Data loading problem

I would like to run RAM model on my own data set. My training set is train.t7 . I have encountered a problem while loading my data set because it is my first time that I have used python. I have tried very hard to understand both the software and package but I need your help in solving this problem.
I run the following code:

    o=torchfile.load('train.t7'

    )

    x= tf.convert_to_tensor(o[b'data'],dtype=tf.uint8)

    y=tf.convert_to_tensor(o[b'labels'],dtype=tf.uint8)

    x=tf.reshape(x,[50000,784])

    dataset = torch.utils.data.TensorDataset(x, y)

But then I got the following error:


`
File "/home/wing92518/.local/lib/python3.6/site-packages/torch/utils/data/dataset.py", line 36, in <genexpr>
    assert all(tensors[0].size(0) == tensor.size(0) for tensor in tensors)
AttributeError: 'Tensor' object has no attribute 'size'
`

pytorch tensor and tf tensors are different things…