espnet2.enh.layers.dcunet.get_activation

source

espnet2.enh.layers.dcunet.get_activation(name)

Get the activation function based on the provided name.

This function returns the corresponding activation function class from PyTorch based on the input string. The following activation functions are supported:

• ‘silu’: SiLU (Sigmoid Linear Unit)
• ‘relu’: ReLU (Rectified Linear Unit)
• ‘leaky_relu’: Leaky ReLU

• Parameters:name (str) – The name of the activation function to retrieve.
• Returns: The corresponding activation function class.
• Return type: Callable
• Raises:
  • NotImplementedError – If the provided name does not match any
  • supported activation function. –

Examples

>>> activation = get_activation("relu")
>>> print(activation)  # Output: <class 'torch.nn.modules.activation.ReLU'>

>>> activation = get_activation("silu")
>>> print(activation)  # Output: <class 'torch.nn.modules.activation.SiLU'>

>>> activation = get_activation("unknown")
Traceback (most recent call last):
    ...
NotImplementedError: Unknown activation: unknown