Best-of Machine Learning with Python
π A ranked list of awesome machine learning Python libraries. Updated weekly.
This curated list contains 920 awesome open-source projects with a total of 5.1M stars grouped into 34 categories. All projects are ranked by a project-quality score, which is calculated based on various metrics automatically collected from GitHub and different package managers. If you like to add or update projects, feel free to open an issue, submit a pull request, or directly edit the projects.yaml. Contributions are very welcome!
π§ββοΈ Discover other best-of lists or create your own.
π« Subscribe to our newsletter for updates and trending projects.
Contents
- Machine Learning Frameworks 64 projects
- Data Visualization 55 projects
- Text Data & NLP 103 projects
- Image Data 64 projects
- Graph Data 36 projects
- Audio Data 29 projects
- Geospatial Data 22 projects
- Financial Data 25 projects
- Time Series Data 29 projects
- Medical Data 19 projects
- Tabular Data 6 projects
- Optical Character Recognition 12 projects
- Data Containers & Structures 1 projects
- Data Loading & Extraction 1 projects
- Web Scraping & Crawling 1 projects
- Data Pipelines & Streaming 2 projects
- Distributed Machine Learning 36 projects
- Hyperparameter Optimization & AutoML 52 projects
- Reinforcement Learning 23 projects
- Recommender Systems 17 projects
- Privacy Machine Learning 7 projects
- Workflow & Experiment Tracking 40 projects
- Model Serialization & Deployment 20 projects
- Model Interpretability 55 projects
- Vector Similarity Search (ANN) 13 projects
- Probabilistics & Statistics 24 projects
- Adversarial Robustness 9 projects
- GPU & Accelerator Utilities 20 projects
- Tensorflow Utilities 16 projects
- Jax Utilities 3 projects
- Sklearn Utilities 19 projects
- Pytorch Utilities 32 projects
- Database Clients 1 projects
- Others 66 projects
Explanation
Machine Learning Frameworks
General-purpose machine learning and deep learning frameworks.
Tensorflow (π₯56 Β· β 200K) - An Open Source Machine Learning Framework for Everyone. Apache-2 
-
GitHub (π¨βπ» 5K Β· π 75K Β· π¦ 540K Β· π 42K - 4% open Β· β±οΈ 30.10.2025):
git clone https://github.com/tensorflow/tensorflow -
PyPi (π₯ 26M / month Β· π¦ 9.6K Β· β±οΈ 13.08.2025):
-
Conda (π₯ 6M Β· β±οΈ 27.10.2025):
conda install -c conda-forge tensorflow -
Docker Hub (π₯ 81M Β· β 2.8K Β· β±οΈ 30.10.2025):
docker pull tensorflow/tensorflow
scikit-learn (π₯53 Β· β 64K) - scikit-learn: machine learning in Python. BSD-3 
-
GitHub (π¨βπ» 3.4K Β· π 26K Β· π₯ 1.1K Β· π¦ 1.3M Β· π 12K - 17% open Β· β±οΈ 30.10.2025):
git clone https://github.com/scikit-learn/scikit-learn -
PyPi (π₯ 140M / month Β· π¦ 35K Β· β±οΈ 09.09.2025):
-
Conda (π₯ 40M Β· β±οΈ 09.09.2025):
conda install -c conda-forge scikit-learn
XGBoost (π₯46 Β· β 28K) - Scalable, Portable and Distributed Gradient Boosting (GBDT, GBRT or.. Apache-2
PaddlePaddle (π₯46 Β· β 23K) - PArallel Distributed Deep LEarning: Machine Learning.. Apache-2 
jax (π₯45 Β· β 34K) - Composable transformations of Python+NumPy programs: differentiate,.. Apache-2
pytorch-lightning (π₯45 Β· β 30K) - Pretrain, finetune ANY AI model of ANY size on 1 or.. Apache-2 
-
GitHub (π¨βπ» 1K Β· π 3.6K Β· π₯ 15K Β· π¦ 48K Β· π 7.4K - 11% open Β· β±οΈ 29.10.2025):
git clone https://github.com/Lightning-AI/lightning -
PyPi (π₯ 9.8M / month Β· π¦ 1.8K Β· β±οΈ 05.09.2025):
pip install pytorch-lightning -
Conda (π₯ 1.7M Β· β±οΈ 05.09.2025):
conda install -c conda-forge pytorch-lightning
StatsModels (π₯45 Β· β 11K) - Statsmodels: statistical modeling and econometrics in Python. BSD-3
LightGBM (π₯42 Β· β 18K) - A fast, distributed, high performance gradient boosting (GBT, GBDT, GBRT,.. MIT
Catboost (π₯42 Β· β 8.6K) - A fast, scalable, high performance Gradient Boosting on Decision.. Apache-2
Flax (π₯38 Β· β 6.9K) - Flax is a neural network library for JAX that is designed for.. Apache-2 
Ignite (π₯36 Β· β 4.7K) - High-level library to help with training and evaluating neural.. BSD-3 
einops (π₯35 Β· β 9.2K) - Flexible and powerful tensor operations for readable and reliable code.. MIT
Jina (π₯33 Β· β 22K Β· π€) - Build multimodal AI applications with cloud-native stack. Apache-2
-
GitHub (π¨βπ» 180 Β· π 2.2K Β· β±οΈ 24.03.2025):
git clone https://github.com/jina-ai/jina -
PyPi (π₯ 120K / month Β· π¦ 29 Β· β±οΈ 24.03.2025):
-
Conda (π₯ 110K Β· β±οΈ 22.04.2025):
conda install -c conda-forge jina-core -
Docker Hub (π₯ 1.8M Β· β 9 Β· β±οΈ 24.03.2025):
Thinc (π₯33 Β· β 2.9K Β· π€) - A refreshing functional take on deep learning, compatible with your.. MIT
Ludwig (π₯32 Β· β 12K Β· π€) - Low-code framework for building custom LLMs, neural networks,.. Apache-2 
tensorflow-upstream (π₯31 Β· β 700) - TensorFlow ROCm port. Apache-2 
Geomstats (π₯30 Β· β 1.4K) - Computations and statistics on manifolds with geometric structures. MIT
pyRiemann (π₯28 Β· β 700) - Machine learning for multivariate data through the Riemannian.. BSD-3 
NuPIC (π₯27 Β· β 6.4K Β· π€) - Numenta Platform for Intelligent Computing is an implementation of.. MIT
Determined (π₯26 Β· β 3.2K Β· π€) - Determined is an open-source machine learning.. Apache-2 

Neural Network Libraries (π₯26 Β· β 2.8K) - Neural Network Libraries. Apache-2
deepinv (π₯26 Β· β 540) - DeepInverse: a PyTorch library for solving imaging inverse problems.. BSD-3
Towhee (π₯23 Β· β 3.4K Β· π€) - Towhee is a framework that is dedicated to making neural data.. Apache-2
Runhouse (π₯21 Β· β 1.1K) - Distribute and run AI workloads magically in Python, like PyTorch.. Apache-2
NeoML (π₯19 Β· β 790) - Machine learning framework for both deep learning and traditional.. Apache-2
chefboost (π₯19 Β· β 480) - A Lightweight Decision Tree Framework supporting regular algorithms:.. MIT
ThunderGBM (π₯18 Β· β 710 Β· π€) - ThunderGBM: Fast GBDTs and Random Forests on GPUs. Apache-2
Show 26 hidden projects...
- dlib (π₯40 Β· β 14K) - A toolkit for making real world machine learning and data analysis..
βοΈBSL-1.0 - MXNet (π₯38 Β· β 21K Β· π) - Lightweight, Portable, Flexible Distributed/Mobile Deep..
Apache-2 - Theano (π₯37 Β· β 10K Β· π) - Theano was a Python library that allows you to define, optimize, and..
BSD-3 - MindsDB (π₯33 Β· β 37K) - Federated query engine for AI - The only MCP Server youll ever need.
βοΈICU - Vowpal Wabbit (π₯33 Β· β 8.6K Β· π) - Vowpal Wabbit is a machine learning system which pushes the..
BSD-3 - Chainer (π₯33 Β· β 5.9K Β· π) - A flexible framework of neural networks for deep learning.
MIT - Turi Create (π₯32 Β· β 11K Β· π) - Turi Create simplifies the development of custom machine..
BSD-3 - tensorpack (π₯32 Β· β 6.3K Β· π) - A Neural Net Training Interface on TensorFlow, with..
Apache-2 - TFlearn (π₯31 Β· β 9.6K Β· π) - Deep learning library featuring a higher-level API for TensorFlow.
MIT - dyNET (π₯31 Β· β 3.4K Β· π) - DyNet: The Dynamic Neural Network Toolkit.
Apache-2 - CNTK (π₯29 Β· β 18K Β· π) - Microsoft Cognitive Toolkit (CNTK), an open source deep-learning toolkit.
MIT - Lasagne (π₯28 Β· β 3.9K Β· π) - Lightweight library to build and train neural networks in Theano.
MIT - SHOGUN (π₯26 Β· β 3.1K Β· π) - Unified and efficient Machine Learning.
BSD-3 - ktrain (π₯26 Β· β 1.3K Β· π) - ktrain is a Python library that makes deep learning and AI..
Apache-2 - NeuPy (π₯25 Β· β 740 Β· π) - NeuPy is a Tensorflow based python library for prototyping and building..
MIT - xLearn (π₯24 Β· β 3.1K Β· π) - High performance, easy-to-use, and scalable machine learning (ML)..
Apache-2 - EvaDB (π₯24 Β· β 2.7K Β· π) - Database system for AI-powered apps.
Apache-2 - neon (π₯22 Β· β 3.9K Β· π) - Intel Nervana reference deep learning framework committed to best..
Apache-2 - ThunderSVM (π₯22 Β· β 1.6K Β· π) - ThunderSVM: A Fast SVM Library on GPUs and CPUs.
Apache-2 - Torchbearer (π₯22 Β· β 640 Β· π) - torchbearer: A model fitting library for PyTorch.
MIT - mace (π₯21 Β· β 5K Β· π) - MACE is a deep learning inference framework optimized for mobile..
Apache-2 - Neural Tangents (π₯21 Β· β 2.4K Β· π) - Fast and Easy Infinite Neural Networks in Python.
Apache-2 - Objax (π₯20 Β· β 770 Β· π) - Objax is a machine learning framework that provides an Object..
Apache-2
- elegy (π₯19 Β· β 480 Β· π) - A High Level API for Deep Learning in JAX.
MIT
- StarSpace (π₯16 Β· β 4K Β· π) - Learning embeddings for classification, retrieval and ranking.
MIT - nanodl (π₯14 Β· β 300 Β· π) - A Jax-based library for building transformers, includes..
MIT
Data Visualization
General-purpose and task-specific data visualization libraries.
