from torch.nn import Parameter

The `Parameter` class in PyTorch is a subclass of the `Tensor` class, specifically designed for values that are to be learned during training. It is a wrapper around a `Tensor` object that is automatically marked as a parameter of the model when it is assigned as an attribute of a `Module` object.

Here's an example of how to use `Parameter`:

import torch

class MyModule(torch.nn.Module):
    def __init__(self):
        super(MyModule, self).__init__()
        self.weight = torch.nn.Parameter(torch.randn(10, 5))
        self.bias = torch.nn.Parameter(torch.zeros(10))

    def forward(self, x):
        y = torch.matmul(x, self.weight) + self.bias
        return y

In this example, `weight` and `bias` are both `Parameter` objects that will be learned during training. They are initialized with random values and zeros, respectively. The `forward` method uses these parameters to perform a linear transformation on the input `x`.

Note that you can also create a `Parameter` object from an existing `Tensor` by calling its `requires_grad_()` method:

x = torch.randn(10, 5)
w = torch.nn.Parameter(x)

This creates a `Parameter` object `w` that wraps the `x` tensor and indicates that it should be updated during training.