Using Google colab to conduct deep learning training research on fast AI in the cloud

Configure hardware requirements before development

Colab provides free GPU usage to the vast number of deep learning/machine learning developers, can be used without any settings, and runs completely in the cloud.
To use Google colab to develop and research fast AI in the cloud, you must have:
1. A computer/mobile phone (I haven’t tried a mobile phone)
2. You must be able to access the YouTube network
3. Have a Google account. If you don’t have one, you can register for one.

---

1. Construction of development environment

1. Connect to the Internet, that is, to be able to access the YouTube network. (I understand everything)
2. Use the Google browser to log in to your Google account, register one if you don’t have one, and there are no special restrictions.
3. Prepare the environment in colab

• ①. Search colab in the Google browser and enter the colab page, as shown in the figure below.
• 

②. On the web page, click File - Create a new file to create a new file to start.
Go to the original link on Github: About fastai tutorial on GitHub .

Here is a related example about coloring old videos and photos - DeOldify - Colorize your own videos ! These videos and photos are black and white, and the processed videos and photos of this model are colored videos and photos. You can try to run other people's training examples using Google colab, and try to feel a sense of accomplishment.

• The .ipynb file created in colab before will have a save record, and you can also view the previously created file in the cloud later, but you need to run it again to have data.

• If you do not operate for too long, the connection will be disconnected. After disconnecting and reconnecting, you must run the previous code again.

• Only one session content can be run at the same time