Matplotlib (π₯49 Β· β 22K) - matplotlib: plotting with Python. βUnlicensed
Plotly (π₯47 Β· β 18K) - The interactive graphing library for Python. MIT
-
GitHub (π¨βπ» 300 Β· π 2.7K Β· π₯ 550 Β· π¦ 460K Β· π 3.3K - 21% open Β· β±οΈ 28.10.2025):
git clone https://github.com/plotly/plotly.py -
PyPi (π₯ 37M / month Β· π¦ 9.7K Β· β±οΈ 02.10.2025):
-
Conda (π₯ 12M Β· β±οΈ 03.10.2025):
conda install -c conda-forge plotly -
npm (π₯ 2.8K / month Β· π¦ 9 Β· β±οΈ 12.01.2021):
PyVista (π₯38 Β· β 3.3K) - 3D plotting and mesh analysis through a streamlined interface for.. MIT 
HoloViews (π₯38 Β· β 2.8K) - With Holoviews, your data visualizes itself. BSD-3 
-
GitHub (π¨βπ» 150 Β· π 410 Β· π¦ 17K Β· π 3.4K - 31% open Β· β±οΈ 29.10.2025):
git clone https://github.com/holoviz/holoviews -
PyPi (π₯ 820K / month Β· π¦ 490 Β· β±οΈ 29.10.2025):
-
Conda (π₯ 2.4M Β· β±οΈ 25.06.2025):
conda install -c conda-forge holoviews -
npm (π₯ 380 / month Β· π¦ 7 Β· β±οΈ 20.06.2025):
npm install @pyviz/jupyterlab_pyviz
PyQtGraph (π₯37 Β· β 4.2K) - Fast data visualization and GUI tools for scientific / engineering.. MIT
pandas-profiling (π₯35 Β· β 13K) - 1 Line of code data quality profiling & exploratory.. MIT 

-
GitHub (π¨βπ» 140 Β· π 1.7K Β· π₯ 490 Β· π¦ 6.9K Β· π 850 - 30% open Β· β±οΈ 19.09.2025):
git clone https://github.com/ydataai/pandas-profiling -
PyPi (π₯ 330K / month Β· π¦ 180 Β· β±οΈ 03.02.2023):
pip install pandas-profiling -
Conda (π₯ 590K Β· β±οΈ 22.04.2025):
conda install -c conda-forge pandas-profiling
cartopy (π₯35 Β· β 1.5K) - Cartopy - a cartographic python library with matplotlib support. BSD-3
VisPy (π₯34 Β· β 3.5K Β· π) - High-performance interactive 2D/3D data visualization library. BSD-3 
-
GitHub (π¨βπ» 210 Β· π 620 Β· π¦ 2.1K Β· π 1.5K - 25% open Β· β±οΈ 13.10.2025):
git clone https://github.com/vispy/vispy -
PyPi (π₯ 190K / month Β· π¦ 200 Β· β±οΈ 19.05.2025):
-
Conda (π₯ 980K Β· β±οΈ 30.08.2025):
conda install -c conda-forge vispy -
npm (π₯ 12 / month Β· π¦ 3 Β· β±οΈ 15.03.2020):
datashader (π₯34 Β· β 3.5K) - Quickly and accurately render even the largest data. BSD-3
lets-plot (π₯34 Β· β 1.7K) - Multiplatform plotting library based on the Grammar of Graphics. MIT
Perspective (π₯33 Β· β 9.5K) - A data visualization and analytics component, especially.. Apache-2 
-
GitHub (π¨βπ» 100 Β· π 1.2K Β· π₯ 12K Β· π¦ 190 Β· π 890 - 12% open Β· β±οΈ 29.10.2025):
git clone https://github.com/finos/perspective -
PyPi (π₯ 17K / month Β· π¦ 31 Β· β±οΈ 28.10.2025):
pip install perspective-python -
Conda (π₯ 2.4M Β· β±οΈ 28.10.2025):
conda install -c conda-forge perspective -
npm (π₯ 600 / month Β· π¦ 6 Β· β±οΈ 03.09.2025):
npm install @finos/perspective-jupyterlab
hvPlot (π₯32 Β· β 1.3K) - A high-level plotting API for pandas, dask, xarray, and networkx built.. BSD-3
mpld3 (π₯31 Β· β 2.4K Β· π) - An interactive data visualization tool which brings matplotlib.. BSD-3
-
GitHub (π¨βπ» 54 Β· π 360 Β· π¦ 7.6K Β· π 370 - 59% open Β· β±οΈ 27.07.2025):
git clone https://github.com/mpld3/mpld3 -
PyPi (π₯ 440K / month Β· π¦ 160 Β· β±οΈ 27.07.2025):
-
Conda (π₯ 280K Β· β±οΈ 28.07.2025):
conda install -c conda-forge mpld3 -
npm (π₯ 900 / month Β· π¦ 11 Β· β±οΈ 27.07.2025):
bqplot (π₯30 Β· β 3.7K) - Plotting library for IPython/Jupyter notebooks. Apache-2 
-
GitHub (π¨βπ» 66 Β· π 480 Β· π¦ 62 Β· π 650 - 42% open Β· β±οΈ 25.08.2025):
git clone https://github.com/bqplot/bqplot -
PyPi (π₯ 230K / month Β· π¦ 110 Β· β±οΈ 21.05.2025):
-
Conda (π₯ 1.9M Β· β±οΈ 02.09.2025):
conda install -c conda-forge bqplot -
npm (π₯ 3K / month Β· π¦ 21 Β· β±οΈ 03.09.2025):
D-Tale (π₯29 Β· β 5K) - Visualizer for pandas data structures. βοΈLGPL-2.1 

Plotly-Resampler (π₯27 Β· β 1.2K) - Visualize large time series data with plotly.py. MIT
-
GitHub (π¨βπ» 14 Β· π 74 Β· π¦ 2K Β· π 190 - 32% open Β· β±οΈ 03.09.2025):
git clone https://github.com/predict-idlab/plotly-resampler -
PyPi (π₯ 370K / month Β· π¦ 38 Β· β±οΈ 29.08.2025):
pip install plotly-resampler -
Conda (π₯ 140K Β· β±οΈ 09.10.2025):
conda install -c conda-forge plotly-resampler
HyperTools (π₯26 Β· β 1.9K) - A Python toolbox for gaining geometric insights into high-dimensional.. MIT
data-validation (π₯25 Β· β 780) - Library for exploring and validating machine learning.. Apache-2 

Chartify (π₯24 Β· β 3.6K Β· π€) - Python library that makes it easy for data scientists to create.. Apache-2
vegafusion (π₯21 Β· β 390) - Serverside scaling for Vega and Altair visualizations. BSD-3
-
GitHub (π¨βπ» 6 Β· π 26 Β· π₯ 6.6K Β· π 150 - 36% open Β· β±οΈ 29.09.2025):
git clone https://github.com/vegafusion/vegafusion -
PyPi (π₯ 770 / month Β· π¦ 2 Β· β±οΈ 09.05.2024):
pip install vegafusion-jupyter -
Conda (π₯ 520K Β· β±οΈ 27.10.2025):
conda install -c conda-forge vegafusion-python-embed -
npm (π₯ 1.9K / month Β· π¦ 3 Β· β±οΈ 09.05.2024):
npm install vegafusion-jupyter
Show 22 hidden projects...
Text Data & NLP
Libraries for processing, cleaning, manipulating, and analyzing text data as well as libraries for NLP tasks such as language detection, fuzzy matching, classification, seq2seq learning, conversational AI, keyword extraction, and translation.
transformers (π₯54 Β· β 150K) - Transformers: the model-definition framework for.. Apache-2 

nltk (π₯47 Β· β 14K) - Suite of libraries and programs for symbolic and statistical natural.. Apache-2
litellm (π₯45 Β· β 30K Β· π) - Python SDK, Proxy Server (LLM Gateway) to call 100+.. MIT o t h e r s
spaCy (π₯43 Β· β 33K Β· π) - Industrial-strength Natural Language Processing (NLP) in Python. MIT
sentence-transformers (π₯42 Β· β 18K) - State-of-the-Art Text Embeddings. Apache-2 
-
GitHub (π¨βπ» 240 Β· π 2.7K Β· π¦ 120K Β· π 2.5K - 51% open Β· β±οΈ 22.10.2025):
git clone https://github.com/UKPLab/sentence-transformers -
PyPi (π₯ 17M / month Β· π¦ 3.7K Β· β±οΈ 22.10.2025):
pip install sentence-transformers -
Conda (π₯ 1M Β· β±οΈ 22.10.2025):
conda install -c conda-forge sentence-transformers
gensim (π₯42 Β· β 16K) - Topic Modelling for Humans. βοΈLGPL-2.1
sentencepiece (π₯42 Β· β 11K) - Unsupervised text tokenizer for Neural Network-based text.. Apache-2
-
GitHub (π¨βπ» 100 Β· π 1.3K Β· π₯ 110K Β· π¦ 120K Β· π 800 - 3% open Β· β±οΈ 04.10.2025):
git clone https://github.com/google/sentencepiece -
PyPi (π₯ 31M / month Β· π¦ 2.4K Β· β±οΈ 12.08.2025):
pip install sentencepiece -
Conda (π₯ 1.7M Β· β±οΈ 22.09.2025):
conda install -c conda-forge sentencepiece
Tokenizers (π₯40 Β· β 10K) - Fast State-of-the-Art Tokenizers optimized for Research and.. Apache-2
haystack (π₯37 Β· β 23K) - AI orchestration framework to build customizable, production-ready.. Apache-2
Opik (π₯37 Β· β 15K) - Debug, evaluate, and monitor your LLM applications, RAG systems, and.. Apache-2
ChatterBot (π₯37 Β· β 14K) - ChatterBot is a machine learning, conversational dialog engine for.. BSD-3
flair (π₯37 Β· β 14K) - A very simple framework for state-of-the-art Natural Language Processing.. MIT 
TextBlob (π₯37 Β· β 9.5K) - Simple, Pythonic, text processing--Sentiment analysis, part-of-speech.. MIT
fairseq (π₯36 Β· β 32K) - Facebook AI Research Sequence-to-Sequence Toolkit written in Python. MIT 
stanza (π₯36 Β· β 7.6K) - Stanford NLP Python library for tokenization, sentence segmentation,.. Apache-2
qdrant (π₯35 Β· β 27K) - Qdrant - High-performance, massive-scale Vector Database and Vector.. Apache-2
-
GitHub (π¨βπ» 140 Β· π 1.9K Β· π₯ 500K Β· π¦ 120 Β· π 1.6K - 22% open Β· β±οΈ 30.09.2025):
git clone https://github.com/qdrant/qdrant
Rasa (π₯34 Β· β 21K) - Open source machine learning framework to automate text- and voice-.. Apache-2 
TensorFlow Text (π₯34 Β· β 1.3K) - Making text a first-class citizen in TensorFlow. Apache-2 
snowballstemmer (π₯34 Β· β 810) - Snowball compiler and stemming algorithms. BSD-3
-
GitHub (π¨βπ» 41 Β· π 190 Β· π¦ 11 Β· π 120 - 17% open Β· β±οΈ 28.10.2025):
git clone https://github.com/snowballstem/snowball -
PyPi (π₯ 24M / month Β· π¦ 550 Β· β±οΈ 09.05.2025):
pip install snowballstemmer -
Conda (π₯ 11M Β· β±οΈ 20.05.2025):
conda install -c conda-forge snowballstemmer
torchtext (π₯32 Β· β 3.6K) - Models, data loaders and abstractions for language processing,.. BSD-3 
jellyfish (π₯32 Β· β 2.2K) - a python library for doing approximate and phonetic matching of strings. MIT
DeepPavlov (π₯31 Β· β 6.9K Β· π€) - An open source library for deep learning end-to-end.. Apache-2 
ftfy (π₯31 Β· β 4K Β· π€) - Fixes mojibake and other glitches in Unicode text, after the fact. Apache-2
SciSpacy (π₯31 Β· β 1.9K) - A full spaCy pipeline and models for scientific/biomedical documents. Apache-2
english-words (π₯29 Β· β 12K Β· π€) - A text file containing 479k English words for all your.. Unlicense
rubrix (π₯29 Β· β 4.7K) - Argilla is a collaboration tool for AI engineers and domain experts.. Apache-2
Dedupe (π₯29 Β· β 4.4K Β· π) - A python library for accurate and scalable fuzzy matching, record.. MIT
TextDistance (π₯28 Β· β 3.5K) - Compute distance between sequences. 30+ algorithms, pure python.. MIT
spacy-transformers (π₯28 Β· β 1.4K) - Use pretrained transformers like BERT, XLNet and GPT-2.. MIT spacy
-
GitHub (π¨βπ» 23 Β· π 170 Β· π₯ 610 Β· π¦ 2.4K Β· β±οΈ 26.05.2025):
git clone https://github.com/explosion/spacy-transformers -
PyPi (π₯ 270K / month Β· π¦ 110 Β· β±οΈ 26.05.2025):
pip install spacy-transformers -
Conda (π₯ 140K Β· β±οΈ 22.04.2025):
conda install -c conda-forge spacy-transformers
detoxify (π₯26 Β· β 1.1K) - Trained models & code to predict toxic comments on all 3 Jigsaw.. Apache-2
scattertext (π₯25 Β· β 2.3K) - Beautiful visualizations of how language differs among document.. Apache-2
T5 (π₯24 Β· β 6.4K) - Code for the paper Exploring the Limits of Transfer Learning with a.. Apache-2 
happy-transformer (π₯23 Β· β 540 Β· π€) - Happy Transformer makes it easy to fine-tune and.. Apache-2 huggingface
Sockeye (π₯21 Β· β 1.2K Β· π€) - Sequence-to-sequence framework with a focus on Neural.. Apache-2 
small-text (π₯20 Β· β 630) - Active Learning for Text Classification in Python. MIT 

textaugment (π₯19 Β· β 430) - TextAugment: Text Augmentation Library. MIT
VizSeq (π₯15 Β· β 450) - An Analysis Toolkit for Natural Language Generation (Translation,.. MIT
Show 59 hidden projects...
