BLUE: Toward Better Language Use in Efficient Vision-Language-Action Models for Autonomous Driving
Paper • 2606.08684 • Published • 2
Official Checkpoint and Evaluation Logs of BLUE
TLDR: Driving VLAs often generate language reasoning that is useless or even harmful to driving. BLUE addresses this by generating language only when it clearly helps, thereby improving driving performance while reducing inference latency. BLUE uses a 0.11M-parameter gate to decide at each frame whether to predict driving actions with or without intermediate language generation.
Links
- Paper: https://arxiv.org/abs/2606.08684
- GitHub Repository: https://github.com/George-Ling3/BLUE
- Project Page: https://blue-website.github.io
This Release Includes
- Evaluation logs
- Bench2Drive evaluation logs
- Longest6 v2 evaluation logs
- BLUE gate weights
- Checkpoint:
blue_simlingo_gate.pt - Metadata:
blue_simlingo_gate.json - Checksum file:
SHA256SUMS
- Checkpoint:
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support