Recently, deep learning has developed very rapidly, and it has the potential to unify the rivers and lakes. After a period of study, I found myself very interested in this magical metaphysics, hoping to study it further. However, it is difficult to understand this kind of research topic simply by reading the paper, so I hope to follow up the code written by Daniel and study it in depth. I found that many source codes given by Daniel were written based on python, so I planned to learn Python. ?
At first, I installed python directly, wrote the script under windows Notepad, and then ran it under cmd. But this is inefficient and difficult to debug. But python shell with comand line and IDLE with graphical interface can only interact line by line, which is also troublesome. So I planned to build an IDE myself, similar to Microsoft's VS, but it took several days and I found it was still too young. In addition, I need to download specific modules from official website of python, which is very troublesome. Finally, I chose anaconda, which is easy to use and simple. The download address of anaconda is here:/files/doc/DocBook/Python _ beginner _ tutorial/release/HTMLS/CH04 _ Python _ dev _ env.html # how _ dev _ python _ in _ window.
Anaconda has its own python and most python modules. Specific modules can be viewed in anaconda official website. My IDE is Spyder, and the interface is as follows:
We can see that the interface design of Spyder is very similar to that of matlab. Students who are familiar with matlab can quickly get used to using Spyder, but there are some differences. Let me add these differences. ?
1. First, shortcut keys. Generally speaking, the shortcut key for debugging will be set to F5, and its debugging will be Ctrl+F5. Step by step debugging is different. In short, for convenience, I changed the shortcut keys for debugging into the customary way. The modification method is Tools on the toolbar-> Modify. Preferences-> Keyboard shortcuts, and then modify the shortcuts accordingly. ?
2. Secondly, how to set breakpoints. Generally speaking, the breakpoint is to click the vertical bar next to the editing area once, but Spyder is twice, and canceling is also a double click. Personally, I can't understand why this is set up. Of course, you can also directly select the shortcut key of the breakpoint. The default shortcut key is Ctrl+F 12, and I changed it to F 12. ?
3. Code prompt, matlab code prompt is to enter the first few letters of the function and then press the tab key, python is to write one first and then press the tab key. If it is a created object, write it directly. There will be a code prompt. ?
4. It is best not to write the path of the script in Chinese, because python probably won't recognize it. I can't tell you the specific reason, but this is a lesson from blood.
Recommended course: Python video tutorial of Zhonggu Education