- AllenNLP (π₯36 Β· β 12K Β· π) - An open-source NLP research library, built on PyTorch.
Apache-2 - fastText (π₯34 Β· β 26K Β· π) - Library for fast text representation and classification.
MIT - OpenNMT (π₯33 Β· β 7K Β· π) - Open Source Neural Machine Translation and (Large) Language Models..
MIT - ParlAI (π₯32 Β· β 11K Β· π) - A framework for training and evaluating AI models on a variety of..
MIT - fuzzywuzzy (π₯31 Β· β 9.3K Β· π) - Fuzzy String Matching in Python.
βοΈGPL-2.0 - Sumy (π₯30 Β· β 3.6K Β· π) - Module for automatic summarization of text documents and HTML pages.
Apache-2 - underthesea (π₯30 Β· β 1.6K) - Underthesea - Vietnamese NLP Toolkit.
βοΈGPL-3.0 - nlpaug (π₯29 Β· β 4.6K Β· π) - Data augmentation for NLP.
MIT - vaderSentiment (π₯28 Β· β 4.9K Β· π) - VADER Sentiment Analysis. VADER (Valence Aware Dictionary..
MIT - textacy (π₯28 Β· β 2.2K Β· π) - NLP, before and after spaCy.
βUnlicensed - PyTextRank (π₯28 Β· β 2.2K Β· π) - Python implementation of TextRank algorithms (textgraphs) for..
MIT - Ciphey (π₯27 Β· β 20K Β· π) - Automatically decrypt encryptions without knowing the key or cipher,..
MIT - fastNLP (π₯27 Β· β 3.1K Β· π) - fastNLP: A Modularized and Extensible NLP Framework. Currently..
Apache-2 - polyglot (π₯27 Β· β 2.3K Β· π) - Multilingual text (NLP) processing toolkit.
βοΈGPL-3.0 - flashtext (π₯26 Β· β 5.7K Β· π) - Extract Keywords from sentence or Replace keywords in sentences.
MIT - langid (π₯26 Β· β 2.4K Β· π) - Stand-alone language identification system.
BSD-3 - pySBD (π₯26 Β· β 880 Β· π) - pySBD (Python Sentence Boundary Disambiguation) is a rule-based sentence..
MIT - neuralcoref (π₯25 Β· β 2.9K Β· π) - Fast Coreference Resolution in spaCy with Neural Networks.
MIT - GluonNLP (π₯25 Β· β 2.6K Β· π) - Toolkit that enables easy text preprocessing, datasets..
Apache-2 - pytorch-nlp (π₯25 Β· β 2.2K Β· π) - Basic Utilities for PyTorch Natural Language Processing..
BSD-3 - whoosh (π₯25 Β· β 640 Β· π) - Pure-Python full-text search library.
βοΈBSD-1-Clause - PyText (π₯24 Β· β 6.3K Β· π) - A natural language modeling framework based on PyTorch.
BSD-3 - textgenrnn (π₯24 Β· β 4.9K Β· π) - Easily train your own text-generating neural network of any..
MIT - OpenPrompt (π₯24 Β· β 4.7K Β· π) - An Open-Source Framework for Prompt-Learning.
Apache-2 - Snips NLU (π₯24 Β· β 3.9K Β· π) - Snips Python library to extract meaning from text.
Apache-2 - MatchZoo (π₯24 Β· β 3.9K Β· π) - Facilitating the design, comparison and sharing of deep..
Apache-2 - promptsource (π₯24 Β· β 3K Β· π) - Toolkit for creating, sharing and using natural language..
Apache-2 - YouTokenToMe (π₯24 Β· β 970 Β· π) - Unsupervised text tokenizer focused on computational efficiency.
MIT - Kashgari (π₯23 Β· β 2.4K Β· π) - Kashgari is a production-level NLP Transfer learning..
Apache-2 - FARM (π₯23 Β· β 1.8K Β· π) - Fast & easy transfer learning for NLP. Harvesting language..
Apache-2 - gpt-2-simple (π₯22 Β· β 3.4K Β· π) - Python package to easily retrain OpenAIs GPT-2 text-..
MIT - Texar (π₯22 Β· β 2.4K Β· π) - Toolkit for Machine Learning, Natural Language Processing, and..
Apache-2 - jiant (π₯22 Β· β 1.7K Β· π) - jiant is an nlp toolkit.
MIT - stop-words (π₯22 Β· β 160) - Get list of common stop words in various languages in Python.
BSD-3 - NLP Architect (π₯21 Β· β 2.9K Β· π) - A model library for exploring state-of-the-art deep..
Apache-2 - Texthero (π₯21 Β· β 2.9K Β· π) - Text preprocessing, representation and visualization from zero to..
MIT - anaGo (π₯21 Β· β 1.5K Β· π) - Bidirectional LSTM-CRF and ELMo for Named-Entity Recognition,..
MIT - lightseq (π₯20 Β· β 3.3K Β· π) - LightSeq: A High Performance Library for Sequence Processing..
Apache-2 - fast-bert (π₯20 Β· β 1.9K Β· π) - Super easy library for BERT based NLP models.
Apache-2 - DELTA (π₯20 Β· β 1.6K Β· π) - DELTA is a deep learning based natural language and speech..
Apache-2 - textpipe (π₯20 Β· β 300 Β· π) - Textpipe: clean and extract metadata from text.
MIT - numerizer (π₯19 Β· β 230 Β· π) - A Python module to convert natural language numerics into ints and..
MIT - pyfasttext (π₯19 Β· β 230 Β· π) - Yet another Python binding for fastText.
βοΈGPL-3.0 - DeepMatcher (π₯18 Β· β 5.2K Β· π) - Python package for performing Entity and Text Matching using..
BSD-3 - nboost (π₯18 Β· β 670 Β· π) - NBoost is a scalable, search-api-boosting platform for deploying..
Apache-2 - fastT5 (π₯18 Β· β 590 Β· π) - boost inference speed of T5 models by 5x & reduce the model size..
Apache-2 - Camphr (π₯18 Β· β 340 Β· π) - Camphr - NLP libary for creating pipeline components.
Apache-2spacy - NeuroNER (π₯17 Β· β 1.7K Β· π) - Named-entity recognition using neural networks. Easy-to-use and..
MIT - OpenNRE (π₯16 Β· β 4.4K Β· π) - An Open-Source Package for Neural Relation Extraction (NRE).
MIT - BLINK (π₯15 Β· β 1.2K Β· π) - Entity Linker solution.
MIT - TextBox (π₯15 Β· β 1.1K Β· π) - TextBox 2.0 is a text generation library with pre-trained language..
MIT - Translate (π₯15 Β· β 830 Β· π) - Translate - a PyTorch Language Library.
BSD-3 - skift (π₯15 Β· β 240 Β· π) - scikit-learn wrappers for Python fastText.
MIT - ONNX-T5 (π₯14 Β· β 260 Β· π) - Summarization, translation, sentiment-analysis, text-generation..
Apache-2 - NeuralQA (π₯14 Β· β 230 Β· π) - NeuralQA: A Usable Library for Question Answering on Large Datasets..
MIT - TransferNLP (π₯13 Β· β 290 Β· π) - NLP library designed for reproducible experimentation..
MIT - Headliner (π₯13 Β· β 230 Β· π) - Easy training and deployment of seq2seq models.
MIT - textvec (π₯12 Β· β 200 Β· π) - Text vectorization tool to outperform TFIDF for classification..
MIT - spacy-dbpedia-spotlight (π₯12 Β· β 110 Β· π) - A spaCy wrapper for DBpedia Spotlight.
MITspacy
Image Data
Libraries for image & video processing, manipulation, and augmentation as well as libraries for computer vision tasks such as facial recognition, object detection, and classification.
PyTorch Image Models (π₯42 Β· β 36K) - The largest collection of PyTorch image encoders /.. Apache-2 
torchvision (π₯42 Β· β 17K) - Datasets, Transforms and Models specific to Computer Vision. BSD-3 
deepface (π₯38 Β· β 21K Β· π) - A Lightweight Face Recognition and Facial Attribute Analysis (Age,.. MIT
InsightFace (π₯37 Β· β 27K) - State-of-the-art 2D and 3D Face Analysis Project. MIT 
Albumentations (π₯36 Β· β 15K) - Fast and flexible image augmentation library. Paper about.. MIT 
-
GitHub (π¨βπ» 170 Β· π 1.7K Β· π 1.5K - 14% open Β· β±οΈ 25.06.2025):
git clone https://github.com/albumentations-team/albumentations -
PyPi (π₯ 4.6M / month Β· π¦ 730 Β· β±οΈ 27.05.2025):
pip install albumentations -
Conda (π₯ 340K Β· β±οΈ 28.05.2025):
conda install -c conda-forge albumentations
opencv-python (π₯36 Β· β 5.1K) - Automated CI toolchain to produce precompiled opencv-python,.. MIT
detectron2 (π₯34 Β· β 34K) - Detectron2 is a platform for object detection, segmentation.. Apache-2 
vit-pytorch (π₯31 Β· β 24K) - Implementation of Vision Transformer, a simple way to achieve.. MIT 
PaddleSeg (π₯31 Β· β 9.2K) - Easy-to-use image segmentation library with awesome pre-.. Apache-2 
sahi (π₯31 Β· β 4.9K) - Framework agnostic sliced/tiled inference + interactive ui + error analysis.. MIT
PaddleDetection (π₯28 Β· β 14K) - Object Detection toolkit based on PaddlePaddle. It.. Apache-2 
mtcnn (π₯27 Β· β 2.4K Β· π€) - MTCNN face detection implementation for TensorFlow, as a PIP.. MIT 
CellProfiler (π₯27 Β· β 1.1K) - An open-source application for biological image analysis. BSD-3
Image Deduplicator (π₯26 Β· β 5.5K) - Finding duplicate images made easy!. Apache-2 
tensorflow-graphics (π₯26 Β· β 2.8K Β· π€) - TensorFlow Graphics: Differentiable Graphics Layers.. Apache-2 
Norfair (π₯26 Β· β 2.5K) - Lightweight Python library for adding real-time multi-object tracking.. BSD-3
pytorchvideo (π₯25 Β· β 3.5K) - A deep learning library for video understanding research. Apache-2 
MMF (π₯24 Β· β 5.6K) - A modular framework for vision & language multimodal research from.. BSD-3 
kubric (π₯22 Β· β 2.6K) - A data generation pipeline for creating semi-realistic synthetic.. Apache-2
icevision (π₯22 Β· β 870 Β· π€) - An Agnostic Computer Vision Framework - Pluggable to any.. Apache-2
PySlowFast (π₯21 Β· β 7.2K) - PySlowFast: video understanding codebase from FAIR for.. Apache-2 
Image Super-Resolution (π₯21 Β· β 4.8K Β· π€) - Super-scale your images and run experiments with.. Apache-2 
-
GitHub (π¨βπ» 11 Β· π 760 Β· π 220 - 48% open Β· β±οΈ 18.12.2024):
git clone https://github.com/idealo/image-super-resolution -
PyPi (π₯ 3.9K / month Β· π¦ 5 Β· β±οΈ 08.01.2020):
-
Docker Hub (π₯ 290 Β· β 1 Β· β±οΈ 01.04.2019):
docker pull idealo/image-super-resolution-gpu
Caer (π₯21 Β· β 800) - A lightweight Computer Vision library. Scale your models, not boilerplate. MIT
scenic (π₯16 Β· β 3.7K) - Scenic: A Jax Library for Computer Vision Research and Beyond. Apache-2 
-
GitHub (π¨βπ» 95 Β· π 460 Β· π 400 - 70% open Β· β±οΈ 06.08.2025):
git clone https://github.com/google-research/scenic
Show 30 hidden projects...
