MaskedLinear

MaskedLinear#

class brainpy.dnn.MaskedLinear(conn, weight, mask_fun=Identity0(mode=NonBatchingMode), sharding=None, mode=None, name=None)[source]#

Synaptic matrix multiplication with masked dense computation.

It performs the computation of:

\[y = x @ M\]

where \(y\) is the postsynaptic value, \(x\) the presynaptic value, \(M\) the synaptic weight using a dense matrix.

>>> import brainpy as bp
>>> l = bp.dnn.MaskedLinear(bp.conn.FixedProb(0.1, pre=100, post=100),
>>>                         weight=0.1)

Parameters:

conn (TwoEndConnector) – The connection.
weight (Union[float, TypeVar(ArrayType, Array, Variable, TrainVar, Array, ndarray), Callable]) – Synaptic weights. Can be a scalar, array, or callable function.
mask_fun (Callable) – Masking function.
sharding (Union[Sequence[str], Device, Sharding, None]) – The sharding strategy.
mode (Optional[Mode]) – The synaptic computing mode.
name (Optional[str]) – The synapse model name.