Currently, we load the GRNN model when the surprisal module is loaded. It is not always necessary and we need to specify which version of the model to use.
As developers, we need to make sure irrelevant data are not loaded when we are just importing functions from the surprisal library (i.e. for analysis), and make it easy to change the path to the model we are using.
Currently, we load the GRNN model when the surprisal module is loaded. It is not always necessary and we need to specify which version of the model to use.
As developers, we need to make sure irrelevant data are not loaded when we are just importing functions from the surprisal library (i.e. for analysis), and make it easy to change the path to the model we are using.