How do I setup Google Datalab (NOT Colab) for the course?

Part 1 (2018)
1

I have a GCP account and would like to use datalab. I saw a lot of posts on various platforms, Paperspace, Clouderizer, Colab etc, but nothing on datalab. Sorry if this is somewhat repetitive, but I wish there was a list of setup instructions/requirements that is not tied to a platform. Many Thanks,

2

Sorry I made a confusion about the earlier post so I edited…

So to me is very straight. When you turn you instance online you probably will have a terminal where you can do things and execute commands. So you can clone the git hub project and execute the command to install all the packages of the environment you need.

3

so cloning the github will also install all libraries/dependencies and all setup steps?! just sounds too good to be true.

All setup instructions i saw are geared towards one platform or another. All I’m asking here is: What are the requirements; a list. Otherwise i’ll just start running the notebooks and install the needed library whenever a cell error happens because of a missing library… this is what I’m trying to avoid

4

Hi again @Kais,

As the word suggests cloning and installing aren’t the same thing, I never implied otherwise. On github there is all the steps you need to follow to install the packages.

If its too complicated for you, try this approach:
1- Launch your VM
2- Verify if conda is there. If conda is not installed you will need to install it;
3- If conda is already installed then you start clonning the repository;
4- Step into the fastai folder (clonned folder);

cd fastai

5- Execute (problems may occur because packages):

conda env update

6- Activate Python environment:

conda activate fastai

7-Also is worth to mention that If any error happens during the installation process and it not finishes with success you will need to deal with the error inside the new created environment.