From Anaconda to tensorflow to jupyter, step on the pit and fill it all the way

1. install anaconda Possible pit
Download address ：https://repo.continuum.io/archive/index.html
I installed Anaconda3-5.2.0, To match python3.6. But then I found out that after the new virtual environment , You can customize what you need for installation python edition , therefore anaconda Version can be installed at will .
You need to add it to the environment variable during installation , I skip 了 vscode Installation , As a result, it won't open navigator, Has been installing Go around there , Don't complain .
Solution 1： Open as administrator . Find or turn around .
Solution 2：anaconda prompt Open as Administrator

conda --version
anaconda navigator

An error is as follows ：

The error reported here is because we can't find vscode The version of , Need to find anaconda_api.py,edit with pycharm And so on. , You cannot double-click , Double click directly and make changes without saving . Need to be in anaconda_api.py Middle handlebar Cong 811 The line vscode Related codes to return The previous part is commented out , That's all right. .vscode After the problem is solved ,anaconda Finally, it can be opened .
2. Installing a virtual environment
Why virtual environment ？ I use... Directly anaconda default root Is it not fragrant ？
I was lazy before and didn't install a new virtual environment , Directly in root Lower installation TensorFlow, Many dependent packages are found to be degraded or upgraded , The more secure, the more disorderly . This Zhihu answers the Lord to solve my confusion , The simplest example assumes that the framework needs to be installed A package , install A The prerequisite of the package is B Bag 2.1 Version and C Bag 1.2 All versions have , At this time, in your original system B The bag could be 3.2 edition , In order to pretend A The package needs to be downgraded , But the original in the system D The bag needs 3.2 Version of B package , therefore B The package has been downgraded , May lead to D The bag cannot be used . So we must have a new virtual environment , Dedicated to deep learning .
（ original text ：https://zhuanlan.zhihu.com/p/136268744）
stay anaconda prompt Open as administrator in ,

conda create --name energy-ai python=3.6

The speed of discovery is extremely slow , The error is also reported as follows ：

It seems that the remote server has timed out transmitting information , So an extended time limit is added ,

conda config --set remote_read_timeout_secs 1000.0

This is a new mistake , Server connection problems abroad , So replace the image source ：

pip install pip -U
pip config set global.index-url https://pypi.tuna.tsinghua.edu.cn/simple
conda create --name energy-ai python=3.6

this time sousou Of , It was installed soon .

activate energy-ai

3. install TensorFlow
The previous installation with the command line was very unsuccessful , This time I'm directly in navigator Of environment I found it directly TensorFlow, Click directly .

4.jupyter Installation and normal use
jupyter Is installed directly on the energy-ai One click installation under the interface . But there are problems in using , Has been submitted to the kernel error, Click runtime , Will jump to the next cell, But no error or result .
The analysis reason ： Then used Anaconda Install multiple versions of python When , Or because of multiple uninstallations anaconda when , Lead to python Of kernel The kernel installation path is chaotic , bring Jupyter notebook Unavailable . I uninstall many times anaconda As a result of .

One 、 Start by opening Anaconda Prompt
Two 、 Input jupyter kernelspec list Check the installed kernel and location
Let's say mine is C:\Users\bonni\AppData\Roaming\jupyter\kernels\python3
3、 ... and 、 Enter the location where the kernel is installed , open kernel.jason, see python Whether the path file of the editor is the same as the installation path , If it is different, you can modify it （ Change it to python.exe In the directory ）
All changed to something like this C:\Users\bonni\Anaconda3\envs\energy-ai\python.exe
, It's your anaconda Of the current virtual environment under the installation path python.exe Location path .
Link to the original text ：https://blog.csdn.net/wewearef/article/details/118761042