A Directory of Python Machine Learning and Data Science Frameworks

All libraries below are free, and most are open-source.

Machine Learning

General purpouse Machine Learning

• scikit-learn – machine learning in Python
• Shogun – machine learning toolbox
• xLearn – High Performance, Easy-to-use, and Scalable Machine Learning Package
• Reproducible Experiment Platform (REP) – Machine Learning toolbox for Humans
• modAL – a modular active learning framework for Python3
• Sparkit-learn – PySpark + Scikit-learn = Sparkit-learn
• mlpack – a scalable C++ machine learning library (Python bindings)
• dlib – A toolkit for making real world machine learning and data analysis applications in C++ (Python bindings)
• MLxtend – extension and helper modules for Python’s data analysis and machine learning libraries
• tick – module for statistical learning, with a particular emphasis on time-dependent modelling
• sklearn-extensions – a consolidated package of small extensions to scikit-learn
• civisml-extensions – scikit-learn-compatible estimators from Civis Analytics
• scikit-multilearn – multi-label classification for python
• tslearn – machine learning toolkit dedicated to time-series data
• seqlearn – seqlearn is a sequence classification toolkit for Python
• pystruct – Simple structured learning framework for python
• sklearn-expertsys – Highly interpretable classifiers for scikit learn, producing easily understood decision rules instead of black box models
• skutil – A set of scikit-learn and h2o extension classes (as well as caret classes for python)
• sklearn-crfsuite – scikit-learn inspired API for CRFsuite
• RuleFit – implementation of the rulefit
• metric-learn – metric learning algorithms in Python
• pyGAM – Generalized Additive Models in Python
• luminol – Anomaly Detection and Correlation library

Automated machine learning

• TPOT – Automated Machine Learning tool that optimizes machine learning pipelines using genetic programming
• auto-sklearn – is an automated machine learning toolkit and a drop-in replacement for a scikit-learn estimator
• MLBox – a powerful Automated Machine Learning python library.

Ensemble methods

• ML-Ensemble – high performance ensemble learning
• brew – Python Ensemble Learning API
• Stacking – Simple and useful stacking library, written in Python.
• stacked_generalization – library for machine learning stacking generalization.
• vecstack – Python package for stacking (machine learning technique)

Imbalanced datasets

• imbalanced-learn – module to perform under sampling and over sampling with various techniques
• imbalanced-algorithms – Python-based implementations of algorithms for learning on imbalanced data.

Random Forests

• rpforest – a forest of random projection trees
• Random Forest Clustering – Unsupervised Clustering using Random Forests
• sklearn-random-bits-forest – wrapper of the Random Bits Forest program written by (Wang et al., 2016)
• rgf_python – Python Wrapper of Regularized Greedy Forest

Extreme Learning Machine

• Python-ELM – Extreme Learning Machine implementation in Python
• Python Extreme Learning Machine (ELM) – a machine learning technique used for classification/regression tasks
• hpelm ![alt text][gpu] – High performance implementation of Extreme Learning Machines (fast randomized neural networks).

Kernel methods

• pyFM – Factorization machines in python
• fastFM – a library for Factorization Machines
• tffm – TensorFlow implementation of an arbitrary order Factorization Machine
• liquidSVM – an implementation of SVMs
• scikit-rvm – Relevance Vector Machine implementation using the scikit-learn API

Gradient boosting

• XGBoost ![alt text][gpu] – Scalable, Portable and Distributed Gradient Boosting
• LightGBM ![alt text][gpu] – a fast, distributed, high performance gradient boosting by Microsoft
• CatBoost ![alt text][gpu] – an open-source gradient boosting on decision trees library by Yandex
• InfiniteBoost – building infinite ensembles with gradient descent
• TGBoost – Tiny Gradient Boosting Tree

