, update project page and citation

#1
by nielsr HF Staff - opened
base: refs/heads/main
←
from: refs/pr/1
Discussion Files changed
+21
-26
Files changed (1) hide show
  1. README.md +21 -26
README.md CHANGED
@@ -1,11 +1,11 @@
1
  ---
2
- pipeline_tag: robotics
3
  library_name: transformers
4
  license: cc-by-nc-sa-4.0
 
5
  tags:
6
- - vision-language-model
7
- - video-language-model
8
- - navigation
9
  ---
10
 
11
  <div id="top" align="center">
@@ -13,18 +13,13 @@ tags:
13
 
14
  </div>
15
 
16
-
17
-
18
-
19
  # InternVLA-N1: An Open Dual-System Navigation Foundation Model with Learned Latent Plans
20
 
21
- [![Code](https://img.shields.io/badge/GitHub-Code-181717?logo=github)](https://github.com/InternRobotics/InternNav)
22
-
23
- Project page: https://internrobotics.github.io/internvla-n1.github.io/
24
-
25
- Technical report: https://internrobotics.github.io/internvla-n1.github.io/static/pdfs/InternVLA_N1.pdf
26
-
27
- Data: https://huggingface.co/datasets/InternRobotics/InternData-N1
28
 
29
 
30
 
@@ -35,13 +30,13 @@ Data: https://huggingface.co/datasets/InternRobotics/InternData-N1
35
  * We recommend using this official release for research and deployment, as it contains the most stable and up-to-date improvements.
36
 
37
  ### Key Difference: Preview vs Official
38
- | Feature | InternVLA-N1-Preview | InternVLA-N1 (official) |
39
- | ------------- | ----------------------------------------- | ------------------------------------------------------------------------ |
40
- | System Design | Dual-System (synchronous) | Dual-System (asynchronous) |
41
- | Training | System 1 trained only at System 2 inferrence step | System 1 trained on denser step (~25 cm), using latest System 2 hidden state |
42
- | Inference | System 1, 2 infered at same frequency (~2 hz) | System 1, 2 infered asynchronously, allowing dynamic obstacle avoidance |
43
- | Performance | Solid baseline in simulation & benchmarks | Improved smoothness, efficiency, and real-world zero-shot generalization |
44
- | Status | Historical preview | Stable official release (recommended)
45
 
46
  ## Highlights
47
 
@@ -66,11 +61,11 @@ Please refer to [InternNav](https://github.com/InternRobotics/InternNav) for its
66
  If you find our work helpful, please consider starring this repo 🌟 and cite:
67
 
68
  ```bibtex
69
- @misc{internvla-n1,
70
- title = {{InternVLA-N1: An} Open Dual-System Navigation Foundation Model with Learned Latent Plans},
71
- author = {InternVLA-N1 Team},
72
- year = {2025},
73
- booktitle={arXiv},
74
  }
75
  ```
76
 
 
1
  ---
 
2
  library_name: transformers
3
  license: cc-by-nc-sa-4.0
4
+ pipeline_tag: robotics
5
  tags:
6
+ - vision-language-model
7
+ - video-language-model
8
+ - navigation
9
  ---
10
 
11
  <div id="top" align="center">
 
13
 
14
  </div>
15
 
 
 
 
16
  # InternVLA-N1: An Open Dual-System Navigation Foundation Model with Learned Latent Plans
17
 
18
+ **Paper:** [Ground Slow, Move Fast: A Dual-System Foundation Model for Generalizable Vision-and-Language Navigation](https://huggingface.co/papers/2512.08186)
19
+ **Code:** [![Code](https://img.shields.io/badge/GitHub-Code-181717?logo=github)](https://github.com/InternRobotics/InternNav)
20
+ **Project Page:** https://internrobotics.github.io/internvla-n1-dualvln.github.io/
21
+ **Technical Report (InternVLA-N1):** https://internrobotics.github.io/internvla-n1.github.io/static/pdfs/InternVLA_N1.pdf
22
+ **Data:** https://huggingface.co/datasets/InternRobotics/InternData-N1
 
 
23
 
24
 
25
 
 
30
  * We recommend using this official release for research and deployment, as it contains the most stable and up-to-date improvements.
31
 
32
  ### Key Difference: Preview vs Official
33
+ | Feature | InternVLA-N1-Preview | InternVLA-N1 (official) |
34
+ |---|---|---|
35
+ | System Design | Dual-System (synchronous) | Dual-System (asynchronous) |
36
+ | Training | System 1 trained only at System 2 inferrence step | System 1 trained on denser step (~25 cm), using latest System 2 hidden state |
37
+ | Inference | System 1, 2 infered at same frequency (~2 hz) | System 1, 2 infered asynchronously, allowing dynamic obstacle avoidance |
38
+ | Performance | Solid baseline in simulation & benchmarks | Improved smoothness, efficiency, and real-world zero-shot generalization |
39
+ | Status | Historical preview | Stable official release (recommended)
40
 
41
  ## Highlights
42
 
 
61
  If you find our work helpful, please consider starring this repo 🌟 and cite:
62
 
63
  ```bibtex
64
+ @article{wei2025ground,
65
+ title={Ground Slow, Move Fast: A Dual-System Foundation Model for Generalizable Vision-and-Language Navigation},
66
+ author={Wei, Meng and Wan, Chenyang and Peng, Jiaqi and Yu, Xiqian and Yang, Yuqiang and Feng, Delin and Cai, Wenzhe and Zhu, Chenming and Wang, Tai and Pang, Jiangmiao and Liu, Xihui},
67
+ journal={arXiv preprint arXiv:2512.08186},
68
+ year={2025}
69
  }
70
  ```
71