Graph Data
Libraries for graph processing, clustering, embedding, and machine learning tasks.
PyTorch Geometric (π₯41 Β· β 23K) - Graph Neural Network Library for PyTorch. MIT 
-
GitHub (π¨βπ» 560 Β· π 3.9K Β· π¦ 11K Β· π 4K - 30% open Β· β±οΈ 29.10.2025):
git clone https://github.com/pyg-team/pytorch_geometric -
PyPi (π₯ 940K / month Β· π¦ 730 Β· β±οΈ 15.10.2025):
pip install torch-geometric -
Conda (π₯ 190K Β· β±οΈ 16.10.2025):
conda install -c conda-forge pytorch_geometric
dgl (π₯36 Β· β 14K) - Python package built to ease deep learning on graph, on top of existing DL.. Apache-2
pygraphistry (π₯29 Β· β 2.4K) - PyGraphistry is a Python library to quickly load, shape,.. BSD-3 
ogb (π₯29 Β· β 2K) - Benchmark datasets, data loaders, and evaluators for graph machine learning. MIT
PyKEEN (π₯28 Β· β 1.9K) - A Python library for learning and evaluating knowledge graph embeddings. MIT
pytorch_geometric_temporal (π₯27 Β· β 2.9K) - PyTorch Geometric Temporal: Spatiotemporal Signal.. MIT 
torch-cluster (π₯24 Β· β 900) - PyTorch Extension Library of Optimized Graph Cluster.. MIT 
-
GitHub (π¨βπ» 40 Β· π 150 Β· π 190 - 16% open Β· β±οΈ 12.08.2025):
git clone https://github.com/rusty1s/pytorch_cluster -
PyPi (π₯ 34K / month Β· π¦ 62 Β· β±οΈ 12.10.2023):
pip install torch-cluster -
Conda (π₯ 440K Β· β±οΈ 22.09.2025):
conda install -c conda-forge pytorch_cluster
Show 28 hidden projects...
Audio Data
Libraries for audio analysis, manipulation, transformation, and extraction, as well as speech recognition and music generation tasks.
speechbrain (π₯38 Β· β 11K) - A PyTorch-based Speech Toolkit. Apache-2 
torchaudio (π₯37 Β· β 2.8K) - Data manipulation and transformation for audio signal.. BSD-2 
SpeechRecognition (π₯34 Β· β 8.9K) - Speech recognition module for Python, supporting several.. BSD-3
-
GitHub (π¨βπ» 56 Β· π 2.4K Β· π¦ 21 Β· π 670 - 48% open Β· β±οΈ 28.10.2025):
git clone https://github.com/Uberi/speech_recognition -
PyPi (π₯ 2.2M / month Β· π¦ 730 Β· β±οΈ 12.05.2025):
pip install SpeechRecognition -
Conda (π₯ 360K Β· β±οΈ 12.05.2025):
conda install -c conda-forge speechrecognition
DeepSpeech (π₯33 Β· β 27K Β· π) - DeepSpeech is an open source embedded (offline, on-.. MPL-2.0 
audioread (π₯33 Β· β 520 Β· π) - cross-library (GStreamer + Core Audio + MAD + FFmpeg) audio.. MIT
audiomentations (π₯32 Β· β 2.2K) - A Python library for audio data augmentation. Useful for.. MIT
pyAudioAnalysis (π₯28 Β· β 6.2K) - Python Audio Analysis Library: Feature Extraction,.. Apache-2
python-soundfile (π₯27 Β· β 800) - SoundFile is an audio library based on libsndfile, CFFI, and.. BSD-3
Show 11 hidden projects...
Geospatial Data
Libraries to load, process, analyze, and write geographic data as well as libraries for spatial analysis, map visualization, and geocoding.
pydeck (π₯43 Β· β 14K) - WebGL2 powered visualization framework. MIT 
-
GitHub (π¨βπ» 310 Β· π 2.2K Β· π¦ 9.2K Β· π 3.3K - 13% open Β· β±οΈ 29.10.2025):
git clone https://github.com/visgl/deck.gl -
PyPi (π₯ 16M / month Β· π¦ 160 Β· β±οΈ 21.03.2025):
-
Conda (π₯ 850K Β· β±οΈ 22.04.2025):
conda install -c conda-forge pydeck -
npm (π₯ 750K / month Β· π¦ 360 Β· β±οΈ 16.10.2025):
pyproj (π₯37 Β· β 1.2K) - Python interface to PROJ (cartographic projections and coordinate.. MIT
ArcGIS API (π₯36 Β· β 2.1K) - Documentation and samples for ArcGIS API for Python. Apache-2
ipyleaflet (π₯33 Β· β 1.5K) - A Jupyter - Leaflet.js bridge. MIT 
-
GitHub (π¨βπ» 94 Β· π 360 Β· π¦ 18K Β· π 660 - 44% open Β· β±οΈ 19.06.2025):
git clone https://github.com/jupyter-widgets/ipyleaflet -
PyPi (π₯ 230K / month Β· π¦ 340 Β· β±οΈ 13.06.2025):
-
Conda (π₯ 1.8M Β· β±οΈ 13.06.2025):
conda install -c conda-forge ipyleaflet -
npm (π₯ 2.7K / month Β· π¦ 9 Β· β±οΈ 13.06.2025):
npm install jupyter-leaflet
EarthPy (π₯28 Β· β 530) - A package built to support working with spatial data using open source.. BSD-3
pymap3d (π₯25 Β· β 430) - pure-Python (Numpy optional) 3D coordinate conversions for geospace ecef.. BSD-2
Mapbox GL (π₯22 Β· β 680 Β· π€) - Use Mapbox GL JS to visualize data in a Python Jupyter notebook. MIT 
Show 7 hidden projects...
- Satpy (π₯34 Β· β 1.1K) - Python package for earth-observing satellite data processing.
βοΈGPL-3.0 - geopy (π₯32 Β· β 4.7K Β· π) - Geocoding library for Python.
MIT - Geocoder (π₯32 Β· β 1.6K Β· π) - Python Geocoder.
MIT - prettymaps (π₯24 Β· β 12K) - Draw pretty maps from OpenStreetMap data! Built with osmnx..
βοΈAGPL-3.0 - Sentinelsat (π₯24 Β· β 1K Β· π) - Search and download Copernicus Sentinel satellite images.
βοΈGPL-3.0 - gmaps (π₯22 Β· β 760 Β· π) - Google maps for Jupyter notebooks.
BSD-3 - geoplotlib (π₯21 Β· β 1K Β· π) - python toolbox for visualizing geographical data and making maps.
MIT
Financial Data
Libraries for algorithmic stock/crypto trading, risk analytics, backtesting, technical analysis, and other tasks on financial data.
Qlib (π₯32 Β· β 33K) - Qlib is an AI-oriented Quant investment platform that aims to use AI tech.. MIT 
Alpha Vantage (π₯27 Β· β 4.6K) - A python wrapper for Alpha Vantage API for financial data. MIT
-
GitHub (π¨βπ» 44 Β· π 760 Β· π 290 - 0% open Β· β±οΈ 27.07.2025):
git clone https://github.com/RomelTorres/alpha_vantage -
PyPi (π₯ 140K / month Β· π¦ 35 Β· β±οΈ 18.07.2024):
pip install alpha_vantage -
Conda (π₯ 10K Β· β±οΈ 22.04.2025):
conda install -c conda-forge alpha_vantage
stockstats (π₯26 Β· β 1.4K) - Supply a wrapper ``StockDataFrame`` based on the.. BSD-3
tf-quant-finance (π₯21 Β· β 5K Β· π€) - High-performance TensorFlow library for quantitative.. Apache-2 
finmarketpy (π₯21 Β· β 3.7K Β· π€) - Python library for backtesting trading strategies &.. Apache-2
Show 17 hidden projects...
- arch (π₯33 Β· β 1.5K) - ARCH models in Python.
βUnlicensed - zipline (π₯32 Β· β 19K Β· π) - Zipline, a Pythonic Algorithmic Trading Library.
Apache-2 - ta (π₯32 Β· β 4.8K Β· π) - Technical Analysis Library using Pandas and Numpy.
MIT - pyfolio (π₯31 Β· β 6.1K Β· π) - Portfolio and risk analytics in Python.
Apache-2 - backtrader (π₯29 Β· β 19K Β· π) - Python Backtesting library for trading strategies.
βοΈGPL-3.0 - IB-insync (π₯28 Β· β 3.1K Β· π) - Python sync/async framework for Interactive Brokers API.
BSD-2 - Alphalens (π₯27 Β· β 4K Β· π) - Performance analysis of predictive (alpha) stock factors.
Apache-2 - Enigma Catalyst (π₯27 Β· β 2.5K Β· π) - An Algorithmic Trading Library for Crypto-Assets in..
Apache-2 - empyrical (π₯27 Β· β 1.4K Β· π) - Common financial risk and performance metrics. Used by..
Apache-2 - Backtesting.py (π₯26 Β· β 7.4K) - Backtest trading strategies in Python.
βοΈAGPL-3.0 - TensorTrade (π₯26 Β· β 5.6K Β· π) - An open source reinforcement learning framework for..
Apache-2 - PyAlgoTrade (π₯25 Β· β 4.6K Β· π) - Python Algorithmic Trading Library.
Apache-2 - FinTA (π₯24 Β· β 2.2K Β· π) - Common financial technical indicators implemented in Pandas.
βοΈLGPL-3.0 - Crypto Signals (π₯22 Β· β 5.4K Β· π) - Github.com/CryptoSignal - Trading & Technical Analysis Bot -..
MIT - FinQuant (π₯22 Β· β 1.6K Β· π) - A program for financial portfolio management, analysis and..
MIT - surpriver (π₯12 Β· β 1.8K Β· π) - Find big moving stocks before they move using machine..
βοΈGPL-3.0 - pyrtfolio (π₯9 Β· β 150 Β· π) - Python package to generate stock portfolios.