Deep Learning

Keras

• Keras – a high-level neural networks API, written in Python and capable of running on top of TensorFlow, CNTK, or Theano
• keras-contrib – Keras community contributions
• Hyperas – Keras + Hyperopt: A very simple wrapper for convenient hyperparameter
• Elephas – Distributed Deep learning with Keras & Spark
• Hera – Train/evaluate a Keras model, get metrics streamed to a dashboard in your browser.
• dist-keras – Distributed Deep Learning, with a focus on distributed training
• Conx – The On-Ramp to Deep Learning
• Keras add-ons

PyTorch

• PyTorch – Tensors and Dynamic neural networks in Python with strong GPU acceleration
• torchvision – Datasets, Transforms and Models specific to Computer Vision
• torchtext – Data loaders and abstractions for text and NLP
• torchaudio – an audio library for PyTorch
• ignite – high-level library to help with training neural networks in PyTorch
• PyToune – a Keras-like framework and utilities for PyTorch
• skorch – a scikit-learn compatible neural network library that wraps pytorch
• PyTorchNet – an abstraction to train neural networks
• Aorun – intend to implement an API similar to Keras with PyTorch as backend.
• pytorch_geometric – Geometric Deep Learning Extension Library for PyTorch

Tensorflow

• TensorFlow – Computation using data flow graphs for scalable machine learning by Google
• TensorLayer – Deep Learning and Reinforcement Learning Library for Researcher and Engineer.
• TFLearn – Deep learning library featuring a higher-level API for TensorFlow
• Sonnet – TensorFlow-based neural network library by DeepMind
• TensorForce – a TensorFlow library for applied reinforcement learning
• tensorpack – a Neural Net Training Interface on TensorFlow
• Polyaxon – a platform that helps you build, manage and monitor deep learning models
• Horovod – Distributed training framework for TensorFlow
• tfdeploy – Deploy tensorflow graphs for fast evaluation and export to tensorflow-less environments running numpy
• hiptensorflow ![alt text][amd] – ROCm/HIP enabled Tensorflow
• TensorFlow Fold – Deep learning with dynamic computation graphs in TensorFlow
• tensorlm – wrapper library for text generation / language models at char and word level with RNN
• TensorLight – a high-level framework for TensorFlow
• Mesh TensorFlow – Model Parallelism Made Easier

Theano

Warning: Theano development has ceased

• Theano – is a Python library that allows you to define, optimize, and evaluate mathematical expressions
• Lasagne – Lightweight library to build and train neural networks in Theano Lasagne add-ons…
• nolearn – scikit-learn compatible neural network library (mainly for Lasagne)
• Blocks – a Theano framework for building and training neural networks
• platoon – Multi-GPU mini-framework for Theano
• NeuPy – NeuPy is a Python library for Artificial Neural Networks and Deep Learning
• scikit-neuralnetwork – Deep neural networks without the learning cliff
• Theano-MPI – MPI Parallel framework for training deep learning models built in Theano

MXNet

• MXNet – Lightweight, Portable, Flexible Distributed/Mobile Deep Learning with Dynamic, Mutation-aware Dataflow Dep Scheduler
• Gluon – a clear, concise, simple yet powerful and efficient API for deep learning (now included in MXNet)
• MXbox – simple, efficient and flexible vision toolbox for mxnet framework.
• gluon-cv – provides implementations of the state-of-the-art deep learning models in computer vision.
• gluon-nlp – NLP made easy
• MXNet ![alt text][amd] – HIP Port of MXNet

Caffe

• Caffe – a fast open framework for deep learning
• Caffe2 – a lightweight, modular, and scalable deep learning framework
• hipCaffe ![alt text][amd] – the HIP port of Caffe

CNTK

• CNTK – Microsoft Cognitive Toolkit (CNTK), an open source deep-learning toolkit

Chainer

• Chainer – a flexible framework for neural networks
• ChainerRL – a deep reinforcement learning library built on top of Chainer.
• ChainerCV – a Library for Deep Learning in Computer Vision
• ChainerMN – scalable distributed deep learning with Chainer
• scikit-chainer – scikit-learn like interface to chainer
• chainer_sklearn – Sklearn (Scikit-learn) like interface for Chainer

