ludwig-ai / model-hub Goto Github PK

In addition to the fields currently tracked in the schema, we should add:

• The type of features that this model can be used for.
• The expected input and output structure that the model uses.

Write unit tests for the happy paths for the model-hub

We want to make sure that users can upload and retrieve Ludwig models (entire ECD pipelines) but also single encoders, combiners and decoders.

So we need to both allow for uploading of single es, cs and ds, as well as specifying that, from an uploaded ecd they only want to retrieve a specific piece.

For instance if I upload a model for email classification with an encoder for the meila text, and encoder for the sender email address and a date encoder, a combiner and a category decoder, another user should be able to retrieve just the encoder of the sender email address.