| # | Week 31, 2023 | August 20, 2023 |
|---|---|---|
| 1 | Jax | A Python library for optimizing scientific computing developed by Google in 2018: It can be regarded as NumPy running on GPU and TPU. jax.numpy provides an API interface very similar to numpy. , Just After eXceution (JAX) is NumPy on CPU, GPU, and TPU with excellent automatic differentiation. It is a Python library for high-performance numerical computing, especially machine learning research. Its numerical API is based on NumPy, a library for scientific computing. Both Python and NumPy are well-known and commonly used programming languages that make JAX straightforward, versatile and easy to implement. In the field of machine learning, everyone may be familiar with TensorFlow and PyTorch, but in addition to these two frameworks, some new forces should not be underestimated. It is JAX launched by Google. What was the starting point for developing JAX? Speaking of this, I have to mention NumPy. NumPy is a basic numerical operation library in Python and is widely used. However, numpy does not support GPU or other hardware accelerators, and there is no built-in support for backpropagation. In addition, the speed limit of Python itself hinders the use of NumPy, so few researchers directly use numpy to train or deploy deep learning models in production environments. . In this case, numerous deep learning frameworks have emerged, such as PyTorch, TensorFlow, etc. However, numpy has unique advantages such as flexibility, convenient debugging, and stable API. The main starting point of JAX is to combine the above advantages of numpy with hardware acceleration. TensorFlow was developed by Google. The initial version can be traced back to the open-source TensorFlow0.1 in 2015. Afterwards, it has developed steadily and has a strong user group, becoming the most popular deep learning framework. However, when users use it, they also expose the shortcomings of TensorFlow, such as insufficient API stability and complex programming of static calculation graphs. Therefore, in the TensorFlow2.0 version, Google included Keras and became tf. keras. PyTorch (Python-Torch) is a machine learning library from Facebook. Use TensorFlow or PyTorch? A year ago, this issue was uncontroversial, and most researchers would choose TensorFlow. But the situation is very different now, and more and more researchers use PyTorch. |
| 2 | KeymouseGo | Mouse and keyboard recording and automatic operation similar to button wizard to simulate clicking and typing. Function: record the user's mouse and keyboard operation, automatically execute the previously recorded operation by triggering the button, and the number of executions can be set, which can be understood as a simplified green version of the button wizard . Uses: When doing some simple, monotonous and repetitive work, using this software can save you a lot of effort. Just do it yourself, and then let the computer do it. Desktop mode 1. Click the record button to start recording. 2. Perform any operation on the computer, such as clicking the mouse or keyboard input, and this part of the action will be recorded. 3. Click the end button to end the recording. 4. Click the start button, and the computer will repeat the actions recorded in step 2. |
| 3 | roop | One-click face swap |
| 4 | youtube-dl | Command line program to download videos from YouTube.com and other video sites |
| 5 | locust | An open source load testing tool. User behavior is defined using Python code, and millions of users can also be simulated. Locust is very easy to use, distributed, user load testing tool. Locust mainly performs load testing for websites or other systems, and can test how many users a system can handle concurrently. Locust is completely time-based, so a single machine supports thousands of concurrent users. |
| 6 | fell | A free and open source painting software released under the GPL license. It has complete functions and is capable of all painting processes from drafting, sketching, coloring to final adjustment. It can draw concept sketches, illustrations, comics, animations, scenes and 3D textures. It supports digital boards, pressure sensitivity, anti-shake, Layers, filters, color management and more |
| 7 | zombie | This Plants vs. Zombies is a shooting version defense game based on pygame. Simple and easy to play, easy to use, general shooting games can use the framework of this game. |
| 8 | facechain | A deep learning model tool that can be used to create a personal digital image. Users only need to provide a minimum of three photos to get their own personal digital avatar. FaceChain supports the use of model training and reasoning capabilities in the gradient interface, and also supports senior developers to use python scripts for training and reasoning. |
| 9 | kityminder | Another masterpiece of the f-cube team (formerly UEditor team) of the Baidu FEX team. As an online brain map editing tool, it has a lot of interactive experience of Native editing tools. KM has the same purpose as UE, which is to focus on the end user's experience. At the same time, it gives full play to the advantages of Web cloud storage, and can directly synchronize the edited mind map to the cloud. In addition, with the original "cloud disk sharing" function, users can directly generate an online link to share the currently edited mind map with other users with one click, achieving seamless communication. KM is implemented based on SVG technology, using JavaScript+html. Support most major browsers. |
Category: Weekly Ranking of Python Open Source ProjectsTags : scientific computing button wizard pygame load testing youtube game
Note: The current article will be updated from time to time. If you have better suggestions for this article, or new materials to recommend, you can click: Welcome to share excellent websites .