An embeddable Datalog database that is Datomic-inspired written in Golang

A high-throughput and memory-efficient inference and serving engine for LLMs

Universal LLM Deployment Engine with ML Compilation

KernelBench: Can LLMs Write GPU Kernels? - Benchmark + Toolkit with Torch -> CUDA (+ more DSLs)

Nano vLLM

An invisible desktop application to help you pass your technical interviews.

DLRover: An Automatic Distributed Deep Learning System

DLRover: An Automatic Distributed Deep Learning System

Scripts for managing a large H100 cluster and fixing hardware issues to ensure smooth model training.

Optimized primitives for collective multi-GPU communication

Lightweight Virtualization Add-on for Kubernetes

NVIDIA Linux open GPU kernel module source

🔮 Execution time predictions for deep neural network training iterations across different GPUs.