A place to store reusable transformer components of my own creation or found on the interwebs

Tilus is a tile-level kernel programming language with explicit control over shared memory and registers.

📚LeetCUDA: Modern CUDA Learn Notes with PyTorch for Beginners🐑, 200+ CUDA Kernels, Tensor Cores, HGEMM, FA-2 MMA.🎉

A Quirky Assortment of CuTe Kernels

Quantized LLM training in pure CUDA/C++.

SkyRL: A Modular Full-stack RL Library for LLMs

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

NVIDIA Inference Xfer Library (NIXL)

Perplexity open source garden for inference technology

Perplexity GPU Kernels

DeepEP: an efficient expert-parallel communication library

AlphaFold 3 inference pipeline.

Fast and memory-efficient exact attention

Nano vLLM

Efficient Triton Kernels for LLM Training

SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformer

NanoGPT (124M) in 3 minutes

Minimal yet performant LLM examples in pure JAX