In this PR, we are moving the label_col argument in the load_data() function to be optional, it allows the usage of this function in training and inference use cases. The TorchTabularTextDataset() accepts the labels as None.
We are also adding the class_weights argument, which is accepted in the TorchTabularTextDataset().
In this PR, we are moving the
label_col
argument in theload_data()
function to be optional, it allows the usage of this function in training and inference use cases. TheTorchTabularTextDataset()
accepts thelabels
as None.We are also adding the
class_weights
argument, which is accepted in theTorchTabularTextDataset()
.