Good evening all,
I’m attempting to create a DataBunch based on the information listed in the data_block docs but running into the following warning and error.
Warning:
UserWarning: It’s not possible to collate samples of your dataset together in a batch.
Error:
TypeError: default_collate: batch must contain tensors, numpy arrays, numbers, dicts or lists; found <class ‘pathlib.PosixPath’>
I’m using the following code to create the DataBunch.
data = (ItemList.from_folder(path, presort=True)
.split_by_folder()
.label_from_folder()
.databunch())
The files are of the .bin extension and contain raw byte data. Do I need to construct a custom ItemList subclass or collate function to get something like this to work?
Best regards,
Kahless