Projects#

Projects are the primary organizational unit in NeMo Studio. They group related resources such as datasets, models, customization jobs, and evaluation results.

The Projects page provides a centralized interface for managing your projects.

Backend Microservices#

In the backend, the UI communicates with NeMo Entity Store to manage project entities.

Projects Page UI Overview#

The following are the main components and features of the Projects page.

Project Listing#

The Projects page displays your projects in a table format with the following columns:

  • Name: The unique name for your project.

  • Description: Optional text describing the project’s purpose.

  • Created: Timestamp showing when the project was created.

  • Last Modified: Timestamp showing the most recent update to the project.

You can sort projects by clicking column headers to organize your view.

Project Management#

You can perform the following actions on the Projects page:

  • Create New Project: Set up new projects to organize your AI development work.

  • Search and Filter: Find projects by name or apply filters.

Resources in a Project#

You can view the following pages in a project by clicking on the project name.

  • Datasets: Upload, view, and manage datasets for training, evaluation, and testing. Refer to Datasets for more information.

  • Models (Playground): Test models interactively with system prompts and input-output examples for prompt tuning. Refer to Models for more information.

  • Customizations: Create and monitor model fine-tuning jobs through an intuitive interface. Refer to Customizations for more information.

  • Evaluations: Assess model performance using the Evaluation page. Refer to Evaluations for more information.

Considerations#

  • Resources in a project are not shared across different projects.

  • Resources created using the Python SDK or the REST APIs without the project parameter specified are not discoverable in the NeMo Studio Projects page. To make them discoverable, create a project and add the project name to the resource creation or update requests. The following is an example of creating a dataset in a project using the SDK:

    from nemo_microservices import NeMoMicroservices

client = NeMoMicroservices(
base_url="http://nemo.test",
inference_base_url="http://nim.test",
)

project = client.projects.create(
    name="your-project-name",
    description="Example project",
    ... # Other project creation parameters
)

dataset = client.datasets.create(
    name="your-dataset-name",
    namespace="default",
    description="Example dataset",
    files_url="hf://datasets/default/sample-basic-test",
    project=project.name
    ... # Other dataset creation parameters
)