Track 1: Driving with Language
-
Updated
Aug 23, 2025 - Python
Track 1: Driving with Language
Fine-tuning Vision-Language Models for autonomous driving VQA: custom i.i.d. DriveLM-nuScenes split, ~132k pseudo-labels generated by Qwen3 from nuScenes sensor priors (3D bounding boxes, LiDAR depth, tracking trajectories), and InternVL2-2B LoRA fine-tuning reaching a DriveLM score of 0.589. Master's thesis, CTU Prague 2026.
Add a description, image, and links to the drivelm topic page so that developers can more easily learn about it.
To associate your repository with the drivelm topic, visit your repo's landing page and select "manage topics."