FastGPU

Quickstart

Let's create your first FastGPU Pod.

1. Setup payment method

To begin with, log in to the FastGPU. After logging in, you will be asked to set up your payment method. You can use your credit card to set up your payment method under the billing tab.

2. Create your first pod

Navigate to the pods page and choose the desired gpu card type of your choice. Once the card type is selected, you can choose the number of GPUs you want to use.

Once the resource is selected, you can choose the storage and image options.

There are some default options available for image. You can choose the image of your choice or pass in a image url to use your own image. If you'd like to use your private image, create a new registry auth and select the registry auth from the dropdown next to the image.

The storage within the pod is meant to be ephemeral and will be deleted once the pod is deleted or restarted. To apply persistent storage, you can choose the folder from file system and mount is as a folder within the pod. For example, if you want to mount the folder /data from the file system to the folder /mnt within the pod, you can choose the folder /data from the file system and mount it as /mnt within the pod.

For more information on storage, refer to the storage page.

3. Connect to your GPU

Once the pod status changes to Running, you can connect to your pod via

Web Terminal

On the pods page, click on the Terminal button to get the Web Terminal link to connect to your pod.

3.1 Connect via SSH and Jupyter Lab

We provided SSH and Jupyter Lab login support, for images in our default options, Installation can be skipped. For your own images, it is required, only Ubuntu based images are supported.

Installation

  1. Download https://public-storage.fastgpu.com/lepton_start.sh to your local editor.
  2. Edit line 4 to 8 to be your SSH password, SSH public key and Jupyter Lab token. Leave it empty to skip installation.
  3. Copy the file, paste into Web terminal and press enter.

SSH

On the pods page, click on the Connect button to get the SSH command to connect to your pod with password.

Jupyter Lab

On the pods page, click on the Connect button to get the Jupyter Lab link to connect to your pod, Jupyter Lab log will be stored at /var/log/jupyter.log