whestbench.
API

run_mlp

Forward pass returning final-layer activations.

function · source

run_mlp(mlp: 'MLP', inputs: 'fnp.ndarray') -> 'fnp.ndarray'
Forward pass returning final-layer activations.

Args:
    mlp: MLP to execute.
    inputs: Input matrix of shape ``(samples, mlp.width)``.

Returns:
    Activations of shape ``(samples, mlp.width)`` after the last layer.