YouTip

# PyTorch torch.empty_like Function The `torch.empty_like` function is used to create an uninitialized tensor with the same shape and data type as the input tensor. ### Syntax ```python torch.empty_like(input, dtype=None, layout=None, device=None, requires_grad=False, memory_format=torch.preserve_format) ``` ### Parameters | Parameter | Description | | --- | --- | | input (Tensor) | Input tensor, the output tensor will have the same shape as this tensor. | | dtype (torch.dtype, optional) | Desired data type of the output tensor. If None, defaults to the dtype of the input tensor. | | layout (torch.layout, optional) | Desired layout of the output tensor. If None, defaults to the layout of the input tensor. | | device (torch.device, optional) | Desired device of the output tensor. If None, defaults to the device of the input tensor. | | requires_grad (bool, optional) | If True, the output tensor requires gradient computation. Default is False. | | memory_format (torch.memory_format, optional) | Desired memory format of the output tensor. Default is torch.preserve_format. | ### Return Value Returns an uninitialized tensor with the same shape as the input tensor. ### Example ```python import torch # Create a tensor x = torch.tensor([[1, 2, 3], [4, 5, 6]]) # Create an uninitialized tensor with the same shape as x y = torch.empty_like(x) print("Original tensor x:") print(x) print(" Uninitialized tensor y:") print(y) ``` Output result: ``` Original tensor x: tensor([[1, 2, 3], [4, 5, 6]]) Uninitialized tensor y: tensor([[ 0, 0, 0], [ 0, 0, 0]]) ``` ### Notes 1. The `torch.empty_like` function does not initialize the values of the tensor, so the values in the output tensor are undefined. 2. If you need a tensor initialized with zeros, use `torch.zeros_like`. 3. If you need a tensor initialized with ones, use `torch.ones_like`. 4. If you need a tensor initialized with a specific value, use `torch.full_like`. ### Related Functions * `torch.empty`: Creates an uninitialized tensor with a specified shape. * `torch.zeros_like`: Creates a tensor filled with zeros with the same shape as the input. * `torch.ones_like`: Creates a tensor filled with ones with the same shape as the input. * `torch.full_like`: Creates a tensor filled with a specified value with the same shape as the input. * `torch.rand_like`: Creates a tensor filled with random numbers from a uniform distribution with the same shape as the input. * `torch.randn_like`: Creates a tensor filled with random numbers from a normal distribution with the same shape as the input.