βοΈGPL-3.0
Time Series Data
Libraries for forecasting, anomaly detection, feature extraction, and machine learning on time-series and sequential data.
sktime (π₯41 Β· β 9.3K) - A unified framework for machine learning with time series. BSD-3 
-
GitHub (π¨βπ» 520 Β· π 1.7K Β· π₯ 110 Β· π¦ 4.7K Β· π 3.1K - 39% open Β· β±οΈ 28.10.2025):
git clone https://github.com/alan-turing-institute/sktime -
PyPi (π₯ 1M / month Β· π¦ 160 Β· β±οΈ 25.09.2025):
-
Conda (π₯ 1.2M Β· β±οΈ 18.09.2025):
conda install -c conda-forge sktime-all-extras
Prophet (π₯34 Β· β 20K) - Tool for producing high quality forecasts for time series data that has.. MIT
StatsForecast (π₯34 Β· β 4.6K) - Lightning fast forecasting with statistical and econometric.. Apache-2
-
GitHub (π¨βπ» 56 Β· π 340 Β· π¦ 2K Β· π 400 - 34% open Β· β±οΈ 29.10.2025):
git clone https://github.com/Nixtla/statsforecast -
PyPi (π₯ 990K / month Β· π¦ 91 Β· β±οΈ 29.10.2025):
pip install statsforecast -
Conda (π₯ 220K Β· β±οΈ 30.10.2025):
conda install -c conda-forge statsforecast
tslearn (π₯33 Β· β 3.1K) - The machine learning toolkit for time series analysis in Python. BSD-2 
skforecast (π₯33 Β· β 1.4K) - Time series forecasting with machine learning models. BSD-3 
Darts (π₯32 Β· β 9K) - A python library for user-friendly forecasting and anomaly detection on.. Apache-2
-
GitHub (π¨βπ» 140 Β· π 970 Β· π 1.8K - 13% open Β· β±οΈ 26.10.2025):
git clone https://github.com/unit8co/darts -
PyPi (π₯ 86K / month Β· π¦ 10 Β· β±οΈ 03.10.2025):
-
Conda (π₯ 94K Β· β±οΈ 05.10.2025):
conda install -c conda-forge u8darts-all -
Docker Hub (π₯ 2.1K Β· β±οΈ 03.10.2025):
pytorch-forecasting (π₯32 Β· β 4.6K) - Time series forecasting with PyTorch. MIT
-
GitHub (π¨βπ» 79 Β· π 710 Β· π¦ 670 Β· π 920 - 59% open Β· β±οΈ 19.10.2025):
git clone https://github.com/jdb78/pytorch-forecasting -
PyPi (π₯ 270K / month Β· π¦ 27 Β· β±οΈ 10.10.2025):
pip install pytorch-forecasting -
Conda (π₯ 87K Β· β±οΈ 05.07.2025):
conda install -c conda-forge pytorch-forecasting
pmdarima (π₯32 Β· β 1.7K Β· π€) - A statistical library designed to fill the void in Pythons time.. MIT
STUMPY (π₯30 Β· β 4K) - STUMPY is a powerful and scalable Python library for modern time series.. BSD-3
NeuralForecast (π₯30 Β· β 3.8K) - Scalable and user friendly neural forecasting algorithms. Apache-2
Show 13 hidden projects...
- NeuralProphet (π₯26 Β· β 4.2K Β· π) - NeuralProphet: A simple forecasting package.
MIT - PyFlux (π₯25 Β· β 2.1K Β· π) - Open source time series library for Python.
BSD-3 - luminol (π₯22 Β· β 1.2K Β· π) - Anomaly Detection and Correlation library.
Apache-2 - ADTK (π₯22 Β· β 1.2K Β· π) - A Python toolkit for rule-based/unsupervised anomaly detection in..
MPL-2.0 - seglearn (π₯21 Β· β 580 Β· π) - Python module for machine learning time series:.
BSD-3 - pydlm (π₯21 Β· β 480 Β· π) - A python library for Bayesian time series modeling.
BSD-3 - tick (π₯20 Β· β 520 Β· π) - Module for statistical learning, with a particular emphasis on time-..
BSD-3 - matrixprofile-ts (π₯19 Β· β 740 Β· π) - A Python library for detecting patterns and anomalies..
Apache-2 - tsflex (π₯19 Β· β 430 Β· π) - Flexible time series feature extraction & processing.
MIT - Auto TS (π₯17 Β· β 760 Β· π) - Automatically build ARIMA, SARIMAX, VAR, FB Prophet and XGBoost..
Apache-2 - tsaug (π₯15 Β· β 360 Β· π) - A Python package for time series augmentation.
Apache-2 - atspy (π₯14 Β· β 520 Β· π) - AtsPy: Automated Time Series Models in Python (by @firmai).
MIT - tslumen (π₯8 Β· β 71 Β· π) - A library for Time Series EDA (exploratory data analysis).
Apache-2
Medical Data
Libraries for processing and analyzing medical data such as MRIs, EEGs, genomic data, and other medical imaging formats.
MNE (π₯37 Β· β 3.1K) - MNE: Magnetoencephalography (MEG) and Electroencephalography (EEG) in Python. BSD-3
NiBabel (π₯34 Β· β 740) - Python package to access a cacophony of neuro-imaging file formats. MIT
DeepVariant (π₯27 Β· β 3.5K) - DeepVariant is an analysis pipeline that uses a deep neural.. BSD-3 
Brainiak (π₯19 Β· β 360 Β· π€) - Brain Imaging Analysis Kit. Apache-2
-
GitHub (π¨βπ» 35 Β· π 140 Β· π 230 - 38% open Β· β±οΈ 06.01.2025):
git clone https://github.com/brainiak/brainiak -
PyPi (π₯ 1.3K / month Β· β±οΈ 07.01.2025):
-
Docker Hub (π₯ 2K Β· β 1 Β· β±οΈ 07.01.2025):
docker pull brainiak/brainiak
Show 10 hidden projects...
Tabular Data
Libraries for processing tabular and structured data.
pytorch_tabular (π₯23 Β· β 1.6K) - A standard framework for modelling Deep Learning Models.. MIT 
upgini (π₯21 Β· β 350) - Data search & enrichment library for Machine Learning Easily find and add.. BSD-3
Show 3 hidden projects...
- miceforest (π₯21 Β· β 390) - Multiple Imputation with LightGBM in Python.
βUnlicensed - carefree-learn (π₯18 Β· β 410 Β· π) - Deep Learning PyTorch.
MIT - deltapy (π₯13 Β· β 550 Β· π) - DeltaPy - Tabular Data Augmentation (by @firmai).
MIT
Optical Character Recognition
Libraries for optical character recognition (OCR) and text extraction from images or videos.
PaddleOCR (π₯44 Β· β 62K) - Turn any PDF or image document into structured data for your.. Apache-2 
OCRmyPDF (π₯37 Β· β 32K) - OCRmyPDF adds an OCR text layer to scanned PDF files, allowing them.. MPL-2.0
Tesseract (π₯32 Β· β 6.2K Β· π€) - Python-tesseract is an optical character recognition (OCR).. Apache-2
MMOCR (π₯27 Β· β 4.7K Β· π€) - OpenMMLab Text Detection, Recognition and Understanding Toolbox. Apache-2 
keras-ocr (π₯25 Β· β 1.5K) - A packaged and flexible version of the CRAFT text detector and.. MIT 
Show 6 hidden projects...
Data Containers & Structures
General-purpose data containers & structures as well as utilities & extensions for pandas.
π best-of-python - Data Containers ( β 4.2K) - Collection of data-container, dataframe, and pandas-..
Data Loading & Extraction
Libraries for loading, collecting, and extracting data from a variety of data sources and formats.
π best-of-python - Data Extraction ( β 4.2K) - Collection of data-loading and -extraction libraries.
Web Scraping & Crawling
Libraries for web scraping, crawling, downloading, and mining as well as libraries.
π best-of-web-python - Web Scraping ( β 2.6K) - Collection of web-scraping and crawling libraries.
Data Pipelines & Streaming
Libraries for data batch- and stream-processing, workflow automation, job scheduling, and other data pipeline tasks.
π best-of-python - Data Pipelines ( β 4.2K) - Libraries for data batch- and stream-processing,..
Show 1 hidden projects...
Distributed Machine Learning
Libraries that provide capabilities to distribute and parallelize machine learning tasks across large-scale compute infrastructure.
Ray (π₯48 Β· β 40K) - Ray is an AI compute engine. Ray consists of a core distributed runtime.. Apache-2
DeepSpeed (π₯41 Β· β 41K) - DeepSpeed is a deep learning optimization library that makes.. Apache-2 
-
GitHub (π¨βπ» 420 Β· π 4.6K Β· π¦ 15K Β· π 3.2K - 34% open Β· β±οΈ 29.10.2025):
git clone https://github.com/microsoft/DeepSpeed -
PyPi (π₯ 990K / month Β· π¦ 350 Β· β±οΈ 23.10.2025):
-
Docker Hub (π₯ 24K Β· β 4 Β· β±οΈ 02.09.2022):
docker pull deepspeed/deepspeed
dask.distributed (π₯39 Β· β 1.7K) - A distributed task scheduler for Dask. BSD-3
horovod (π₯36 Β· β 15K) - Distributed training framework for TensorFlow, Keras, PyTorch, and.. Apache-2
metrics (π₯36 Β· β 2.3K) - Machine learning metrics for distributed, scalable PyTorch.. Apache-2 
H2O-3 (π₯34 Β· β 7.3K) - H2O is an Open Source, Distributed, Fast & Scalable Machine Learning.. Apache-2
ColossalAI (π₯33 Β· β 41K) - Making large AI models cheaper, faster and more accessible. Apache-2
-
GitHub (π¨βπ» 200 Β· π 4.5K Β· π¦ 530 Β· π 1.8K - 26% open Β· β±οΈ 26.09.2025):
git clone https://github.com/hpcaitech/colossalai
FairScale (π₯31 Β· β 3.4K) - PyTorch extensions for high performance and large scale training. BSD-3 
BigDL (π₯30 Β· β 8.4K) - Accelerate local LLM inference and finetuning (LLaMA, Mistral,.. Apache-2
-
GitHub (π¨βπ» 120 Β· π 1.4K Β· π₯ 710 Β· π 3K - 40% open Β· β±οΈ 14.10.2025):
git clone https://github.com/intel-analytics/BigDL -
PyPi (π₯ 15K / month Β· π¦ 2 Β· β±οΈ 24.03.2024):
-
Maven (π¦ 5 Β· β±οΈ 20.04.2021):
<dependency> <groupId>com.intel.analytics.bigdl</groupId> <artifactId>bigdl-SPARK_2.4</artifactId> <version>[VERSION]</version> </dependency>
petastorm (π₯29 Β· β 1.9K) - Petastorm library enables single machine or distributed training.. Apache-2
Hivemind (π₯26 Β· β 2.3K) - Decentralized deep learning in PyTorch. Built to train models on.. MIT
Apache Singa (π₯23 Β· β 3.6K Β· π€) - a distributed deep learning platform. Apache-2
-
GitHub (π¨βπ» 98 Β· π 1.3K Β· π¦ 6 Β· π 140 - 35% open Β· β±οΈ 26.03.2025):
git clone https://github.com/apache/singa -
Conda (π₯ 1.2K Β· β±οΈ 25.03.2025):
conda install -c nusdbsystem singa -
Docker Hub (π₯ 9.9K Β· β 3 Β· β±οΈ 31.05.2022):
analytics-zoo (π₯22 Β· β 2.6K) - Distributed Tensorflow, Keras and PyTorch on Apache.. Apache-2 
Show 17 hidden projects...
Hyperparameter Optimization & AutoML
Libraries for hyperparameter optimization, automl and neural architecture search.
