State-of-the-Art Deep Learning scripts organized by models - easy to train and deploy with reproducible accuracy and performance on enterprise-grade infrastructure.

QLoRA: Efficient Finetuning of Quantized LLMs

Medusa: Simple Framework for Accelerating LLM Generation with Multiple Decoding Heads

Solutions of Reinforcement Learning, An Introduction

A throughput-oriented high-performance serving framework for LLMs

Disaggregated serving system for Large Language Models (LLMs).

DeepSeek-V3/R1 inference performance simulator

The repo to host all the web data including images for documents in dmlc projects.