Others

• SKIL Skymind’s platform for distributed training of machine learning models, tracking machine learning experiments, deploying models to production and managing them over their lifecycle.
• Neon – Intel Nervana™ reference deep learning framework committed to best performance on all hardware
• Tangent – Source-to-Source Debuggable Derivatives in Pure Python
• autograd – Efficiently computes derivatives of numpy code
• Myia – deep learning framework (pre-alpha)
• nnabla – Neural Network Libraries by Sony

Model explanation

• Auralisation – auralisation of learned features in CNN (for audio)
• CapsNet-Visualization – a visualization of the CapsNet layers to better understand how it works
• lucid – a collection of infrastructure and tools for research in neural network interpretability.
• Netron – visualizer for deep learning and machine learning models (no Python code, but visualizes models from most Python Deep Learning frameworks)
• FlashLight – visualization Tool for your NeuralNetwork
• tensorboard-pytorch – tensorboard for pytorch (and chainer, mxnet, numpy, …)
• anchor – code for “High-Precision Model-Agnostic Explanations” paper
• aequitas – Bias and Fairness Audit Toolkit
• Contrastive Explanation – Contrastive Explanation (Foil Trees)
• yellowbrick – visual analysis and diagnostic tools to facilitate machine learning model selection
• scikit-plot – an intuitive library to add plotting functionality to scikit-learn objects
• shap – a unified approach to explain the output of any machine learning model
• ELI5 – a library for debugging/inspecting machine learning classifiers and explaining their predictions
• Lime – Explaining the predictions of any machine learning classifier
• FairML – FairML is a python toolbox auditing the machine learning models for bias
• L2X – Code for replicating the experiments in the paper Learning to Explain: An Information-Theoretic Perspective on Model Interpretation
• PDPbox – partial dependence plot toolbox
• pyBreakDown – Python implementation of R package breakDown
• PyCEbox – Python Individual Conditional Expectation Plot Toolbox
• Skater – Python Library for Model Interpretation
• tensorflow/model-analysis – Model analysis tools for TensorFlow
• themis-ml – a library that implements fairness-aware machine learning algorithms
• treeinterpreter [alt text][skl] -interpreting scikit-learn’s decision tree and random forest predictions

Reinforcement Learning

• OpenAI Gym – a toolkit for developing and comparing reinforcement learning algorithms.

Distributed computing systems

• PySpark – exposes the Spark programming model to Python
• Veles – Distributed machine learning platform by Samsung
• Jubatus – Framework and Library for Distributed Online Machine Learning
• DMTK – Microsoft Distributed Machine Learning Toolkit
• PaddlePaddle – PArallel Distributed Deep LEarning by Baidu
• dask-ml – Distributed and parallel machine learning
• Distributed – Distributed computation in Python

Probabilistic methods

• pomegranate ![alt text][cp] – probabilistic and graphical models for Python
• pyro – a flexible, scalable deep probabilistic programming library built on PyTorch.
• ZhuSuan – Bayesian Deep Learning
• PyMC – Bayesian Stochastic Modelling in Python
• PyMC3 – Python package for Bayesian statistical modeling and Probabilistic Machine Learning
• sampled – Decorator for reusable models in PyMC3
• Edward – A library for probabilistic modeling, inference, and criticism.
• InferPy – Deep Probabilistic Modelling Made Easy
• GPflow – Gaussian processes in TensorFlow
• PyStan – Bayesian inference using the No-U-Turn sampler (Python interface)
• gelato – Bayesian dessert for Lasagne
• sklearn-bayes – Python package for Bayesian Machine Learning with scikit-learn API
• bayesloop – Probabilistic programming framework that facilitates objective model selection for time-varying parameter models
• PyFlux – Open source time series library for Python
• skggm – estimation of general graphical models
• pgmpy – a python library for working with Probabilistic Graphical Models.
• skpro – supervised domain-agnostic prediction framework for probabilistic modelling by The Alan Turing Institute
• Aboleth – a bare-bones TensorFlow framework for Bayesian deep learning and Gaussian process approximation
• PtStat – Probabilistic Programming and Statistical Inference in PyTorch
• PyVarInf – Bayesian Deep Learning methods with Variational Inference for PyTorch
• emcee – The Python ensemble sampling toolkit for affine-invariant MCMC
• hsmmlearn – a library for hidden semi-Markov models with explicit durations
• pyhsmm – bayesian inference in HSMMs and HMMs
• GPyTorch – a highly efficient and modular implementation of Gaussian Processes in PyTorch
• Bayes – Python implementations of Naive Bayes algorithm variants

