Creating a New Project

The Data Science Lab (DSLab) enables users to create projects with customized environments, resources, and Git integration. This guide explains the step-by-step process of creating a new DSLab project.

Access the Project Creation Page

Navigation path: Data Science Lab > Projects > Create

1. Navigate to the Projects page in the DSLab module.

2. Click the Create option.

3. The Create Project form opens.

Provide Project Details

Fill in the following information:

• Project Name: Enter a name for the new project.

• Project Description: (Optional) Provide a short description.

• Select Algorithms: Choose from the available algorithms using the drop-down menu.

Environment Selection

Choose the execution environment for the project. Currently supported environments:

• Python TensorFlow – Preconfigured with TensorFlow and Keras. No manual installation required in notebooks.

• Python PyTorch – Preconfigured with PyTorch and Torchvision. No manual installation required.

• PySpark – Preconfigured for distributed processing with Apache Spark.

• Python – A general-purpose Python environment suitable for traditional machine learning, data analysis, and scripting.

• BaseDS – A lightweight base environment that provides core data science libraries and utilities for custom experimentation.

• Selecting the right environment ensures that required frameworks are preinstalled and available directly inside notebooks.

Resource Allocation

Specify the resource limits for the project workspace container:

• Options: Low, Medium, High.

• Includes allocation of CPU, GPU, and memory.

Idle Shutdown

Define the idle time after which the notebook session disconnects and the project deactivates:

• Options: 30 minutes, 1 hour, 2 hours.

• To continue work, you must reactivate the project.

External Libraries

• Provide names of external libraries to install in the project environment.

• If a specific version is required, include the version number (e.g., pandas==2.0.3).

• Separate multiple library names with commas without spaces.

• This field is optional.

Advanced Configuration

After filling in the mandatory fields, additional modifiable options appear with preselected values:

• Image Name: Default container image.

• Image Version: Preselected image version.

• Limit (CPU/Memory): Maximum resources assigned.

• Request (CPU/Memory): Minimum resources requested.

GPU Configuration

• GPU Type: Currently supports NVIDIA GPUs.

• GPU Limit: Set maximum GPU resources (visible only if GPU type is selected).

Git Integration

• Git Project: Select a Git project from the drop-down.

• Git Branch: Select a branch (supported: main, migration, version).

• Sync Git repo at project creation:

  • Enable this option to automatically sync the project with Git.

  • Projects created with Git sync are labeled as Repo Sync Projects in the Project List with a branch icon.

Please note: If Git sync is not enabled, you can still configure Git access later by specifying the Git repository and branch. These projects display the branch icon without the drop-down selector.

Nodepool

• Select a Nodepool option for efficient execution of workloads.

Save and Confirm

1. After completing all fields, click Save.

2. A confirmation message appears.

3. The newly created project is added to the Project List page.

Viewing Project Details

• Click the new project entry in the Project List page.

• A panel on the right displays Project Details, Configurations, and External Libraries.

• You have successfully created a new DSLab project.

Notes

• Repo Sync Projects:

  • Appear in the Project List with a branch icon.

  • Git branch selection is enabled.

• Non-Repo Sync Projects (Git Access enabled):

  • Also display the branch icon, but without a branch selector.