Hi, saw your comment on another issue, I'll review it later, but <class 'NoneType'> means you are feeding None as label, so there will be some issue with your preprocessing of data

from mltu.

I updated my code. In my local I change some line in dataProvider.py

Validate dataset

self._dataset = self.validate(dataset)
#if not skip_validation:
#self._dataset = self.validate(dataset)
#else:
#self.logger.info("Skipping Dataset validation...")
it shows "Validating Dataset". I think it is ok. In the colab, i cannot change it, I think it show error
I double check train_data_provider._dataset and val_data_provider._dataset, it has data
Although, there doesnt have any error when running in local, my model.fit still has error

from mltu.

@pythonlessons Do you have any update for my issue. Please, I need your help

from mltu.

Hi, you need to be more attentive when you refactor code and write it your self. You are using from mltu.torch.dataProvider import DataProvider when you should use from mltu.tensorflow.dataProvider import DataProvider with TensorFlow

from mltu.

I assume its solved

from mltu.