A Beginner's Guide to Python Machine Learning and Data Science Frameworks

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

Table of contents:

• Machine Learning
  • General Purpose ML
  • Automated Machine Learning
  • Ensemble methods
  • Imbalanced datasets
  • Random Forests
  • Extreme Learning Machine
  • Kernel methods
  • Gradient boosting
• Deep Learning
  • Keras
  • TensorFlow
  • Theano
  • PyTorch
  • MXnet
  • Caffe
  • CNTK
  • Chainer
  • Others
• Model explanation
• Reinforcement Learning
• Distributed computing systems
• Probabilistic methods
• Genetic Programming
• Optimization
• Natural Language Processing
• Computer Audition
• Computer Vision
• Feature engineering
• Data manipulation & pipelines
• Statistics
• Experiments tools
• Visualization
• Evaluation
• Computations
• Spatial analysis
• Quantum computing
• Conversion

Learn to build AI apps now »

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

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

• 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.optimize interface
• 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.DataFrame with 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