gnes.encoder.numeric.pooling module

class gnes.encoder.numeric.pooling.PoolingEncoder(pooling_strategy='REDUCE_MEAN', backend='numpy', *args, **kwargs)[source]

Bases: gnes.encoder.base.BaseNumericEncoder

encode(data, *args, **kwargs)[source]
masked_reduce_max(x, m)[source]
masked_reduce_mean(x, m, jitter=1e-10)[source]
minus_mask(x, m, offset=1e+30)[source]
mul_mask(x, m)[source]
post_init()[source]

Declare class attributes/members that can not be serialized in standard way

train(*args, **kwargs)

Train the model, need to be overrided