Stars
A version of CloverLeaf using NVIDIA's CUDA
Agentic framework for computational chemistry and materials science workflows
Paper list of agent for science
Runtime provenance for AI and scientific workflows—capture, enrich, and query workflow data via observability adapters and code annotation across edge, cloud, and HPC.
KV-Direct: Bounded-Memory Transformer Inference via Residual Stream Checkpointing
Run OpenClaw more securely inside NVIDIA OpenShell with managed inference
A high-performance and light-weight router for vLLM large scale deployment
collection of benchmarks to measure basic GPU capabilities
Harnessing distributed, tiered storage for context management
The academic meta-prompting framework for AI agents like Claude Code, Gemini CLI, OpenCode. Features citation-aware drafting, hallucination checks, and rigorous structural planning. Built for PhDs …
Tile-Based Runtime for Ultra-Low-Latency LLM Inference
TileFusion is an experimental C++ macro kernel template library that elevates the abstraction level in CUDA C for tile processing.
cuTile is a programming model for writing parallel kernels for NVIDIA GPUs
ArcticInference: vLLM plugin for high-throughput, low-latency inference
Official Implementation of APB (ACL 2025 main Oral) and Spava (ACL 2026 main).
[ICLR'26] The official code implementation for "Cache-to-Cache: Direct Semantic Communication Between Large Language Models"
Code, Data and Model for COLM 2025 Paper "E2-RAG: Towards Editable Efficient RAG by Editing Compressed KV Caches"
gLLM: Global Balanced Pipeline Parallelism System for Distributed LLM Serving with Token Throttling
Demystify AI agents by building them yourself. Local LLMs, no black boxes, real understanding of function calling, memory, and ReAct patterns.
agentUniverse is a LLM multi-agent framework that allows developers to easily build multi-agent applications.
Autonomous Agents (LLMs) research papers. Updated Daily.
Lightweight coding agent that runs in your terminal