Stars
gpt-oss-120b and gpt-oss-20b are two open-weight language models by OpenAI
Ray is an AI compute engine. Ray consists of a core distributed runtime and a set of AI Libraries for accelerating ML workloads.
[CVPR 2025 Best Paper Award] VGGT: Visual Geometry Grounded Transformer
CoTracker is a model for tracking any point (pixel) on a video.
Schedule-Free Optimization in PyTorch
Biological foundation modeling from molecular to genome scale
Formalizing and benchmarking open problems in single-cell genomics
Fast and Accurate ML in 3 Lines of Code
Awesome LLM compression research papers and tools.
PyTorch Extension Library of Optimized Autograd Sparse Matrix Operations
Accessible large language models via k-bit quantization for PyTorch.
DSPy: The framework for programming—not prompting—language models
A playbook for systematically maximizing the performance of deep learning models.
AutoBound automatically computes upper and lower bounds on functions.
Hackable and optimized Transformers building blocks, supporting a composable construction.
Lightning fast C++/CUDA neural network framework
A Python-level JIT compiler designed to make unmodified PyTorch programs faster.