local_server

Deploy ColabSaprot on local server

For users who want to deploy ColabSaprot on their local server, please follow the instructions below:

For Linux OS

Step1: Install the environment

cd local_server # This is important because the system will find files using relative path!
bash install.sh

Step2: Start jupyter notebook

bash run.sh

Step3: Connect to local server on the colab platform

See here.

For Windows

Step1: Preparation

1. Download SaprotHubImage.tar and WindowsDeploy.bat in one folder.
2. Make sure you have installed Docker Desktop and it is started

Step2: Installation

Click the WindowsDeploy.bat until the Ubuntu bash shell is shown like the figure below

Step3: Start jupyter notebook

Entering the commands below into the bash shell

cd ~/SaprotHub/local_server
bash run.sh

Step4: Connect to local server on the colab platform

See here.

Connect to local server on the colab platform

Open the SaprotHub.ipynb and click Connect to a local runtime, then input the address of the local server: http://localhost:12315/?token=SaprotHub.

(Optional) SSH port forwarding

If the GPU is deployed on your remote server, you can use SSH port forwarding to connect to the local server.

ssh -NL 12315:localhost:12315 user@remote_server

After that, you can connect to the local server on the colab platform.