Together.ai

Integrating Together AI with Lakehousecat brings robust AI models from a leading provider into your workspace. Together AI is considered experimental due to the variety of models available, meaning some may function optimally while others might not meet anticipated performance levels.

Prerequisites

Before you proceed with the integration, ensure that you have:

• Administrator access to the Lakehousecat workspace.
• A together.ai account and API key.
• Access to the Together AI documentation for model IDs and API details.

Experimental Notice

Given the experimental nature of the Together AI integration, it's important to understand that not all models have undergone extensive testing. Performance can vary depending on the model selected.

Set Up the Together.ai

1. Navigate to the Admin Workspace

   • Access your Lakehousecat admin workspace and go to the 'Models' section.

2. Select Provider Models

   • Click on 'Provider Models' to explore different model providers.

3. Choose the Provider

   • Click the "+" icon to add a new provider and select "Together AI".

4. Model Type Selection

   • Select the desired model type from Together AI. Options include Chat, Embedding, Text-to-Speech, and Speech-to-Text. For this example, choose "Chat".

Configuration Details

• Configuration Name
  • Input a meaningful and user-specific name for your configuration. This facilitates easy identification and management.
• Together AI API Key
  • Enter your Together AI API key, necessary for authentication and access.
• Model ID
  • Provide the model ID as given in the Together AI documentation, ensuring it aligns with your selected model type.

Save the Provider Configuration

Once all information is correctly entered:

1. Save Configuration
   • Click "Save" to store your provider configuration within the Lakehousecat platform.

Best Practices

• Naming Conventions
  • Use consistent naming conventions to improve manageability and clarity for users accessing different models.
• Documentation Checking
  • Continually refer to Together AI documentation for updates and details on API key acquisition and model IDs.

Troubleshooting

Due to the experimental nature of the Together AI integration:

• Model Performance
  • Be prepared for variable performance across different models. Thorough testing of configurations is recommended to ensure functionality meets expectations.

Conclusion

The integration of Together AI entails diverse model options, though users should evaluate model compatibility and performance individually within the Lakehousecat environment. For further specifics on model configuration and troubleshooting, please review the dedicated sections on model management.