Getting started with Deque tools is a breeze. When the user logs into the app, they are taken to the home page. To the left of the home page is the sidebar, which has a few key buttons including Home/Recent, Tools, Drive and Manage.
Check out this video to get started:
This sidebar is available on all screens within the app and offers a quick way to navigate within the app.
If the user's intent is to use pre-downloaded data then the user can upload that data into the Drive. To get started the user can select the DRIVE
button on the sidebar. The Drive is a very familiar file browser interface. Deque allows the user to create their own folder and file hierarchy with no limitations. At the top left the user can click on the more icon
and select the upload folder or upload file. The upload happens asynchronously in the background. It is important to note that the entire Drive is automatically mounted across the users Notebooks, Jobs and Predictions.
Once the data upload is in progress, the User can launch their first Notebook, by clicking the Plus icon on the top right and selecting Notebook. This will open a “Create Notebook Settings Window '' in which the user can select a Notebook file from their computer or check out from GitHub. Additionally, the user can select the instance type they would like to launch the Notebook on, the disc size and even optionally enable experiment tracking by selecting between Tensorboard https://www.tensorflow.org/tensorboardand MLflow https://www.mlflow.org under the experiment settings section. Notebook creation is almost instantaneous and once the Notebook is created, the user will see the content of the Notebook. At this point the user can click on the “Start” button at the top left to start the Notebook instance.
The Notebook instance can take several minutes to start depending on the cloud provider. Once the instance is up and running the Notebook cells are activated and the user can start editing the code and running the cells. A user can refer to their Drive by simply typing /Drive in their code and access the data as if exists on the Notebook instance. They can also save their trained models in the Drive in a similar fashion. Congratulations!
You have launched your first Notebook successfully with access to your Drive.