Building blocks

📄️ Connect to superduper

Note that this is only relevant if you are running superduper in development mode.

📄️ Create datatype

SuperduperDB supports automatic data conversion, so users don’t need to worry about the compatibility of different data formats (PIL.Image, numpy.array, pandas.DataFrame, etc.) with the database.

📄️ Get useful sample data

📄️ Insert data

In order to create data, we need to create a Schema for encoding our special Datatype column(s) in the databackend.

📄️ Compute features

📄️ Build text embedding model

📄️ Build image embedding model

Construct a neural network architecture to project high-dimensional image data into a lower-dimensional, dense vector representation

📄️ Build multimodal embedding models

Some embedding models such as CLIP come in pairs of model and compatible_model.

📄️ Build LLM

📄️ Create vector-index

📄️ Perform a vector search

Once we have this search target, we can execute a search as follows:

📄️ Connecting listeners

📄️ Build and train classifier