Model Metrics and Overview
The Model Dashboard provides several key metrics to monitor your models’ performance:- Active - Number of models currently deployed and ready for use
- Speed - Rate at which tokens are processed across your models
- Response - Average time taken to generate responses
- Success - Percentage of successful requests across all models
Search and Filtering
Use the search bar to quickly locate specific models by name. Filter your models by status:- All Models - Display all models regardless of status
- Ready - Show only models available for deployment
- Inline Tuned - View models that have been inline-tuned
- Training - Display models currently in training
- Archived - Show archived models
View Options
Toggle between two display modes to suit your preference:- Card view - Visual grid layout showing models as cards
- Table view - Compact tabular format for detailed information
Creating a Model
- Click the ‘Create Model’ button to begin creating your model.
- Run Through the Model Setup Wizard
- Enter a name for your model, provide a description, and optionally apply a persona, then click ‘Next’.
- Select the foundation model that best fits your needs.
- Upload training data to customize the model. Options include uploading new data via file or integration, or using an existing dataset. Click ‘Add Files’.
The screenshot provided uses a Hugging Face Integration dataset.
- Click ‘Next’ to finalize the model creation.
- Review your configuration details and click ‘Confirm & Train‘ to finalize the configuration and begin creating your model.
