Popular repositories Loading
-
drivefusion
drivefusion PublicDriveFusion is an open-source multimodal Vision–Language–Action model for autonomous driving that fuses visual perception, language understanding, and driving context (GPS and speed) to describe dr…
Python 2
-
data-preprocessing
data-preprocessing PublicData preprocessing pipeline for the DriveFusionQA. It converts multiple autonomous-driving QA datasets into unified LLaMA and LLaVA-style instruction formats, with modular dataset preprocessors, JS…
Python
-
Evaluate-Models
Evaluate-Models PublicEvaluation framework for the DriveFusion DriveFusionQA vision-language model, benchmarking Q&A performance on driving datasets using metrics like Lingo-Judge, BLEU, and BERTScore.
Python
-
drivefusion-train
drivefusion-train PublicTraining framework for the DriveFusion project that fine-tunes and train LLMs and multimodal vision-language models for driving tasks. Built on LLaMAFactory, it adds dataset processing, distributed…
Python
-
carla-data-collection
carla-data-collection PublicAutonomous-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.
Python
-
car-deployment
car-deployment PublicAutonomous 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…
Python
Repositories
- drivefusion Public
DriveFusion is an open-source multimodal Vision–Language–Action model for autonomous driving that fuses visual perception, language understanding, and driving context (GPS and speed) to describe driving scenes and predict future trajectories and target speeds.
DriveFusion/drivefusion’s past year of commit activity - carla-data-collection Public
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.
DriveFusion/carla-data-collection’s past year of commit activity - drivefusion-train Public
Training framework for the DriveFusion project that fine-tunes and train LLMs and multimodal vision-language models for driving tasks. Built on LLaMAFactory, it adds dataset processing, distributed training workflows, optimization for scalable autonomous-driving model development.
DriveFusion/drivefusion-train’s past year of commit activity - car-deployment Public
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.
DriveFusion/car-deployment’s past year of commit activity - data-preprocessing Public
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.
DriveFusion/data-preprocessing’s past year of commit activity - Evaluate-Models Public
Evaluation framework for the DriveFusion DriveFusionQA vision-language model, benchmarking Q&A performance on driving datasets using metrics like Lingo-Judge, BLEU, and BERTScore.
DriveFusion/Evaluate-Models’s past year of commit activity
Top languages
Loading…
Most used topics
Loading…