Yufeng Li

Open Source Continuous Inference Benchmarking Qwen3.5, DeepSeek, GPTOSS - GB200 NVL72 vs MI355X vs B200 vs GB300 NVL72 vs H100 & soon™ TPUv6e/v7/Trainium2/3

Benchmark SGLang on SLURM

Official Implementation of EAGLE-1 (ICML'24), EAGLE-2 (EMNLP'24), and EAGLE-3 (NeurIPS'25).

Fast and memory-efficient exact attention

Transformer related optimization, including BERT, GPT

Fine-tuning & Reinforcement Learning for LLMs. 🦥 Train OpenAI gpt-oss, Qwen3, Llama 4, DeepSeek-R1, Gemma 3, TTS 2x faster with 70% less VRAM.

Fully open reproduction of DeepSeek-R1

CUDA Templates for Linear Algebra Subroutines

🤗 Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX.

8-bit CUDA functions for PyTorch

Development repository for the Triton language and compiler

Inference code for LLaMA models

An innovation library for efficient LLM inference via low-bit quantization and sparsity

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

Robust Speech Recognition via Large-Scale Weak Supervision

🏎️ Accelerate training and inference of 🤗 Transformers with easy to use hardware optimization tools

🤗 Diffusers: State-of-the-art diffusion models for image and audio generation in PyTorch

No description provided.

MatMul Performance Benchmarks for a Single CPU Core comparing both hand engineered and codegen kernels.

ONNX Runtime: cross-platform, high performance scoring engine for ML models

Open Neural Network Exchange

Tensors and Dynamic neural networks in Python with strong GPU acceleration

Tutorials for creating and using ONNX models

Samples for Windows ML.