What you can do on the AI learning environment construction service portal site

For data scientists working on AI development as a team, AI learning environment construction service proposes and builds optimal infrastructure for AI development according to customer's development environment and provides UI for developers.

 

This article introduces an overview of the AI learning environment construction service. https://www.macnica.co.jp/business/semiconductor/manufacturers/nvidia/products/133865/

This article explains in detail about the portal site provided by this service.

Reasons why a simple UI that can be operated intuitively is required

In model development, it is necessary to shorten the cycle of data collection, model creation, learning, and evaluation until the expected accuracy and results are obtained. However, when performing these flows, trial and error is performed until the expected accuracy is achieved, so in many cases containers are used to set up the development environment, and it is necessary to acquire knowledge about infrastructure first.

 

Therefore, in model development, if an interactive development environment is immediately set up and there is a UI that can be operated in a browser until batch job submission with trial and error of hyperparameters, development can be started without infrastructure knowledge such as containers and container orchestration. We have proceeded with the construction of the portal site by reflecting the customer's voice that it would be possible to

 

First of all, you can see the image when you actually use it in the video.

We will introduce the functions that you can actually use.

·container

A dedicated container repository is prepared for the company, and it is used for sharing container images between teams and storing containers that have been worked on.

·catalog

It is a feature that can be started by simply filling in the requirements of a containerized application.

·notebook

Simply select the framework you want to use, the number of GPUs, and the development environment you want to use (Jupyter Lab or Vistual Studio Code) to launch an interactive development environment.

・Batch job

It assists in performing high-volume, repetitive data jobs such as processing data and finding model hyperparameters.

・Git

Gitlab dedicated to the company is provided, and code management using Git can be performed.

·storage

You can upload, delete, and download data from your browser.

・Metrics

You can visualize and check the usage status of the managed computing servers.

·User Management

We provide a function to create and delete users from the UI for administrators.

You can touch the trial environment for free

I have explained the portal site provided by the AI learning environment construction service, how was it? We have prepared a trial environment where you can try the operation of the portal site for free, so please contact us if you are interested.