AutoGluon (π₯35 Β· β 9.6K) - Fast and Accurate ML in 3 Lines of Code. Apache-2 

-
GitHub (π¨βπ» 140 Β· π 1.1K Β· π¦ 1.2K Β· π 1.8K - 24% open Β· β±οΈ 29.10.2025):
git clone https://github.com/autogluon/autogluon -
PyPi (π₯ 260K / month Β· π¦ 38 Β· β±οΈ 23.10.2025):
-
Conda (π₯ 45K Β· β±οΈ 30.07.2025):
conda install -c conda-forge autogluon -
Docker Hub (π₯ 20K Β· β 19 Β· β±οΈ 16.06.2025):
docker pull autogluon/autogluon
Bayesian Optimization (π₯34 Β· β 8.4K) - A Python implementation of global optimization with.. MIT
Hyperopt (π₯34 Β· β 7.5K Β· π€) - Distributed Asynchronous Hyperparameter Optimization in Python. BSD-3
featuretools (π₯32 Β· β 7.6K Β· π€) - An open source python library for automated feature.. BSD-3
lazypredict (π₯28 Β· β 3.2K) - Lazy Predict help build a lot of basic models without much code.. MIT 
mljar-supervised (π₯28 Β· β 3.2K) - Python package for AutoML on Tabular Data with Feature.. MIT
-
GitHub (π¨βπ» 30 Β· π 420 Β· π¦ 170 Β· π 680 - 21% open Β· β±οΈ 07.07.2025):
git clone https://github.com/mljar/mljar-supervised -
PyPi (π₯ 8.7K / month Β· π¦ 6 Β· β±οΈ 07.07.2025):
pip install mljar-supervised -
Conda (π₯ 52K Β· β±οΈ 08.07.2025):
conda install -c conda-forge mljar-supervised
Hyperactive (π₯24 Β· β 530) - An optimization and data collection toolbox for convenient and fast.. MIT
Auto ViML (π₯20 Β· β 540 Β· π€) - Automatically Build Multiple ML Models with a Single Line of.. Apache-2
featurewiz (π₯18 Β· β 670 Β· π€) - Use advanced feature engineering strategies and select best.. Apache-2
Show 36 hidden projects...
Reinforcement Learning
Libraries for building and evaluating reinforcement learning & agent-based systems.
Dopamine (π₯27 Β· β 11K Β· π€) - Dopamine is a research framework for fast prototyping of.. Apache-2 
TF-Agents (π₯27 Β· β 3K) - TF-Agents: A reliable, scalable and easy to use TensorFlow.. Apache-2 
PARL (π₯24 Β· β 3.4K) - A high-performance distributed training framework for Reinforcement.. Apache-2 
Show 15 hidden projects...
Recommender Systems
Libraries for building and evaluating recommendation systems.
Recommenders (π₯33 Β· β 21K) - Best Practices on Recommendation Systems. MIT
RecBole (π₯25 Β· β 4.1K Β· π€) - A unified, comprehensive and efficient recommendation library. MIT 
TF Recommenders (π₯25 Β· β 2K) - TensorFlow Recommenders is a library for building.. Apache-2 
Show 11 hidden projects...
Privacy Machine Learning
Libraries for encrypted and privacy-preserving machine learning using methods like federated learning & differential privacy.
PySyft (π₯31 Β· β 9.8K) - Perform data science on data that remains in someone elses server. Apache-2 
TensorFlow Privacy (π₯24 Β· β 2K) - Library for training machine learning models with.. Apache-2 
Show 2 hidden projects...
- TFEncrypted (π₯24 Β· β 1.2K Β· π) - A Framework for Encrypted Machine Learning in..
Apache-2 - PipelineDP (π₯19 Β· β 280) - PipelineDP is a Python framework for applying differentially..
Apache-2
Workflow & Experiment Tracking
Libraries to organize, track, and visualize machine learning experiments.
mlflow (π₯47 Β· β 23K) - The open source developer platform to build AI/LLM applications and.. Apache-2
wandb client (π₯44 Β· β 10K) - The AI developer platform. Use Weights & Biases to train and fine-.. MIT
Tensorboard (π₯41 Β· β 7K) - TensorFlows Visualization Toolkit. Apache-2 
SageMaker SDK (π₯41 Β· β 2.2K) - A library for training and deploying machine learning.. Apache-2 

tensorboardX (π₯35 Β· β 8K) - tensorboard for pytorch (and chainer, mxnet, numpy, ...). MIT
PyCaret (π₯34 Β· β 9.6K Β· π€) - An open-source, low-code machine learning library in Python. MIT
ClearML (π₯34 Β· β 6.3K) - ClearML - Auto-Magical CI/CD to streamline your AI workload... Apache-2
-
GitHub (π¨βπ» 100 Β· π 710 Β· π₯ 3.5K Β· π¦ 1.9K Β· π 1.2K - 45% open Β· β±οΈ 27.10.2025):
git clone https://github.com/allegroai/clearml -
PyPi (π₯ 500K / month Β· π¦ 78 Β· β±οΈ 22.10.2025):
-
Docker Hub (π₯ 31K Β· β±οΈ 05.10.2020):
docker pull allegroai/trains
snakemake (π₯34 Β· β 2.6K) - This is the development home of the workflow management system.. MIT
aim (π₯32 Β· β 5.8K) - Aim An easy-to-use & supercharged open-source experiment tracker. Apache-2
AzureML SDK (π₯31 Β· β 4.3K Β· π€) - Python notebooks with ML and deep learning examples with Azure.. MIT
sacred (π₯29 Β· β 4.3K) - Sacred is a tool to help you configure, organize, log and reproduce.. MIT
Neptune.ai (π₯29 Β· β 620) - The experiment tracker for foundation model training. Apache-2
-
GitHub (π¨βπ» 57 Β· π 66 Β· π¦ 920 Β· π 260 - 12% open Β· β±οΈ 09.06.2025):
git clone https://github.com/neptune-ai/neptune-client -
PyPi (π₯ 480K / month Β· π¦ 77 Β· β±οΈ 15.04.2025):
pip install neptune-client -
Conda (π₯ 390K Β· β±οΈ 22.04.2025):
conda install -c conda-forge neptune-client
livelossplot (π₯25 Β· β 1.3K Β· π€) - Live training loss plot in Jupyter Notebook for Keras,.. MIT 
ml-metadata (π₯25 Β· β 660) - For recording and retrieving metadata associated with ML.. Apache-2
Labml (π₯24 Β· β 2.3K) - Monitor deep learning model training and hardware usage from your mobile.. MIT
gokart (π₯24 Β· β 330) - Gokart solves reproducibility, task dependencies, constraints of good code,.. MIT
TensorWatch (π₯22 Β· β 3.5K) - Debugging, monitoring and visualization for Python Machine Learning.. MIT
Show 13 hidden projects...
Model Serialization & Deployment
Libraries to serialize models to files, convert between a variety of model formats, and optimize models for deployment.
huggingface_hub (π₯40 Β· β 3K) - The official Python client for the Hugging Face Hub. Apache-2
-
GitHub (π¨βπ» 280 Β· π 830 Β· π 1.3K - 11% open Β· β±οΈ 30.10.2025):
git clone https://github.com/huggingface/huggingface_hub -
PyPi (π₯ 120M / month Β· π¦ 4.1K Β· β±οΈ 28.10.2025):
pip install huggingface_hub -
Conda (π₯ 4.2M Β· β±οΈ 28.10.2025):
conda install -c conda-forge huggingface_hub
BentoML (π₯36 Β· β 8.2K) - The easiest way to serve AI apps and models - Build Model Inference.. Apache-2
Core ML Tools (π₯35 Β· β 5K) - Core ML tools contain supporting tools for Core ML model.. BSD-3
TorchServe (π₯33 Β· β 4.4K Β· π€) - Serve, optimize and scale PyTorch models in production. Apache-2 
-
GitHub (π¨βπ» 220 Β· π 890 Β· π₯ 8K Β· π¦ 900 Β· π 1.7K - 25% open Β· β±οΈ 17.03.2025):
git clone https://github.com/pytorch/serve -
PyPi (π₯ 97K / month Β· π¦ 26 Β· β±οΈ 30.09.2024):
-
Conda (π₯ 570K Β· β±οΈ 25.03.2025):
conda install -c pytorch torchserve -
Docker Hub (π₯ 1.5M Β· β 32 Β· β±οΈ 30.09.2024):
docker pull pytorch/torchserve
mmdnn (π₯25 Β· β 5.8K) - MMdnn is a set of tools to help users inter-operate among different deep.. MIT
Hummingbird (π₯24 Β· β 3.5K) - Hummingbird compiles trained ML models into tensor computation for.. MIT
-
GitHub (π¨βπ» 40 Β· π 290 Β· π₯ 930 Β· π 330 - 21% open Β· β±οΈ 17.07.2025):
git clone https://github.com/microsoft/hummingbird -
PyPi (π₯ 7.6K / month Β· π¦ 7 Β· β±οΈ 25.10.2024):
pip install hummingbird-ml -
Conda (π₯ 64K Β· β±οΈ 22.04.2025):
conda install -c conda-forge hummingbird-ml
tfdeploy (π₯15 Β· β 360 Β· π€) - Deploy tensorflow graphs for fast evaluation and export to.. BSD-3 
Show 10 hidden projects...
Model Interpretability
Libraries to visualize, explain, debug, evaluate, and interpret machine learning models.
shap (π₯42 Β· β 25K) - A game theoretic approach to explain the output of any machine learning model. MIT
Netron (π₯36 Β· β 32K) - Visualizer for neural network, deep learning and machine learning.. MIT 

evaluate (π₯34 Β· β 2.4K) - Evaluate: A library for easily evaluating machine learning models.. Apache-2
InterpretML (π₯33 Β· β 6.7K) - Fit interpretable models. Explain blackbox machine learning. MIT 
DoWhy (π₯30 Β· β 7.8K) - DoWhy is a Python library for causal inference that supports explicit.. MIT
shapash (π₯30 Β· β 3K) - Shapash: User-friendly Explainability and Interpretability to.. Apache-2 
explainerdashboard (π₯30 Β· β 2.5K) - Quickly build Explainable AI dashboards that show the inner.. MIT
-
GitHub (π¨βπ» 23 Β· π 340 Β· π¦ 650 Β· π 240 - 16% open Β· β±οΈ 01.08.2025):
git clone https://github.com/oegedijk/explainerdashboard -
PyPi (π₯ 42K / month Β· π¦ 15 Β· β±οΈ 03.06.2025):
pip install explainerdashboard -
Conda (π₯ 75K Β· β±οΈ 04.06.2025):
conda install -c conda-forge explainerdashboard
dtreeviz (π₯28 Β· β 3.1K Β· π€) - A python library for decision tree visualization and model.. MIT
Model Analysis (π₯27 Β· β 1.3K) - Model analysis tools for TensorFlow. Apache-2 

Fairness 360 (π₯26 Β· β 2.7K) - A comprehensive set of fairness metrics for datasets and.. Apache-2
imodels (π₯26 Β· β 1.5K) - Interpretable ML package for concise, transparent, and accurate.. MIT
LIT (π₯25 Β· β 3.6K Β· π€) - The Learning Interpretability Tool: Interactively analyze ML models.. Apache-2
responsible-ai-widgets (π₯25 Β· β 1.6K Β· π€) - Responsible AI Toolbox is a suite of tools providing.. MIT 


Explainability 360 (π₯24 Β· β 1.7K Β· π€) - Interpretability and explainability of data and.. Apache-2
random-forest-importances (π₯19 Β· β 620 Β· π€) - Code to compute permutation and drop-column.. MIT 
fairness-indicators (π₯18 Β· β 360) - Tensorflows Fairness Evaluation and Visualization.. Apache-2 

Show 32 hidden projects...
Vector Similarity Search (ANN)
Libraries for Approximate Nearest Neighbor Search and Vector Indexing/Similarity Search.
π ANN Benchmarks ( β 5.5K) - Benchmarks of approximate nearest neighbor libraries in Python.
