Public repository of the Micro QuickJS Javascript Engine

⚡️SwanLab - an open-source, modern-design AI training tracking and visualization tool. Supports Cloud / Self-hosted use. Integrated with PyTorch / Transformers / verl / LLaMA Factory / ms-swift / U…

A Survey of Reinforcement Learning for Large Reasoning Models

Allow torch tensor memory to be released and resumed later

Mirage Persistent Kernel: Compiling LLMs into a MegaKernel

Out-of-tree LLVM passes in Rust

A GUI client for Windows, Linux and macOS, support Xray and sing-box and others

slime is an LLM post-training framework for RL Scaling.

Multi-Level Triton Runner supporting Python, IR, PTX, and cubin.

An extremely fast Python type checker and language server, written in Rust.

verl: Volcano Engine Reinforcement Learning for LLMs

TritonParse: A Compiler Tracer, Visualizer, and Reproducer for Triton Kernels

Web-Scale Blockchain for fast, secure, scalable, decentralized apps and marketplaces.

The most widely used Python to C compiler

CUDA Python: Performance meets Productivity

Backward compatible ML compute opset inspired by HLO/MHLO

A Python-embedded DSL that makes it easy to write fast, scalable ML kernels with minimal boilerplate.

Nim is a statically typed compiled systems programming language. It combines successful concepts from mature languages like Python, Ada and Modula. Its design focuses on efficiency, expressiveness,…

A Datacenter Scale Distributed Inference Serving Framework

TORCH_LOGS parser for PT2

C Intermediate Language

FlashMLA: Efficient Multi-head Latent Attention Kernels

DeepGEMM: clean and efficient FP8 GEMM kernels with fine-grained scaling

Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels

Ecosystem of libraries and tools for writing and executing fast GPU code fully in Rust.

🔥 A cross-platform build utility based on Lua

The Nexus zkVM: The zero-knowledge virtual machine

A distributed, fast open-source graph database featuring horizontal scalability and high availability