Avx2 Python Packages

ctranslate2

Fast inference engine for Transformer models

pic-scale

High performance image scale in pure Rust

nanovector

The SQLite of Vector Search & Episodic Memory for AI Agents in ~120KB. Pure C99, AVX2+FMA, ARM NEON, FASM x64, Native MCP Server.

nanogemm

Minimalist, bare-metal SIMD & Assembly GEMM engine for Python. Sub-microsecond CPU matrix multiplication & Native MCP Server (nanogemm-mcp).

corrfunc

⚡️⚡️⚡️Blazing fast correlation functions on the CPU.

addernet

Neural "networks" with zero multiplications at inference. AdderNet for embedded systems.

voxrun-ctranslate2

Fast inference engine for Transformer models

myers-batch

Batched infix edit distance: bit-parallel Myers for aarch64 NEON and x86-64 AVX2, 7.5-10x faster than edlib single-threaded on the reference machine

osaca

Open Source Architecture Code Analyzer

turbocpp

Fast CPU LLM inference in pure C++17. AVX2+FMA matmul, Q4/Q8 quantization, TurboQuant-style 4/3-bit KV cache, GQA, BPE, mmap loader. Zero deps.

adaptq

High-performance C++17 KV cache compression engine for LLM inference with Python bindings, SIMD acceleration, and multi-backend support.

ctranslate2-arty

Fast inference engine for Transformer models

eastat

CSV column statistics powered by Ea SIMD kernels.

easobel

Fast Sobel edge detection. SIMD kernels via Eä, called from Python. 5× faster than OpenCV single-threaded.

eavec

Fast vector similarity search. SIMD kernels via Eä, called from Python. Beats FAISS 4–8× at dim=384.

pyflagstats

Efficient C functions to compute the summary statistics (flagstats) for sequencing read sets.

detect-simd

Python library to detect CPU SIMD capabilities.

voxrt-asr

On-device streaming ASR (speech-to-text) for Linux x86_64 + aarch64 — one hardened .so behind Python (PyPI), Node.js (npm), Go (go get), C/C++ (tarball + CMake). AVX2 on servers, NEON on edge. RNN-T + CTC, RTF 0.06–0.08 on x86_64. Same runtime as voxrt-asr-{android,ios}.

voxrt-kws

On-device 14-way keyword spotter for Linux x86_64 + aarch64 — one glibc-2.17 .so behind Python (PyPI), Node.js (npm), Go (go get), C / C++ (tarball + CMake). Streaming Conformer, RTF 1 % on WSL AVX2, 32 % on Pi Zero 2 W NEON. Same runtime as voxrt-kws-{android,ios} (upcoming).