gpu-tpu-lab-server

gpu_tpu_lab_server

Running a jupyter lab or vs code server environment from colab with google drive support.

references:

fns.colab. Most of the localtunnel colab code is taken from this repo.

Motivation for creating this package

• For some reason in fns.colab python package version >=0.5, fns.colab supports jupyter notebook not jupyter lab, so here I am trying to use the previous version of that fns.colab code to support jupyter lab environment and to fix frequent time out and connection issues.> 1. It was boring to copy paste the whole code everytime to start the jupyter lab environment in colab, so I created a package of it.
• Also, I wanted to try how to create python package using nbdev, this sounded like a good excuse.

Install

pip install gpu_tpu_lab_server

How to use

from gpu_tpu_lab_server.core import jupyter_lab, vscode, run_process_in_background, run_process_in_foreground

For starting the server in google colab with google_drive support

To enable jupyter_lab_environment

• execute jupyter_lab()
• an authorization window will open with a link, go to the link > select your google account > accept all the permissions > authorization code will be displayed > copy and paste the code in the window
• an localtunnel link will be returned, go to that link, wait for 10 second, then press open tunnel button, it will open an jupyter lab environment

To enable vscode() environment

• execute vscode() optional parameter any text :- the generated link will have the text or by default fastdaima
• an authorization window will open with a link, go to the link > select your google account > accept all the permissions > authorization code will be displayed > copy and paste the code in the window
• an localtunnel link will be returned, go to that link it will open an vscode server environment