Ongoing research training transformer models at scale

Agent-friendly GPU profile-query CLI

OpenShell is the safe, private runtime for autonomous AI agents.

Stable and Efficient Reinforcement Learning for Trillion-Parameter LLMs

https://github.com/gpu-mode/reference-kernels

https://github.com/gpu-mode/reference-kernels

Spherelab Research Blog Website

GPU Functional Descriptor for memory access

TokenSpeed is a speed-of-light LLM inference engine.

WeChat group member activity analyzer - generate Excel with avatars, activity stats and cleanup suggestions. Based on Wetrace API.

Rust implementation of RaBitQ + IVF and MSTG (multi-scale tree graph)

self use tilelang scripts.

Implementation of <Orthogonal Model Merging>

OpenViking is an open-source context database designed specifically for AI Agents(such as openclaw). OpenViking unifies the management of context (memory, resources, and skills) that Agents need th…

A data processing framework

cuTile is a programming model for writing parallel kernels for NVIDIA GPUs

lab of map reduce

MineContext is your proactive context-aware AI partner（Context-Engineering+ChatGPT Pulse）

Offline optimization of your disaggregated Dynamo graph

A list of papers for Graph Retrieval-Augmented Generation (GraphRAG).

🔥 Comprehensive survey on Context Engineering: from prompt engineering to production-grade AI systems. hundreds of papers, frameworks, and implementation guides for LLMs and AI agents.

CUDA Core Compute Libraries