Genetic Programming

• gplearn – Genetic Programming in Python
• DEAP – Distributed Evolutionary Algorithms in Python
• karoo_gp – A Genetic Programming platform for Python with GPU support
• monkeys – A strongly-typed genetic programming framework for Python
• sklearn-genetic – Genetic feature selection module for scikit-learn

Optimization

• Spearmint – Bayesian optimization
• SMAC3 – Sequential Model-based Algorithm Configuration
• Optunity – is a library containing various optimizers for hyperparameter tuning.
• hyperopt – Distributed Asynchronous Hyperparameter Optimization in Python
• hyperopt-sklearn – hyper-parameter optimization for sklearn
• sklearn-deap – use evolutionary algorithms instead of gridsearch in scikit-learn
• sigopt_sklearn – SigOpt wrappers for scikit-learn methods
• Bayesian Optimization – A Python implementation of global optimization with gaussian processes.
• SafeOpt – Safe Bayesian Optimization
• scikit-optimize – Sequential model-based optimization with a scipy.optimizeinterface
• Solid – A comprehensive gradient-free optimization framework written in Python
• PySwarms – A research toolkit for particle swarm optimization in Python
• Platypus – A Free and Open Source Python Library for Multiobjective Optimization
• GPflowOpt – Bayesian Optimization using GPflow
• POT – Python Optimal Transport library
• Talos – Hyperparameter Optimization for Keras Models

Natural Language Processing

• NLTK – modules, data sets, and tutorials supporting research and development in Natural Language Processing
• CLTK – The Classical Language Toolkik
• gensim – Topic Modelling for Humans
• PSI-Toolkit – a natural language processing toolkit by Adam Mickiewicz University in Poznań
• pyMorfologik – Python binding for Morfologik (Polish morphological analyzer)
• skift – scikit-learn wrappers for Python fastText.
• Phonemizer – Simple text to phonemes converter for multiple languages

Computer Audition

• librosa – Python library for audio and music analysis
• Yaafe – Audio features extraction
• aubio – a library for audio and music analysis
• Essentia – library for audio and music analysis, description and synthesis
• LibXtract – is a simple, portable, lightweight library of audio feature extraction functions
• Marsyas – Music Analysis, Retrieval and Synthesis for Audio Signals
• muda – a library for augmenting annotated audio data
• madmom – Python audio and music signal processing library

Computer Vision

• OpenCV – Open Source Computer Vision Library
• scikit-image – Image Processing SciKit (Toolbox for SciPy)
• imgaug – image augmentation for machine learning experiments
• imgaug_extension – additional augmentations for imgaug
• Augmentor – Image augmentation library in Python for machine learning
• albumentations – fast image augmentation library and easy to use wrapper around other libraries

Feature engineering

• Featuretools – automated feature engineering
• scikit-feature – feature selection repository in python
• skl-groups – scikit-learn addon to operate on set/”group”-based features
• Feature Forge – a set of tools for creating and testing machine learning feature
• boruta_py – implementations of the Boruta all-relevant feature selection method
• BoostARoota – a fast xgboost feature selection algorithm
• few – a feature engineering wrapper for sklearn
• scikit-rebate – a scikit-learn-compatible Python implementation of ReBATE, a suite of Relief-based feature selection algorithms for Machine Learning
• scikit-mdr – a sklearn-compatible Python implementation of Multifactor Dimensionality Reduction (MDR) for feature construction.
• tsfresh – Automatic extraction of relevant features from time series

