Hi, currently I have implemented following Dataset class that accepts input and output data for creating batches for training.
class My_dataset(Dataset):
def __init__(self, data_in, data_out):
self._data_in = data_in
self._data_out = data_out
def __len__(self):
return len(self._data_in)
def __getitem__(self, index):
return self._data_in[index], self._data_out[index]
Now I would like to add a few helper variables that has the same length with data to improve training results and they should also be split into same batches with input and output data. Does anyone know if it is possible to create a subclass of Dataset that accepts variable number of arrays in the constructor (e.g. it could be called as My_dataset(data_in, data_out, helper_1)
, My_dataset(data_in, data_out, helper_1, helper_2)
, etc) and generates batches accordingly?
Thank you!