worv-ai/D2E

Scales vision-action pretraining on desktop data for embodied AI transfer

View on index · View in 3D Map

Related repos

• fast vision-language model for on-device AI
• Trains agents to act from videos
• Open-source computer vision tool that generates AI code from prompts
• Open source AI assistant for desktop automation
• vision-language model using reinforcement learning for visual understanding tasks
• End-to-end autonomous driving with vision-language-action models
• Web interface for creating AI images and videos
• DeepSeek-VL2, a vision-language AI modelfor tasks like visual question answeringand OCR
• Interactive deep learning textbook with executable Jupyter notebooks
• Builds real-time vision AI agents