PyTorch 06: a range and tensors like
Using torch.arange(), torch.zeros_like() and torch.ones_like()
Sometimes you might want a range of numbers and for that you can use torch.arange(start, end, step):
# Create a range of numbers from 1 to 9 with a step of 1
range = torch.arange(start = 1, end = 10, step = 1)
range
tensor([1, 2, 3, 4, 5, 6, 7, 8, 9])
# You can simplify the call by omitting parameter names
torch.arange(1, 10, 1)
Let’s move on to another concept called tensor like.
Sometimes you might want a tensor full of zeros or ones with the same shape as some other tensor and to do so you can make use of torch.zeros_like(input) or torch.ones_like(output).
# Create a tensor of zeros with the same shape as 'range'
like_tensor_zeros = torch.zeros_like(range)
# Create a tensor of ones with the same shape as 'range'
like_tensor_ones = torch.ones_like(range)
# Display tensors of zeros and ones
like_tensor_zeros, like_tensor_ones
(tensor([0, 0, 0, 0, 0, 0, 0, 0, 0]), tensor([1, 1, 1, 1, 1, 1, 1, 1, 1]))