Skip to content

Conversation

@kew6688
Copy link
Collaborator

@kew6688 kew6688 commented Nov 3, 2025

✅ Summary of Changes:
After refactor to new vln_multi_evaluator, the old two evaluators (vln_multi and vln_pe) are merged together. However, the vis_output and save_to_json commit is missed to merge back.

In this pr,

  • The visualization feature is added back.
  • The internnav.model.basemodel.rdp.utils is a bad design and redesign to internnav.utils.geometry_utils, as it contains most isaac sim geometry convert utilities and used across modules (model, agent, evaluator visualize)
  • save_to_json flag is for online evaluation in eval.ai, keep for future open leaderboard

Fix #141

@kew6688 kew6688 requested a review from yuqiang-yang November 4, 2025 02:50
@kew6688 kew6688 requested a review from Tai-Wang November 4, 2025 03:48
@Tai-Wang Tai-Wang merged commit e0d265c into InternRobotics:main Nov 4, 2025
1 check passed
heyi301 pushed a commit to heyi301/InternNav that referenced this pull request Dec 16, 2025
…eometry utils (InternRobotics#145)

* add back visualize feature; refactor rdp utils to geometry utils

* remove json file
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

[Question]: 进行eval评测时,系统没有记录video

3 participants