Data manipulation & pipelines

• pandas – powerful Python data analysis toolkit
• sklearn-pandas – Pandas integration with sklearn
• alexander – wrapper that aims to make scikit-learn fully compatible with pandas
• blaze – NumPy and Pandas interface to Big Data
• pandasql – allows you to query pandas DataFrames using SQL syntax
• pandas-gbq – Pandas Google Big Query
• xpandas – universal 1d/2d data containers with Transformers functionality for data analysis by The Alan Turing Institute
• Fuel – data pipeline framework for machine learning
• Arctic – high performance datastore for time series and tick data
• pdpipe – sasy pipelines for pandas DataFrames.
• SSPipe – Python pipe () operator with support for DataFrames and Numpy and Pytorch
• meza – a Python toolkit for processing tabular data
• pandas-ply – functional data manipulation for pandas
• Dplython – Dplyr for Python
• pysparkling – a pure Python implementation of Apache Spark’s RDD and DStream interfaces
• quinn – pyspark methods to enhance developer productivity
• Dataset – helps you conveniently work with random or sequential batches of your data and define data processing
• swifter – a package which efficiently applies any function to a pandas dataframe or series in the fastest available manner

Statistics

• statsmodels – statistical modeling and econometrics in Python
• stockstats – Supply a wrapper StockDataFrame based on the pandas.DataFramewith inline stock statistics/indicators support.
• simplestatistics – simple statistical functions implemented in readable Python.
• weightedcalcs – pandas-based utility to calculate weighted means, medians, distributions, standard deviations, and more
• scikit-posthocs – Pairwise Multiple Comparisons Post-hoc Tests
• pysie – provides python implementation of statistical inference engine

Experiments tools

• Sacred – a tool to help you configure, organize, log and reproduce experiments by IDSIA
• Xcessiv – a web-based application for quick, scalable, and automated hyperparameter tuning and stacked ensembling
• Persimmon – A visual dataflow programming language for sklearn

Visualization

• Matplotlib – plotting with Python
• seaborn – statistical data visualization using matplotlib
• Bokeh – Interactive Web Plotting for Python
• HoloViews – stop plotting your data – annotate your data and let it visualize itself
• Alphalens – performance analysis of predictive (alpha) stock factors by Quantopian
• python-ternary – ternary plotting library for python with matplotlib
• Naarad – framework for performance analysis & rating of sharded & stateful services.

Evaluation

• kaggle-metrics – Metrics for Kaggle competitions
• Metrics – machine learning evaluation metric
• sklearn-evaluation – scikit-learn model evaluation made easy: plots, tables and markdown reports

Computations

• numpy – the fundamental package needed for scientific computing with Python.
• Dask – parallel computing with task scheduling
• bottleneck – Fast NumPy array functions written in C
• minpy – NumPy interface with mixed backend execution
• CuPy – NumPy-like API accelerated with CUDA
• scikit-tensor – Python library for multilinear algebra and tensor factorizations
• numdifftools – solve automatic numerical differentiation problems in one or more variables
• quaternion – Add built-in support for quaternions to numpy
• adaptive – Tools for adaptive and parallel samping of mathematical functions

Spatial analysis

• GeoPandas – Python tools for geographic data
• PySal – Python Spatial Analysis Library

Quantum Computing

• QML – a Python Toolkit for Quantum Machine Learning

Conversion

• sklearn-porter – transpile trained scikit-learn estimators to C, Java, JavaScript and others
• ONNX – Open Neural Network Exchange
• MMdnn – a set of tools to help users inter-operate among different deep learning frameworks.

See Also

• Java AI
• Scala AI
• Clojure AI

Ref
https://skymind.ai