Tutorial
  • Research Skill Bootcamp
    • Overview
    • Introduction to Research
    • Literature Review
    • Latex, Overleaf and Template
    • Latex, use TikZ to draw diagram
    • Introduction to Pytorch
    • Introduction to Neural Networks, MLP, CNN, RNN, Transformer
    • Problem Formulation and Experiment Design
    • Before Vision Becomes Reality
  • AI Engineer Bootcamp
    • Poster
    • Introduction
    • Development Environment Setup
    • Docker
    • Git, GitHub and Agile
    • Introduction to RAG
    • Full Stack Intro, Demo and Setup
    • Python Package Development
    • Databases
    • React
    • Django
    • GraphQL and Hasura
    • Authentication and Authorization
    • Deploy and CI/CD
    • Project Demo
    • External Resources
    • Knowledge Graph and GraphRAG
  • Dev Setup
    • How to install Docker
    • Docker 101
  • GPU Resources
    • HPC 101
    • Kaya
      • Demo project
      • Interactively use Kaya with JetBrains Gateway
      • Battle notes
      • Run experiment on Multiple GPU in Kaya
    • DUG
      • DUG HPC FastX connection Guide for Linux
  • Background Knowledge
    • Public Available LLM
      • Self-Hosted LLM
      • Quantization demo
      • Public Open LLM API
Powered by GitBook
On this page
  • Why recommend JetBrains tools
  • How to connect
  • Develop interactively
  1. GPU Resources
  2. Kaya

Interactively use Kaya with JetBrains Gateway

https://www.jetbrains.com/remote-development/gateway/

PreviousDemo projectNextBattle notes

Last updated 9 months ago

Like the VSCode Remote SSH function, JetBrains also provides similar tools.

Why recommend JetBrains tools

Use which tool to do development will be up to you, but I do suggest new beginners to use Jetbrain products to start, as their intelligence engine will help you learn how to implement your code with best practice. For example, if you have named your variable duplicated with a global variable, which potential will introduce a ghost bug for you later, it will recommend you not to do that. And some pre-check of the syntax can help you to reduce the bugs.

VSCode do can achieve the same functionalities with proper setup, however, the beneifit I mentioned above are all out-of-box, which is the reason why I say new beginners should use it.

Student License

As long as you are a student, you can apply for a free educational license to start with.

And when you graduate with good practice regarding programming, you may want to switch to vscode.

How to connect

After you have Jetbrain Gateway installed, you can open it and create a connection via SSH:

After that you will see something like this:

And then you can create a project via different IDE they provide, for us, most common one is PyCharm.

Select the IDE you want, and then select the project directory, you can start the IDE and do the development.

At this stage, there will be not much difference between use the interface and your local PyCharm

As we mentioned before, when you doing some edit here with the code, because of the storage setup of the Kaya, the changes will be reflected on the compute node.

Develop interactively

salloc -p gpu --mem=16G -N 1 -n 8 --gres=gpu:p100:1

In the IDE terminal, type the command above to request a P100 node (You can adjust the command to use your specific required GPU)

Thenyou should be able to change the code in IDE and then run the latest changes in the terminal to verify your changes now.

You can also queue the job via sbatch here, and check the results from the left file list panel.

Free Educational Licenses - Community SupportJetBrains
Logo