jaxsnn.event.hardware.transforms

Functions

jaxsnn.event.hardware.transforms.linear_saturating(weight: jax.Array, scale: float, min_weight: float = - 63.0, max_weight: float = 63.0, as_int: bool = True) → jax.Array

Scale all weights according to:

w <- clip(scale * w, min_weight, max_weight)

Parameters

• weight – The weight array to be transformed.

• scale – A constant the weight array is scaled with.

• min_weight – The minimum value, smaller values are clipped to after scaling.

• max_weight – The maximum value, bigger values are clipped to after scaling.

• as_int – Round to nearest int and return as int type.

Returns

The transformed weight tensor.