Llama

Meta’s SAM image segmentation project supports prompt inputs such as points and boxes, and can also automatically generate candidate masks for an entire image. It is suitable for interactive annotation and image processing workflows.

A library for efficient similarity search and clustering of dense vectors.

A PyTorch-based audio generation research library containing components such as MusicGen, AudioGen, and EnCodec. It supports text-conditioned generation, some melody-conditioned tasks, and model training workflows.

The repository provides code for running inference with the Meta Segment Anything Model 2 (SAM 2), links for downloading the trained model checkpoints, and example notebooks that show how to use the model.

Visual Geometry Grounded Transformer predicts camera parameters, depth maps, point maps, and point tracks from scene images, making it suitable for preliminary geometric estimation in 3D vision research and reconstruction workflows.

Meta’s self-supervised vision models provide pretrained backbones and task heads for feature extraction, visual retrieval and downstream adaptation.

A configuration framework for Python applications that manages experiments through configuration composition, command-line overrides, and multirun execution. It is suitable for organizing data, model, and training parameters into reusable configuration structures.