gnes.encoder.video.incep_mixture module

class gnes.encoder.video.incep_mixture.IncepMixtureEncoder(model_dir_inception: str, model_dir_mixture: str, select_layer: str = 'PreLogitsFlatten', use_cuda: bool = False, feature_size: int = 300, vocab_size: int = 28, cluster_size: int = 256, method: str = 'fvnet', input_size: int = 1536, vocab_size_2: int = 174, max_frames: int = 30, multitask_method: str = 'Attention', *args, **kwargs)[source]

Bases: gnes.encoder.base.BaseVideoEncoder

batch_size = 64
encode(data: List[np.ndarray], *args, **kwargs) → numpy.ndarray[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