gnes.encoder.numeric.tf_pq module

class gnes.encoder.numeric.tf_pq.TFPQEncoder(num_bytes: int, cluster_per_byte: int = 255, *args, **kwargs)[source]

Bases: gnes.encoder.numeric.pq.PQEncoder

batch_size = 8192
close()[source]

Release the resources as model is destroyed

encode(vecs: numpy.ndarray, *args, **kwargs) → numpy.ndarray[source]
post_init()[source]

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

classmethod pre_init()[source]
train(vecs: numpy.ndarray, *args, **kwargs)