Stars
Tile-Based Runtime for Ultra-Low-Latency LLM Inference
Hydra is a framework for elegantly configuring complex applications
Unofficial description of the CUDA assembly (SASS) instruction sets.
Reference implementation and examples of the CuTe Layout representation and algebra.
AI agent toolkit: unified LLM API, agent loop, TUI, coding agent CLI
FlyDSL is the Python front‑end of the project: a Flexible Layout Python DSL for expressing tiling, partitioning, data movement, and kernel structure at a high level.
H-Net: Hierarchical Network with Dynamic Chunking
CUDA Templates and Python DSLs for High-Performance Linear Algebra
Kimi K2 is the large language model series developed by Moonshot AI team
A high-performance library for compressed ND arrays and columnar tables, with compute and indexing engines
Development repository for the Triton language and compiler
FlashAttention (Metal Port)
Fast and memory-efficient exact attention
OSLO: Open Source framework for Large-scale model Optimization
Machine learning metrics for distributed, scalable PyTorch applications.
Flexible Python configuration system. The last one you will ever need.
PyTorch Lightning + Hydra. A very user-friendly template for ML experimentation. ⚡🔥⚡
KErnel OPerationS, on CPUs and GPUs, with autodiff and without memory overflows
Composable transformations of Python+NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more
Emacs client/library for the Language Server Protocol
lsp-mode ❤️ Microsoft's python language server
Research workflows made easy, locally and in the Cloud.
PyTorch Extension Library of Optimized Scatter Operations