drop_path

dragon.nn.drop_path(
  inputs,
  ratio=0.2,
  inplace=False,
  **kwargs
)[source]

Set the examples over the input to zero randomly. [Larsson et.al, 2016].

The DropPath function is defined as:

\[\text{DropPath}(x_{ij}) = x_{ij} * (r_{i} \sim \mathcal{B}(1, 1 - \text{ratio})) \]

Examples:

x = dragon.ones((5, 2), 'float32')
print(dragon.nn.drop_path(x, ratio=0.5))
Parameters:
  • inputs (dragon.Tensor) – The input tensor.
  • ratio (Union[float, dragon.Tensor], optional, default=0.2) – The probability to zero an example.
  • inplace (bool, optional, default=False) – Call in-place or return a new tensor.
Returns:

dragon.Tensor – The output tensor.