flax.linen.Dropout

class flax.linen.Dropout(rate, broadcast_dims=(), deterministic=None, parent=<flax.linen.module._Sentinel object>, name=None)[source]

Create a dropout layer.

Parameters
  • rate (float) –

  • broadcast_dims (Sequence[int]) –

  • deterministic (Optional[bool]) –

  • parent (Optional[Union[Type[flax.linen.module.Module], Type[flax.core.scope.Scope], Type[flax.linen.module._Sentinel]]]) –

  • name (str) –

Return type

None

rate

the dropout probability. (_not_ the keep rate!)

Type

float

broadcast_dims

dimensions that will share the same dropout mask

Type

Sequence[int]

deterministic

if false the inputs are scaled by 1 / (1 - rate) and masked, whereas if true, no mask is applied and the inputs are returned as is.

Type

Optional[bool]

__call__(inputs, deterministic=None)[source]

Applies a random dropout mask to the input.

Parameters
  • inputs – the inputs that should be randomly masked.

  • deterministic (Optional[bool]) – if false the inputs are scaled by 1 / (1 - rate) and masked, whereas if true, no mask is applied and the inputs are returned as is.

Returns

The masked inputs reweighted to preserve mean.

Methods