Autonomous vehicle deployment system for DriveFusion combining ROS-based control with AI vision powered by the Qwen 2.5 vision-language model, enabling real-time driving, single-shot decisions, and video/image processing.
Data preprocessing pipeline for the DriveFusionQA. It converts multiple autonomous-driving QA datasets into unified LLaMA and LLaVA-style instruction formats, with modular dataset preprocessors, JSON creators, and validation tools to support training and evaluation of vision-language models.
Evaluation framework for the DriveFusion DriveFusionQA vision-language model, benchmarking Q&A performance on driving datasets using metrics like Lingo-Judge, BLEU, and BERTScore.
Autonomous-driving data pipeline for DriveFusion project built on the CARLA Simulator, generating cleaned multi-modal sensor data and VQA annotations for training vision-language action models.
Add a description, image, and links to the drivefusion topic page so that developers can more easily learn about it.
To associate your repository with the drivefusion topic, visit your repo's landing page and select "manage topics."