NumPy

Python plugin for the IOOS Compliance Checker to validate NCEI NetCDF Templates.

Polars is a fast, Rust-backed DataFrame library with Python bindings, offering lazy and eager data processing on tabular data with Arrow-based memory.

Polars-u64-idx is a blazing-fast DataFrame library with Python bindings, delivering Rust-powered performance for tabular data using Apache Arrow.

JAX ROCm60 plugin enables ROCm-backed execution for JAX on AMD GPUs.

XLA-backed binary library powering JAX's numerical kernels, providing native bindings and the PJRT runtime for efficient array operations.

MediaPipe provides cross-platform ML pipelines and solutions with Python bindings for on-device AI, including models and tooling.

Intel-optimized NumPy provides a drop-in replacement for NumPy that uses Intel MKL to accelerate array computations in Python.

NumPy-mips64 is a Python numerical array library providing an N-dimensional array object, broadcasting, linear algebra, FFT, and integration with C/C++ and Fortran.

SHAP provides model-agnostic explanations for Python ML models, with visualizations for notebooks and support across tree ensembles, linear models, and deep learning.

ONNX is an open standard for representing ML models, enabling interoperability across frameworks and runtimes in Python.

Prophet is a Python library for automatic time series forecasting using an additive model with seasonality and holidays, designed for robust forecasts with minimal data preparation.

AKShare is a Python library that provides a simple interface to fetch diverse financial data sources for stocks, futures, funds, and more.

LightGBM Python-package provides fast gradient-boosting with a Python interface. It accepts NumPy inputs and offers integration with pandas and scikit-learn.

Python bindings for C++ code via pybind11, enabling seamless native extensions in CPython.

ydata-profiling generates a profiling report for pandas DataFrames, providing descriptive statistics, data quality insights, and visuals for quick exploratory data analysis.

Pandas-profiling generates a quick, automated EDA report for Pandas DataFrames with statistics and visuals; it's deprecated in favor of ydata-profiling.

Python framework to evaluate language models across benchmarks, enabling automated experimentation with transformer-based models.

Python bindings for COLMAP enabling Python access to COLMAP's SfM/MVS pipeline; uses NumPy arrays for data interchange.

Python data quality and testing framework for validating data in pipelines using expectations.

Taichi Lang is a Python-embedded language for high-performance numerical computation with GPU acceleration, featuring a JIT compiler and tight NumPy/PyTorch integration.