10 Popular Python Machine learning libraries __python

1.Pipenv

Pipenv is a Kenneth Reitz amateur project designed to integrate other software packages, such as NPM and yarn, into Python. It does not need to install virtualenv, Virtualenvwrapper, do not manage requirements.txt files, and does not have to ensure the reproducibility of dependent versions. With pipenv, you can specify the dependencies in the Pipfile. The tool generates a Pipfile.lock file that makes your build more deterministic and avoids bugs that are hard to find.

2.PyTorch

Pytorch is the Facebook depth learning framework, derived from the Torch framework, while improving the Torch framework, based on the Ython language, the Pytorch has become one of the preferred frameworks for many researchers because of the implementation of the dynamic graph paradigm, and can compute gradients, and very quickly, Scalability is strong.

3. Caffe2

CAFFE2 supports distributed training, deployment (even on mobile platforms), new CPUs, and hardware that supports CUDA. Pytorch may be more suitable for research, and CAFFE2 is more suitable for large-scale deployments, as seen on Facebook. In addition, you can build and train the model in Pytorch and deploy it with Caffe2.

4. Pendulum

One of the advantages of pendulum is that it is a Python standard datetime substitute, so you can easily integrate it with existing code and use it when you need it. Pendulum's authors pay special attention to the processing of time partitions, which are available by default in each instance, and are timed in UTC. You can also get extended timedelta to simplify the calculation of DateTime.

5. The Dash

The Dash is a pure Python open Source Library that can build Web applications, especially data visualization Web applications. It builds on flask, plotly and react, and provides functional abstraction interfaces for these frameworks, so developers don't have to learn these frameworks and develop them efficiently. These applications can be used in browsers and mobile devices.

6. Pyflux

Pyflux is a Python Open source library designed specifically for time series development. The study of time series is a sub domain of statistics and economics, which is used to describe the behavior of time series, and also to predict the future behavior of the sequence.

7. Fire

Fire is an open source library that can automatically generate 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, and even no arguments.

8. Imbalanced-learn

Imbalanced-learn is a Python library that provides the relevant technology to solve the problem of data imbalance. In addition, it is compatible with Scikit-learn and is very useful as part of the Scikit-learn-contrib project.

9. Flashtext

Flashtext proves the importance of algorithms and data structure design, even for simple problems, better algorithms can easily surpass the simple implementation of running on the fast CPU.

Luminoth

Luminoth is an Open-source computer Vision Python Toolkit built with TensorFlow and Sonnet. It can directly support object detection, and the model behind it is faster r-cnn.