Allow metadata to be saved with custom language models

We need to create and manage custom language models for a potentially large number of different clients.

We would like to be able to save metadata with a custom language model that would help us to identify things like what client account the custom language model is for.

This metadata could be simple arbitrary key, value pairs saved with a custom language model at create time, and retrieved with the custom language model when a get is done.

For example:

 request = { "name": model_name,"base_model_name": base_model,"description" : model_name,
"metadata" : { "customer_id": "123456",
               "type": "financial",
                "billing_id" : "7865421"} }


requests.post(uri, auth=(username,password), verify=False, headers=headers, data=json.dumps(request).encode('utf-8'))

Currently we need to maintain this sort of data (association between a custom language model and the customer that it was created for) externally (e.g. in a database table), which needs to be kept synchronized with the actual custom language models.

Additionally, it would be nice to be able to retrieve custom language models by name or by metadata (though this is not essential since it's always possible to perform a GET to get all custom language models and then filter on the client side).

Post comment

By clicking the "Post Comment" or "Submit Idea" button, you are agreeing to the IBM Ideas Portal Terms of Use.
Do not place IBM confidential, company confidential, or personal information into any field.

Please enter your email address

RELATED IDEAS

Allow metadata to be saved with custom language models