nnsvs.multistream.multi_stream_mlpg

nnsvs.multistream.multi_stream_mlpg(inputs, variances, windows, stream_sizes=None, has_dynamic_features=None, streams=None)[source]

Split streams and do apply MLPG if stream has dynamic features

Parameters:

inputs (array like) – input 3-d or 2-d array
variances (array like) – variances of input features
windows (list) – windows for parameter generation
stream_sizes (list) – stream sizes
has_dynamic_features (list) – binary flags that indicates if steams have dynamic features
streams (list, optional) – Streams of interests. Returns all streams if streams is None. Defaults to None.

Raises:

RuntimeError – if stream sizes are wrong

Returns:

generated static features

Return type:

array like