A Python framework for high performance GPU simulation and graphics

An extremely fast Python package and project manager, written in Rust.

CUDA Templates for Linear Algebra Subroutines

HERRO is a highly-accurate, haplotype-aware, deep-learning tool for error correction of Nanopore R10.4.1 or R9.4.1 reads (read length of >= 10 kbps is recommended).

Rust bindings for Metal

Distribute and run LLMs with a single file.

Apple AMX Instruction Set

Apple GPU microarchitecture

Chromium running inside your terminal

A header-only thread pool for C++

Perf monitoring CLI tool for Apple Silicon

Oxford Nanopore's Basecaller

Pod5: a high performance file format for nanopore reads.

Blitzing Fast CTC Beam Search Decoder

The canonical spec for ulid

Mold: A Modern Linker 🦠

WebAssembly modules for genomics

Simple task/command runner with declarative goals and dependencies

A PyTorch Basecaller for Oxford Nanopore Reads

ldd as a tree

Implementation of Fast Transformer in Pytorch

Benchmarking bonito basecaller using E. coli assemblies

GPU Accelerated JavaScript

Tensors and Dynamic neural networks in Python with strong GPU acceleration