Milvus (π₯43 Β· β 38K) - Milvus is a high-performance, cloud-native vector database built for.. Apache-2
-
GitHub (π¨βπ» 330 Β· π 3.5K Β· π₯ 290K Β· π 15K - 5% open Β· β±οΈ 30.10.2025):
git clone https://github.com/milvus-io/milvus -
PyPi (π₯ 3.3M / month Β· π¦ 350 Β· β±οΈ 19.09.2025):
-
Docker Hub (π₯ 72M Β· β 90 Β· β±οΈ 30.10.2025):
docker pull milvusdb/milvus
Faiss (π₯42 Β· β 38K Β· π) - A library for efficient similarity search and clustering of dense vectors. MIT
Annoy (π₯35 Β· β 14K) - Approximate Nearest Neighbors in C++/Python optimized for memory usage.. Apache-2
USearch (π₯33 Β· β 3.2K) - Fast Open-Source Search & Clustering engine for Vectors & Arbitrary.. Apache-2
-
GitHub (π¨βπ» 81 Β· π 230 Β· π₯ 110K Β· π¦ 210 Β· π 250 - 32% open Β· β±οΈ 29.10.2025):
git clone https://github.com/unum-cloud/usearch -
PyPi (π₯ 140K / month Β· π¦ 44 Β· β±οΈ 04.09.2025):
-
npm (π₯ 18K / month Β· π¦ 23 Β· β±οΈ 29.10.2025):
-
Docker Hub (π₯ 480 Β· β 1 Β· β±οΈ 29.10.2025):
NMSLIB (π₯32 Β· β 3.5K) - Non-Metric Space Library (NMSLIB): An efficient similarity search.. Apache-2
PyNNDescent (π₯28 Β· β 950) - A Python nearest neighbor descent for approximate nearest neighbors. BSD-2
NGT (π₯22 Β· β 1.3K) - Nearest Neighbor Search with Neighborhood Graph and Tree for High-.. Apache-2
Show 5 hidden projects...
- hnswlib (π₯32 Β· β 5K Β· π) - Header-only C++/python library for fast approximate nearest..
Apache-2 - NearPy (π₯22 Β· β 770 Β· π) - Python framework for fast (approximated) nearest neighbour search in..
MIT - N2 (π₯22 Β· β 580 Β· π) - TOROS N2 - lightweight approximate Nearest Neighbor library which runs..
Apache-2 - Magnitude (π₯20 Β· β 1.7K Β· π) - A fast, efficient universal vector embedding utility package.
MIT - PySparNN (π₯11 Β· β 920 Β· π) - Approximate Nearest Neighbor Search for Sparse Data in Python!.
BSD-3
Probabilistics & Statistics
Libraries providing capabilities for probabilistic programming/reasoning, bayesian inference, gaussian processes, or statistics.
tensorflow-probability (π₯35 Β· β 4.4K) - Probabilistic reasoning and statistical analysis in.. Apache-2 
-
GitHub (π¨βπ» 500 Β· π 1.1K Β· π¦ 4 Β· π 1.5K - 48% open Β· β±οΈ 22.10.2025):
git clone https://github.com/tensorflow/probability -
PyPi (π₯ 880K / month Β· π¦ 620 Β· β±οΈ 08.11.2024):
pip install tensorflow-probability -
Conda (π₯ 200K Β· β±οΈ 22.04.2025):
conda install -c conda-forge tensorflow-probability
GPyTorch (π₯34 Β· β 3.8K) - A highly efficient implementation of Gaussian Processes in PyTorch. MIT 
Pyro (π₯32 Β· β 8.9K) - Deep universal probabilistic programming with Python and PyTorch. Apache-2 
SALib (π₯31 Β· β 960) - Sensitivity Analysis Library in Python. Contains Sobol, Morris, FAST, and.. MIT
hmmlearn (π₯30 Β· β 3.3K Β· π€) - Hidden Markov Models in Python, with scikit-learn like API. BSD-3 
pomegranate (π₯26 Β· β 3.5K Β· π€) - Fast, flexible and easy to use probabilistic modelling in Python. MIT
scikit-posthocs (π₯24 Β· β 380) - Multiple Pairwise Comparisons (Post Hoc) Tests in Python. MIT 
-
GitHub (π¨βπ» 18 Β· π 41 Β· π₯ 67 Β· π¦ 1.2K Β· π 72 - 6% open Β· β±οΈ 11.09.2025):
git clone https://github.com/maximtrp/scikit-posthocs -
PyPi (π₯ 120K / month Β· π¦ 73 Β· β±οΈ 29.03.2025):
pip install scikit-posthocs -
Conda (π₯ 1.1M Β· β±οΈ 22.04.2025):
conda install -c conda-forge scikit-posthocs
Baal (π₯22 Β· β 910) - Bayesian active learning library for research and industrial usecases. Apache-2
Orbit (π₯21 Β· β 2K) - A Python package for Bayesian forecasting with object-oriented design.. Apache-2
TorchUncertainty (π₯20 Β· β 440 Β· π) - Open-source framework for uncertainty and deep.. Apache-2 
Show 6 hidden projects...
Adversarial Robustness
Libraries for testing the robustness of machine learning models against attacks with adversarial/malicious examples.
ART (π₯34 Β· β 5.6K) - Adversarial Robustness Toolbox (ART) - Python Library for Machine Learning.. MIT
-
GitHub (π¨βπ» 140 Β· π 1.2K Β· π¦ 770 Β· π 910 - 1% open Β· β±οΈ 17.10.2025):
git clone https://github.com/Trusted-AI/adversarial-robustness-toolbox -
PyPi (π₯ 29K / month Β· π¦ 25 Β· β±οΈ 07.07.2025):
pip install adversarial-robustness-toolbox -
Conda (π₯ 85K Β· β±οΈ 07.07.2025):
conda install -c conda-forge adversarial-robustness-toolbox
TextAttack (π₯28 Β· β 3.3K) - TextAttack is a Python framework for adversarial attacks, data.. MIT
Show 7 hidden projects...
GPU & Accelerator Utilities
Libraries that require and make use of CUDA/GPU or other accelerator hardware capabilities to optimize machine learning tasks.
optimum (π₯37 Β· β 3.1K) - Accelerate inference and training of Transformers, Diffusers, TIMM.. Apache-2
Apex (π₯32 Β· β 8.8K) - A PyTorch Extension: Tools for easy mixed precision and distributed.. BSD-3 
gpustat (π₯29 Β· β 4.3K) - A simple command-line utility for querying and monitoring GPU status. MIT
CuPy (π₯27 Β· β 11K) - NumPy & SciPy for GPU. MIT
-
GitHub (π¨βπ» 340 Β· π 950):
git clone https://github.com/cupy/cupy -
PyPi (π₯ 39K / month Β· π¦ 400 Β· β±οΈ 18.08.2025):
-
Conda (π₯ 7.2M Β· β±οΈ 14.09.2025):
conda install -c conda-forge cupy -
Docker Hub (π₯ 92K Β· β 14 Β· β±οΈ 18.08.2025):
DALI (π₯25 Β· β 5.5K) - A GPU-accelerated library containing highly optimized building blocks.. Apache-2
-
GitHub (π¨βπ» 99 Β· π 650 Β· π 1.7K - 15% open Β· β±οΈ 30.10.2025):
git clone https://github.com/NVIDIA/DALI
Vulkan Kompute (π₯23 Β· β 2.4K) - General purpose GPU compute framework built on Vulkan to.. Apache-2
Show 9 hidden projects...
Tensorflow Utilities
Libraries that extend TensorFlow with additional capabilities.
TensorFlow Datasets (π₯39 Β· β 4.5K) - TFDS is a collection of datasets ready to use with.. Apache-2 
-
GitHub (π¨βπ» 660 Β· π 1.6K Β· π¦ 25K Β· π 1.5K - 47% open Β· β±οΈ 17.10.2025):
git clone https://github.com/tensorflow/datasets -
PyPi (π₯ 1.8M / month Β· π¦ 340 Β· β±οΈ 28.05.2025):
pip install tensorflow-datasets -
Conda (π₯ 51K Β· β±οΈ 22.04.2025):
conda install -c conda-forge tensorflow-datasets
tensorflow-hub (π₯31 Β· β 3.5K Β· π€) - A library for transfer learning by reusing parts of.. Apache-2 
TFX (π₯31 Β· β 2.2K Β· π€) - TFX is an end-to-end platform for deploying production ML.. Apache-2 
TF Model Optimization (π₯29 Β· β 1.6K) - A toolkit to optimize ML models for deployment for.. Apache-2 
TensorFlow I/O (π₯29 Β· β 730) - Dataset, streaming, and file system extensions.. Apache-2 
TensorFlow Transform (π₯26 Β· β 990) - Input pipeline framework. Apache-2 
Neural Structured Learning (π₯24 Β· β 1K Β· π€) - Training neural models with structured signals. Apache-2 
TensorFlow Cloud (π₯21 Β· β 380) - The TensorFlow Cloud repository provides APIs that.. Apache-2 
TF Compression (π₯20 Β· β 900) - Data compression in TensorFlow. Apache-2 
Show 7 hidden projects...
Jax Utilities
Libraries that extend Jax with additional capabilities.
equinox (π₯33 Β· β 2.6K) - Elegant easy-to-use neural networks + scientific computing in.. Apache-2 
Show 2 hidden projects...
Sklearn Utilities
Libraries that extend scikit-learn with additional capabilities.
scikit-learn-intelex (π₯35 Β· β 1.3K) - Extension for Scikit-learn is a seamless way to speed.. Apache-2 
-
GitHub (π¨βπ» 86 Β· π 180 Β· π¦ 14K Β· π 250 - 15% open Β· β±οΈ 28.10.2025):
git clone https://github.com/intel/scikit-learn-intelex -
PyPi (π₯ 89K / month Β· π¦ 74 Β· β±οΈ 22.10.2025):
pip install scikit-learn-intelex -
Conda (π₯ 650K Β· β±οΈ 30.10.2025):
conda install -c conda-forge scikit-learn-intelex
imbalanced-learn (π₯33 Β· β 7.1K) - A Python Package to Tackle the Curse of Imbalanced.. MIT 
-
GitHub (π¨βπ» 89 Β· π 1.3K Β· π 630 - 8% open Β· β±οΈ 14.08.2025):
git clone https://github.com/scikit-learn-contrib/imbalanced-learn -
PyPi (π₯ 14M / month Β· π¦ 600 Β· β±οΈ 14.08.2025):
pip install imbalanced-learn -
Conda (π₯ 750K Β· β±οΈ 14.08.2025):
conda install -c conda-forge imbalanced-learn
category_encoders (π₯31 Β· β 2.5K Β· π€) - A library of sklearn compatible categorical variable.. BSD-3 
-
GitHub (π¨βπ» 71 Β· π 400 Β· π¦ 4.1K Β· π 300 - 13% open Β· β±οΈ 24.03.2025):
git clone https://github.com/scikit-learn-contrib/category_encoders -
PyPi (π₯ 2.1M / month Β· π¦ 310 Β· β±οΈ 15.03.2025):
pip install category_encoders -
Conda (π₯ 370K Β· β±οΈ 22.04.2025):
conda install -c conda-forge category_encoders
scikit-lego (π₯28 Β· β 1.4K) - Extra blocks for scikit-learn pipelines. MIT 
scikit-opt (π₯26 Β· β 6.2K) - Genetic Algorithm, Particle Swarm Optimization, Simulated.. MIT 
iterative-stratification (π₯21 Β· β 880 Β· π€) - scikit-learn cross validators for iterative.. BSD-3 
scikit-tda (π₯19 Β· β 550) - Topological Data Analysis for Python. MIT 
Show 11 hidden projects...
