PyTorch 张量

Question

I am new to PyTorch. Can anyone explain to me why we are supposed to convert the data frame to a torch tensor, before performing Regression? (I've split the dataset, as you can see). And I also need an explanation to this piece of code:

x_train_tensor = torch.tensor(x_train.values.reshape(-1, 1), dtype = torch.float)
y_train_tensor = torch.tensor(y_train.values.reshape(-1, 1), dtype = torch.float)

x_test_tensor = torch.tensor(x_test.values.reshape(-1, 1), dtype = torch.float)
y_test_tensor = torch.tensor(y_test.values.reshape(-1, 1), dtype = torch.float)

Thank you!

Answer 1

All Pytorch operations are defined on Tensor s which are multi-dimensional matrixes in essence. So you need to be using tensors if you want to use Pytorch.

Answer 2

All operations or methods in pytorch libraries are performed on tensors only.

Here in the above code you are first converting your tensor to a column vector using .reshape(-1,1)..... it's a way of telling that -1 is where the compiler automatically decided the value and then these are converted to tensors.