Exception not handled in train.py when --val_set is not provided

[djshen]

simpleNN/Python/train.py

Lines 305 to 308 in ad58af4

	full_batch[0], mean_tr = normalize_and_reshape(full_batch[0], dim=config.dim, mean_tr=None)
	val_batch[0], _ = normalize_and_reshape(val_batch[0], dim=config.dim, mean_tr=mean_tr)

If --val_set is not provided, val_data will be set to None, which will cause an exception like:

TypeError: 'NoneType' object is not subscriptable

This case is handled in the gradient_trainer function.

[quanpr]

Hi,

Thanks for your interest in trying our work. The exception is now handled. For your information, if no validation set is provided during the training phase, we will store the model in the last iteration for future prediction.

Best,
Pengrui