snn.initializers¶

All initializer functions take a shape tuple and an optional seed integer and return a NumPy array of that shape.

Zero / Constant¶

snn.initializers.zeros(shape)[source]¶

snn.initializers.ones(shape)[source]¶

Random¶

snn.initializers.random_normal(shape, mean=0.0, stddev=0.05, seed=None)[source]¶

snn.initializers.random_uniform(shape, minval=-0.05, maxval=0.05, seed=None)[source]¶

Variance Scaling¶

snn.initializers.glorot_uniform(shape, seed=None)[source]¶

snn.initializers.glorot_normal(shape, seed=None)[source]¶

snn.initializers.he_uniform(shape, seed=None)[source]¶

snn.initializers.he_normal(shape, seed=None)[source]¶

snn.initializers.lecun_uniform(shape, seed=None)[source]¶

snn.initializers.lecun_normal(shape, seed=None)[source]¶

Factory¶

snn.initializers.get(identifier)[source]¶