facebookresearch/eb_jepa

Train world models for prediction and planning

View on index · View in 3D Map

Related repos

• Environment for training agents to use apps
• VGGT is a neural network that predicts 3D scene attributes like camera parameters, depth, and point maps from images
• Open vision-language models that compete with closed-source alternatives using fully open data
• Sequence modeling toolkit with a cleaner, more modular API
• Library for benchmarking multi-agent reinforcement learning algorithms
• toolkit for time series analysis, forecasting, and anomaly detection
• Balances data samples for inference
• PyTorch models for high-res vision tasks using self-supervised learning
• Benchmark for agents to reproduce scientific findings
• Agent for research across model providers and MCP servers