Privately Hosted Embedding Models
Last updated
Last updated
When configuring a connection to a privately hosted embedding model, it’s essential to refer to your own system’s documentation and ensure that you have access to the necessary details. This setup typically involves connecting to an endpoint you host, and you’ll need to provide the required fields to securely integrate it.
When configuring your privately hosted embedding model, you will need to provide the following details:
Name Field Type: Text Description: Enter a unique name for this privately hosted connection configuration. This helps identify the connection when managing multiple configurations.
Embedding Provider Field Type: Dropdown Description: Select the Privately Hosted option.
Embedding Model Field Type: Text Description: Enter the name of the embedding model you are using. You can specify a custom model name or the one you have hosted on your system.
Description Field Type: Text (optional) Description: Add a manual description of the embedding model. You can enter any relevant details to describe the model or its use case, or you may use the "Generate description" option to auto-generate a description based on the model you’ve selected.
Dimension Field Type: Numeric Description: Enter the number of dimensions for the embedding model. The dimension size should correspond to the model you are using. Make sure to check your system's documentation to determine the correct dimension value.
Endpoint URL Field Type: URL Description: Enter the endpoint URL where your embedding model is hosted. This URL points to the location where requests will be sent to interact with the model. Ensure this is correctly set up in your system.
API Key Field Type: Text Description: Enter your API Key for authentication. This key is necessary for securely accessing the embedding model on your privately hosted system. Be sure to store it securely and avoid sharing it publicly.
Additional Configuration Params
Field Type: Key-Value Pairs
Description: Enter any additional configuration parameters required for your specific setup. These could be custom settings or parameters needed by your system, such as timeouts, custom headers, or other optional settings. You can enter them in the format key=value
, for more than one value, enter the first then put in the next value.
Here’s an example of how the configuration might look when filled in:
Name
My_Private_Embedding_Connection
Embedding Provider
Privately Hosted
Embedding Model
custom-embedding-model-v1
Description
This model is optimized for processing specialized text data.
Dimension
1024
Endpoint URL
https://myserver.com/api/embedding
API Key
abc123xyz456secureapikey
Additional Configuration Params
timeout=30,retries=5
Once you've filled in the necessary details, click Save or Test Connect to establish the connection to your privately hosted embedding model. Make sure the endpoint URL is accessible and that the API key and configuration parameters are valid.
Refer to Your System's Documentation: Make sure to check your system or server's documentation for specific setup details, especially for the endpoint URL and API key.
Security: Be sure to keep your API key and other sensitive information secure. Do not expose these details in public repositories or unsecured environments.