Thanks again to Ahmed Anis for his contribution to the collection of these data and to the rest of KDnuggets staff for their comments, insights and suggestions.
By Matthew Mayo
Original link /www.kdnuggets.com/2020/11/top…
Note that the following diagram, drawn by Gregory Piatetsky, represents each library by type, by star and contributor, and its sign size is expressed as the logarithm of the number of times the library has been committed on Github.
Figure 1: Top-level Python libraries for deep learning, natural language processing, and computer vision
Plotted by star rating and number of contributors; The relative size is expressed in logarithms of submissions
So, without further ado, here are 30 top Python libraries for deep learning, natural language processing and computer vision, handpicked by KDnuggets staff.
Because the jump link of wechat public account can only set the tweets of the public account, the link needs to be pasted into the browser for access.Copy the codeDeep learning
1. TensorFlow (github.com/tensorflow/…
Star: 149000, number of submissions: 97741, Contributors: 2754
TensorFlow is an open source, end-to-end platform for machine learning. It has a comprehensive and flexible ecosystem of tools, libraries, and community resources that help researchers advance advanced machine learning technologies and make it easy for developers to build and deploy machine learning-based applications.
2. Keras (github.com/keras-team/…
Star: 50000, submissions: 5349, Contributors: 864
Keras is a deep learning API written in Python that runs on top of TensorFlow, the top platform for machine learning.
3. PyTorch (github.com/pytorch/pyt…
Star: 43200, submissions: 30696, Contributors: 1619
Powerful GPU acceleration, and tensors and dynamic neural networks implemented using Python.
4. Fastai (github.com/fastai/fast…
Star: 19800, number of submissions: 1450, Contributors: 607
Using current best practices, FASTAI simplifies the training process and speeds up neural networks very quickly.
5. PyTorch Lightning
(github.com/PyTorchLigh…
Star: 9600, submissions: 3594, Contributors: 317
PyTorch is a lightweight package for high-performance ARTIFICIAL intelligence research. Can shrink your model rather than provide small templates.
6. JAX (github.com/google/jax)
Star: 10000, submissions: 5708, contributors: 221
Combinable conversions of Python+NumPy programs: differentiation, vectorization, JIT on GPU/TPU, etc.
7. MXNet (github.com/apache/incu…
Star: 19100, submissions: 11387, Contributors: 839
Lightweight, convenient, flexible distributed/mobile machine learning library with dynamic and mutation-aware data flow management scheduler: support for Python, R, Julia, Scala, Go, JavaScript, and more.
8. Ignite (github.com/pytorch/ign…
Star: 3100, submissions: 747, Contributors: 112
An advanced library to help train and evaluate neural networks for flexible and transparent use in PyTorch.
Natural language processing
9. FastText (github.com/facebookres…
Star: 21700, submissions: 379, Contributors: 47
FastText is a library for learning word ideation and sentence classification efficiently.
10. SpaCy (github.com/explosion/s…
Star: 17400, submissions: 11628, contributors: 482
Used to implement industrial level natural language processing (NLP) through the use of Python and Cython.
11. Gensim (github.com/RaRe-Techno…
Star: 11200, submissions: 4024, Contributors: 361
Gensim is a Python library for topic modeling, document indexing, and similarity retrieval with a large corpus. The target audience is the natural language processing (NLP) and information retrieval (IR) communities.
12. NLTK (github.com/nltk/nltk)
Star: 9300, submissions: 13990, Contributors: 319
NLTK — Natural Language Toolkit — is an open source set of Python modules, data sets, and tutorials for research and development in support of natural language processing.
13. Datasets (Huggingface)
(github.com/huggingface…
Star: 4300, submissions: 568, Contributors: 64
For natural language processing, using tools such as PyTorch, TensorFlow, NumPy and Pandas, with fast, efficient, open-access datasets and natural language processing evaluation metrics.
14. Tokenizers (Huggingface) (github.com/huggingface…
Star: 3800, submissions: 1252, Contributors: 30
Fast, top of the line word divider optimized for research and production.
15. Transformers (Huggingface)
(github.com/huggingface…
Star: 3500, submissions: 5480, Contributors: 585
Transformer: Top of the line natural language processing libraries for Pytorch and TensorFlow 2.0.
16. Stanza (github.com/stanfordnlp…
Star: 4800, submissions: 1514, Contributors: 19
The official standard NLP Python library for many human languages.
17. TextBlob (github.com/sloria/text…
Star: 7300, submissions: 542, Contributors: 24
Simplicity, Python style, text processing — sentiment analysis, part of speech tagging, noun phrase extraction, translation, etc.
18. PyTorch-NLP
(github.com/PetrochukM/…
Star: 1800, submissions: 442, Contributors: 15
PyTorch is a basic tool for natural Language processing (NLP).
19. Textacy (github.com/chartbeat-l…
Star: 1500, submissions: 1324, Contributors: 23
A Python library for performing various natural language processing (NLP) tasks, built on top of a high-performance spaCy library.
20. Finetune (github.com/IndicoDataS…
Star: 626, submissions: 1405, Contributors: 13
Finetune is a library that allows users to perform a wide range of downstream tasks using state-of-the-art pre-trained NLP models.
21. TextHero (github.com/jbesomi/tex…
Star: 1900, submissions: 266, contributors: 17
Text preprocessing, presentation and visualization, from zero to mastery.
22. Spark NLP (github.com/JohnSnowLab…
Star: 1700, submissions: 4363, Contributors: 50
Spark NLP is a natural language processing library built on top of Apache Spark ML.
23. GluonNLP (github.com/dmlc/gluon-…
Star: 2200, submissions: 712, Contributors: 72
GluonNLP is a toolkit that supports simple text preprocessing, data set loading, and neural model building to help you speed up natural language processing (NLP) research.
Computer vision
24. Pillow (github.com/python-pill…
Star: 7800, submissions: 10799, Contributors: 303
Pillow is a very user friendly branch of PIL. PIL is Python’s image library.
25. OpenCV (github.com/opencv/open…
Star: 49600, submissions: 29453, Contributors: 1234
Open source computer vision library.
26. Scikit – image (github.com/scikit-imag…
Star: 4000, submissions: 12352, Contributors: 403
Image processing in Python.
27. Mahotas (github.com/luispedro/m…
Star: 644, submissions: 1273, Contributors: 25
Mahotas is a library that contains the fastest computer vision algorithms (all implemented in C++ for speed), running against numpy arrays.
28. Simple – CV (github.com/sightmachin…
Star: 2400, submissions: 2625, Contributors: 69
SimpleCV is an open source machine vision framework that uses the OpenCV and Python programming languages.
29. GluonCV (github.com/dmlc/gluon-…
Star: 4300, submissions: 774, Contributors: 101
GluonCV provides the implementation of the most advanced (SOTA) deep learning model in computer vision.
30. Torchvision (github.com/pytorch/vis…
Star: 7500, submissions: 1286, Contributors: 334
The TorchVision package consists of popular data sets, model architectures, and image transformation methods commonly used in computer vision.