Skip to content

Releases: benmagnifico/DriveJev

DriveJev 1.0.1

Choose a tag to compare

@benmagnifico benmagnifico released this 03 Oct 09:52

DriveJev 1.0.1

This release makes DriveJev citable: it is archived on Zenodo with a DOI.

  • CITATION.cff for GitHub's "Cite this repository" button, and .zenodo.json with the archive metadata (authors, license, links to the model weights and JevPilot).
  • Citation: DriveJev: Real-Time Driving Behaviour Selection with a Vision-Language Decision Model, Jingguang Li, Kailang Ma, Zuyi Guo, Yebo Wu and Heye Huang.

The code and the model are the same as in DriveJev 1.0. The weights stay at benmagnifico/DriveJev-4B (tag v1.0).

See the README for installation and the evaluation docs for the full results.

DriveJev 1.0

Choose a tag to compare

@benmagnifico benmagnifico released this 02 Oct 13:57

An open-source, efficient System I decision model for autonomous driving.

DriveJev looks at three camera frames and a short description of the situation, scores the behaviours the car can carry out at that instant (cruise, stop at the line, hold, start, turn, yield, brake) and returns a probability for each in one forward pass. A local executor drives the chosen behaviour in the JevPilot town, city and interstate.

Highlights

  • Held-out closed-loop tests: 28 of 35 interaction episodes and 25 of 28 base episodes driven cleanly; 241 of 242 scripted hazards met without contact.
  • 111 ms per decision (9 Hz) on one RTX 5090.
  • Code, weights, JevPilot closed-loop harness, test suites and a live demo, all running locally on one GPU.

Contents

  • Model: [benmagnifico/DriveJev-4B](https://huggingface.co/benmagnifico/DriveJev-4B) on Hugging Face (backbone and decision head in one folder).
  • drivejev/: inference (prompt compiler, image patching, decision head, DrivingPolicy), serve/serve.py: HTTP model service.
  • simulator/, eval/: JevPilot closed-loop harness, hazard and multi-agent interaction scenarios, test suites, summarizer.
  • demo/: interactive JevPilot demo with the decision panel.

See the [README](https://github.com/benmagnifico/DriveJev#install) for installation and the [evaluation docs](https://github.com/benmagnifico/DriveJev/blob/main/docs/evaluation.md) for the full results.