PyTorch wrapper for DeepSpeech PaddlePaddle Beam Search Decoder. Args: labels (list): The tokens/vocab used to train your model.

Class using for maintain different chunks of data in one beam algorithm corresponding to one unique source. Note: after using State you shoul

Base encoder class

PyTorch wrapper for DeepSpeech PaddlePaddle Beam Search Decoder with interface for online decoding. Args: labels (list): The tokens/v

Struct for the beam search output, containing the tokens based on the vocabulary indices, and the timesteps * for each token in the beam search outpu

Trie tree for prefix storing and manipulating, with a dictionary in * finite-state transducer for spelling correction. */

Extract random crop of the video. Args: size (sequence or int): Desired output size for the crop in format (h, w). crop_posit

Resize video bysoomingin and out. Args: rate (float): Video is scaled uniformly between [1 - rate, 1 + rate]. interp

Rotate entire clip randomly by a random angle within given bounds Args: degrees (sequence or int): Range of degrees to select from

Resize video bysoomingin and out. Args: rate (float): Video is scaled uniformly between [1 - rate, 1 + rate]. interp

Implement a callback to retrive the dictionary of language model.

External scorer to query score for n-gram or sentence, including language * model scoring and word insertion. * * Example: * Scorer scorer(alp