Feature Details#
The following are more details about the pages in NeMo Studio.
Projects#
The Projects page serves as the central hub for organizing your AI development workflow. In each project, you can manage datasets, models, and jobs. The resources are scoped to the project and not shared with other projects.
Key Features:
Create and manage projects
Search and filter projects by name
View project metadata (creation date, last modified)
Organize datasets, models, and jobs within projects
Learn More: Projects
Datasets#
The Datasets page provides a centralized interface for managing all your data files used in customization, evaluation, and other workflows.
Key Features:
Upload and download datasets
View dataset contents and metadata
Manage dataset versions
Filter and search datasets by type and attributes
Learn More: Datasets
Models#
The Models page provides a playground interface for testing and experimenting with models. You can try out different models, configure system prompts, and provide input-output examples for prompt-tuning before committing to full model fine-tuning or deployment.
Key Features:
Select and test base models or custom models
Configure system prompts to define model behavior
Add input-output examples for prompt tuning (few-shot learning)
Configure external tools for enhanced model capabilities
Adjust hyperparameters (temperature, maximum tokens)
Real-time interaction and response testing
Learn More: Models
Customization#
The Customizations page enables you to create and manage model fine-tuning jobs through an intuitive interface.
Key Features:
Create new customization jobs
Select base models and training datasets
Configure customization parameters such as learning rate, epochs, and batch size
Monitor job progress and status
View training logs and metrics
Access resulting custom models
Learn More: Customizations
Evaluation#
The Evaluations page allows you to assess model performance on various benchmarks and custom datasets.
Key Features:
Create evaluation jobs
Select models and evaluation datasets
Configure evaluation metrics and parameters
Monitor job progress and completion
View detailed results and scores
Compare performance across models
Learn More: Evaluations