Pytorch Utilities
Libraries that extend Pytorch with additional capabilities.
accelerate (π₯43 Β· β 9.2K) - A simple way to launch, train, and use PyTorch models on.. Apache-2 
tinygrad (π₯33 Β· β 30K) - You like pytorch? You like micrograd? You love tinygrad!. MIT 
-
GitHub (π¨βπ» 420 Β· π 3.6K Β· π¦ 20 Β· π 1K - 12% open Β· β±οΈ 30.10.2025):
git clone https://github.com/geohot/tinygrad
PML (π₯33 Β· β 6.2K) - The easiest way to use deep metric learning in your application. Modular,.. MIT 
-
GitHub (π¨βπ» 45 Β· π 660 Β· π¦ 2.9K Β· π 530 - 14% open Β· β±οΈ 17.08.2025):
git clone https://github.com/KevinMusgrave/pytorch-metric-learning -
PyPi (π₯ 2.3M / month Β· π¦ 68 Β· β±οΈ 17.08.2025):
pip install pytorch-metric-learning -
Conda (π₯ 13K Β· β±οΈ 25.03.2025):
conda install -c metric-learning pytorch-metric-learning
torchdiffeq (π₯31 Β· β 6.2K) - Differentiable ODE solvers with full GPU support and.. MIT 
torchsde (π₯30 Β· β 1.7K Β· π€) - Differentiable SDE solvers with GPU support and efficient.. Apache-2 
torch-scatter (π₯26 Β· β 1.7K) - PyTorch Extension Library of Optimized Scatter Operations. MIT 
PyTorch Sparse (π₯25 Β· β 1.1K) - PyTorch Extension Library of Optimized Autograd Sparse.. MIT 
Pytorch Toolbelt (π₯24 Β· β 1.6K) - PyTorch extensions for fast R&D prototyping and Kaggle.. MIT 
pytorchviz (π₯14 Β· β 3.4K Β· π€) - A small package to create visualizations of PyTorch execution.. MIT
-
GitHub (π¨βπ» 6 Β· π 280 Β· π 72 - 47% open Β· β±οΈ 30.12.2024):
git clone https://github.com/szagoruyko/pytorchviz
Show 22 hidden projects...
- pretrainedmodels (π₯29 Β· β 9.1K Β· π) - Pretrained ConvNets for pytorch: NASNet, ResNeXt,..
BSD-3 - EfficientNet-PyTorch (π₯28 Β· β 8.2K Β· π) - A PyTorch implementation of EfficientNet.
Apache-2 - lightning-flash (π₯27 Β· β 1.7K Β· π) - Your PyTorch AI Factory - Flash enables you to easily..
Apache-2 - pytorch-optimizer (π₯26 Β· β 3.1K Β· π) - torch-optimizer -- collection of optimizers for..
Apache-2 - TabNet (π₯26 Β· β 2.9K Β· π) - PyTorch implementation of TabNet paper :..
MIT - EfficientNets (π₯25 Β· β 1.6K Β· π) - Pretrained EfficientNet, EfficientNet-Lite, MixNet,..
Apache-2 - pytorch-summary (π₯24 Β· β 4.1K Β· π) - Model summary in PyTorch similar to
model.summary()..MIT - Higher (π₯23 Β· β 1.6K Β· π) - higher is a pytorch library allowing users to obtain higher..
Apache-2 - micrograd (π₯22 Β· β 14K Β· π) - A tiny scalar-valued autograd engine and a neural net library..
MIT - SRU (π₯22 Β· β 2.1K Β· π) - Training RNNs as Fast as CNNs (https://arxiv.org/abs/1709.02755).
MIT - Antialiased CNNs (π₯22 Β· β 1.7K Β· π) - pip install antialiased-cnns to improve stability and..
βοΈCC BY-NC-SA 4.0 - AdaBound (π₯21 Β· β 2.9K Β· π) - An optimizer that trains as fast as Adam and as good as SGD.
Apache-2 - reformer-pytorch (π₯21 Β· β 2.2K Β· π) - Reformer, the efficient Transformer, in Pytorch.
MIT - Torchmeta (π₯21 Β· β 2K Β· π) - A collection of extensions and data-loaders for few-shot..
MIT - Poutyne (π₯21 Β· β 580) - A simplified framework and utilities for PyTorch.
βοΈLGPL-3.0 - Performer Pytorch (π₯19 Β· β 1.2K Β· π) - An implementation of Performer, a linear attention-..
MIT - Torch-Struct (π₯19 Β· β 1.1K Β· π) - Fast, general, and tested differentiable structured..
MIT - Lambda Networks (π₯17 Β· β 1.5K Β· π) - Implementation of LambdaNetworks, a new approach to..
MIT - Pywick (π₯17 Β· β 400 Β· π) - High-level batteries-included neural network training library for..
MIT - TorchDrift (π₯15 Β· β 320 Β· π) - Drift Detection for your PyTorch Models.
Apache-2 - Tez (π₯14 Β· β 1.2K Β· π) - Tez is a super-simple and lightweight Trainer for PyTorch. It..
Apache-2 - Tensor Sensor (π₯14 Β· β 810 Β· π) - The goal of this library is to generate more helpful..
MIT
Database Clients
Libraries for connecting to, operating, and querying databases.
π best-of-python - DB Clients ( β 4.2K) - Collection of database clients for python.
Others
scipy (π₯51 Β· β 14K) - Ecosystem of open-source software for mathematics, science, and engineering. BSD-3
PennyLane (π₯37 Β· β 2.9K) - PennyLane is a cross-platform Python library for quantum.. Apache-2
PyOD (π₯36 Β· β 9.6K) - A Python Library for Outlier and Anomaly Detection, Integrating Classical.. BSD-2
Datasette (π₯35 Β· β 10K) - An open source multi-tool for exploring and publishing data. Apache-2
DeepChem (π₯34 Β· β 6.3K Β· π) - Democratizing Deep-Learning for Drug Discovery, Quantum.. MIT 
agate (π₯34 Β· β 1.2K) - A Python data analysis library that is optimized for humans instead of.. MIT
anomalib (π₯31 Β· β 5.1K Β· π) - An anomaly detection library comprising state-of-the-art.. Apache-2
pyjanitor (π₯31 Β· β 1.5K) - Clean APIs for data cleaning. Python implementation of R package.. MIT
causalml (π₯30 Β· β 5.6K) - Uplift modeling and causal inference with machine learning.. Apache-2
dstack (π₯30 Β· β 1.9K) - dstack is an open-source control plane for running development,.. MPL-2.0
metricflow (π₯29 Β· β 1.3K) - MetricFlow allows you to define, build, and maintain metrics in.. Apache-2
Prince (π₯28 Β· β 1.4K) - Multivariate exploratory data analysis in Python PCA, CA, MCA, MFA,.. MIT 
adapter-transformers (π₯27 Β· β 2.8K) - A Unified Library for Parameter-Efficient and Modular.. Apache-2 huggingface
avalanche (π₯26 Β· β 2K Β· π€) - Avalanche: an End-to-End Library for Continual Learning based on.. MIT
gplearn (π₯26 Β· β 1.8K) - Genetic Programming in Python, with a scikit-learn inspired API. BSD-3 
TabPy (π₯26 Β· β 1.6K Β· π€) - Execute Python code on the fly and display results in Tableau.. MIT
MONAILabel (π₯22 Β· β 760) - MONAI Label is an intelligent open source image labeling and.. Apache-2
apricot (π₯22 Β· β 520) - apricot implements submodular optimization for the purpose of selecting.. MIT
pykale (π₯21 Β· β 470) - Knowledge-Aware machine LEarning (KALE): accessible machine learning.. MIT 
SUOD (π₯21 Β· β 390 Β· π€) - (MLSys 21) An Acceleration System for Large-scare Unsupervised.. BSD-2
pymdp (π₯16 Β· β 570) - A Python implementation of active inference for Markov Decision Processes. MIT
Show 31 hidden projects...
- pyopencl (π₯31 Β· β 1.1K) - OpenCL integration for Python, plus shiny features.
βUnlicensed - pysc2 (π₯30 Β· β 8.2K Β· π) - StarCraft II Learning Environment.
Apache-2 - modAL (π₯30 Β· β 2.3K Β· π) - A modular active learning framework for Python.
MIT - datalad (π₯30 Β· β 620 Β· π) - Keep code, data, containers under control with git and git-..
βUnlicensed - cleanlab (π₯29 Β· β 11K) - Cleanlabs open-source library is the standard data-centric AI..
βοΈAGPL-3.0 - alibi-detect (π₯29 Β· β 2.4K) - Algorithms for outlier, adversarial and drift detection.
βοΈIntel - minisom (π₯28 Β· β 1.6K) - MiniSom is a minimalistic implementation of the Self Organizing..
βοΈCC-BY-3.0 - PySwarms (π₯28 Β· β 1.4K Β· π) - A research toolkit for particle swarm optimization in Python.
MIT - kmodes (π₯28 Β· β 1.3K Β· π) - Python implementations of the k-modes and k-prototypes clustering..
MIT - pyclustering (π₯28 Β· β 1.2K Β· π) - pyclustering is a Python, C++ data mining library.
BSD-3 - Cython BLIS (π₯28 Β· β 230) - Fast matrix-multiplication as a self-contained Python library no..
BSD-3 - Feature Engine (π₯26 Β· β 2.1K Β· π) - Feature engineering package with sklearn like functionality.
BSD-3 - metric-learn (π₯26 Β· β 1.4K Β· π) - Metric learning algorithms in Python.
MIT - pandas-ai (π₯25 Β· β 22K) - Chat with your database or your datalake (SQL, CSV, parquet)...
βUnlicensed - Mars (π₯24 Β· β 2.7K Β· π) - Mars is a tensor-based unified framework for large-scale data..
Apache-2 - AstroML (π₯24 Β· β 1.1K Β· π) - Machine learning, statistics, and data mining for astronomy..
BSD-2 - PaddleHub (π₯22 Β· β 13K Β· π) - 400+ AI Models: Rich, high-quality AI models, including..
Apache-2 - opyrator (π₯22 Β· β 3.1K Β· π) - Turns your machine learning code into microservices with web API,..
MIT - mlens (π₯22 Β· β 860 Β· π) - ML-Ensemble high performance ensemble learning.
MIT - BioPandas (π₯22 Β· β 740 Β· π) - Working with molecular structures in pandas DataFrames.
BSD-3 - benchmark_VAE (π₯21 Β· β 2K Β· π) - Unifying Variational Autoencoder (VAE)..
Apache-2 - impyute (π₯21 Β· β 360 Β· π) - Data imputations library to preprocess datasets with missing data.
MIT - StreamAlert (π₯20 Β· β 2.9K Β· π) - StreamAlert is a serverless, realtime data analysis..
Apache-2 - rrcf (π₯20 Β· β 520 Β· π) - Implementation of the Robust Random Cut Forest algorithm for anomaly..
MIT - scikit-rebate (π₯20 Β· β 420 Β· π) - A scikit-learn-compatible Python implementation of..
MIT - baikal (π₯18 Β· β 590 Β· π) - A graph-based functional API for building complex scikit-learn..
BSD-3 - pandas-ml (π₯16 Β· β 320 Β· π) - pandas, scikit-learn, xgboost and seaborn integration.
BSD-3 - KD-Lib (π₯15 Β· β 650 Β· π) - A Pytorch Knowledge Distillation library for benchmarking and..
MIT - NeuralCompression (π₯14 Β· β 580 Β· π) - A collection of tools for neural compression enthusiasts.
MIT - traingenerator (π₯13 Β· β 1.4K Β· π) - A web app to generate template code for machine learning.
MIT - nylon (π₯12 Β· β 82 Β· π) - An intelligent, flexible grammar of machine learning.
MIT
Related Resources
- Papers With Code: Discover ML papers, code, and evaluation tables.
- Sotabench: Discover & compare open-source ML models.
- Google Dataset Search: Dataset search engine by Google.
- Dataset List: List of the biggest ML datasets from across the web.
- Awesome Public Datasets: A topic-centric list of open datasets.
- Best-of lists: Discover other best-of lists with awesome open-source projects on all kinds of topics.
- best-of-python-dev: A ranked list of awesome python developer tools and libraries.
- best-of-web-python: A ranked list of awesome python libraries for web development.
Contribution
Contributions are encouraged and always welcome! If you like to add or update projects, choose one of the following ways:
- Open an issue by selecting one of the provided categories from the issue page and fill in the requested information.
- Modify the projects.yaml with your additions or changes, and submit a pull request. This can also be done directly via the Github UI.
If you like to contribute to or share suggestions regarding the project metadata collection or markdown generation, please refer to the best-of-generator repository. If you like to create your own best-of list, we recommend to follow this guide.
For more information on how to add or update projects, please read the contribution guidelines. By participating in this project, you agree to abide by its Code of Conduct.