Please click here to enter a description of the picture

1.Pipenv

Pipenv is Kenneth Reitz’s side project to integrate other packages (such as NPM and YARN) into Python. It does not need to install Virtualenv, virtualenvwrapper, manage the requirements. TXT file, and ensure that dependent versions are reproducible. With Pipenv, you can specify dependencies in a Pipfile. The tool generates a pipfile. lock file to make your build more deterministic and avoid hard-to-find bugs.

2.PyTorch

PyTorch is a Facebook deep learning framework derived from and improved on the Torch framework. Based on the Ython language, PyTorch has become one of the preferred frameworks for many researchers due to its implementation of the dynamic computational graph paradigm. PyTorch can compute gradients very fast and extensible.

3. Caffe2

Caffe2 supports distributed training, deployment (even on mobile platforms), new cpus and CUDA-enabled hardware. PyTorch is probably better suited for research, while Caffe2 is better suited for large-scale deployments, as seen on Facebook. Alternatively, you can build and train models in PyTorch while deploying with Caffe2.

4. Pendulum

One of the advantages of Pendulum is that it is a standard Python alternative to DateTime, so you can easily integrate it with existing code and use the functionality only when you need it. The authors of the Pendulum paid particular attention to the handling of time partitions, which are available by default in each instance and are timed in UTC. You can also get the timedelta extension to simplify datetime calculations.

5. Dash

Dash is a pure Python open source library for building Web applications, especially data visualization Web applications. It is built on top of Flask, Plotly, and React, and provides a functional abstraction interface for these frameworks so developers don’t have to learn them to develop efficiently. These applications are available in browsers and mobile devices.

6. PyFlux

PyFlux is an open source Python library developed specifically for time series. Time series research is a subfield of statistics and economics. Its purpose is to describe the behavior of time series and to predict the future behavior of time series.

7. Fire

Fire is an open source library that automatically generates a command line interface for any Python project. You hardly need to write any code or documentation, you just need to call a Fire method and pass it to the command line interface you want: a function, an object, a class, a library, or even pass no arguments.

8. imbalanced-learn

Imbalanced-learn is a Python library that provides techniques for dealing with data imbalances. In addition, it is compatible with Scikit-learn and is useful as part of the Scikit-learn-contrib project.

9. FlashText

FlashText demonstrates the importance of algorithm and data structure design, and that even for simple problems, better algorithms can easily outperform naive implementations running on fast cpus.

10. Luminoth

Luminoth is an open source Python kit for computer vision built with TensorFlow and Sonnet. It supports object detection directly, and the model behind it is Faster R-CNN.


These are 10 popular machine learning libraries. Which ones have you heard about or learned about? If you’re interested, take a closer look.