Become a beta tester to receive 25€ in free credits and get access to discounted credit packs

Join the beta

How to train a machine learning model with Tresorio ?

Why choose Tresorio?

Train your models on GPU-powered resources in just a few clicks and access the power of the Tresorio Cloud in seconds. Enjoy:

  • a simplified configuration
  • GPU to optimize your computations
  • real time monitoring

Moreover, thanks to our micro data centers that act like digital boiler, 100% of the computing power you consume is turned into hot water and helps heat buildings.

Tutorial

Now let’s see how to train an AI model on Tresorio!

Take a look at our video tutorial or read our article below

Create your project

Click on “create your first projet“.

create a new cloud computing project on tresorio

Link your provider

Link a provider to your project (Github/Gitlab)

link your provider on tresorio

Choose your repository (Github/Gitlab)

Now that you have linked your provider, select your repository and click on the « Confirm » button.

select your provider on tresorio
select your repository on tresorio

Set up your project

Select your powerpack, write your script and set your timeout.

set up your cloud computing project on tresorio

Click on the « create project » button. You’ll be redirected to your project page.

cloud computing project created on tresorio

Choose or upload your datasets

Click on the “+” button to add a dataset.

If you don’t have any available dataset, click on the link on the pop-up to upload a new one.

choose from your dataset on tresorio

Add a new dataset

Click on “Add a new dataset”.

add a new dataset on tresorio

Name your dataset, upload your file or folder, and click on the « save » button.

upload your dataset on tresorio

Your dataset has been successfully added ! You can now add your dataset to your project.

Link your dataset to your project

In the navigation menu, select « Your projects ». Click on the project that you have previously created.

your cloud computing project on tresorio

Under the name of your project, you will find a « + » button. Click on it to link one or more datasets to your project.

link dataset on tresorio

Choose from your datasets

A pop up will appear asking you to choose from your uploaded datasets. Select your dataset and click on the « save » button.

Your project is now complete and you are ready to start a new training.

new training on tresorio

Launch a new training

On your project page, click on the « new training » button. Select or modifiy your powerpack, set or modify the timeout and complete your project commands.

Find your datasets in /app/datasets. Use /app/artifacts/ to host your models, datas and more. Then click on the « launch training » button.

launch your training on tresorio

Your training will run instantly.

Training status

Once your training is running, you will reach its dedicated page. Here you can check resources usage and monitor your training.

monitor your training on tresorio
resources usage on tresorio

Check your logs in real time.

check your logs on tresorio

Training completed !

Your training is completed when the « resources » tab becomes « logs » and the « logs » tab become « artifacts ». You’ll also see a notification on the right of your screen telling you that your training has been successfully completed.

Download your outputs

Click on the « Download training results » button to download your outputs.

download training results tresorio

Once downloaded, you will find a zip file in your « Downloads » folder. Unzip it and find your logs and artifacts in the output folder.

You will also like

    Tresorio press release
    Press releases 20 December 2019

    TRESORIO raises €480,000 to accelerate its development

    Tresorio press release
    Press releases 20 December 2019

    Dalkia recovers heat from Tresorio’s computer servers to heat part of the hot water at Mercy’s hospital in Metz.

    VPS Tresorio
    Tutorials 11 December 2019

    How to deploy a Virtual Machine on Tresorio Cloud?