PyTorch .empty()

The .empty() method creates a tensor with uninitialized data. This means that the tensor is allocated memory without setting its values, which may contain arbitrary data (such as NaNs or other undefined values). The shape of the tensor must be specified as an argument.

Syntax

torch.empty(size, out=None, dtype=None, layout=torch.strided, device=None, requires_grad=False, pin_memory=False, memory_format=torch.contiguous_format)

The parameters are as follows:

• size: Specifies the shape of the tensor. It can be an integer or a tuple of integers representing the dimensions
• out(Optional): The output Tensor, defaults to None.
• dtype(Optional): Specifies the desired data type of the tensor.
• layout(Optional): Specifies the layout (torch.layout) of the output tensor, defaults to torch.strided.
• device(Optional): Specifies the device (torch.device) of the output tensor, defaults to None.
• requires_grad(Optional): A boolean indicating whether autograd will record operations on the output tensor, defaults to False.
• pin_memory(Optional): A boolean indicating whether the tensor is allocated in pinned memory. This only works for CPU tensors. Defaults to False.
• memory_format(Optional): Specifies the memory format (torch.memory_format) of the output tensor, defaults to torch.contiguous_format.

Example

The example below uses the .empty() method:

import torch

t0 = torch.empty((1, 3))
print(t0)

Copy to clipboard
Copy to clipboard

The returned tensor is as follows:

tensor([1.245e+11, NaN, -2.956e-52])

Copy to clipboard
Copy to clipboard

Codebyte Example

Run the following code to see how the .empty() method works:

Visit us
Visit us
Hide code
Code
Output
Hide output
Copy to your clipboard
Copy Codebyte
Run