Gastron/asr-crdnn-librispeech

The repository provides all the necessary tools to perform automatic speech-recognition from an end-to-end system pretrained on LibriSpeech (EN) within SpeechBrain . The ASR system is composed with 3 different but linked blocks: The unigram (unigram) that transforms words into subword units and trained with the train transcriptions of LibiSpeech . The CRDNN architecture is made of blocks of convolutional neural networks with normalisation and pooling on the.frequency domain . The SpeechBrain EncoderDecoderASR() class also provides an easy way to encode the speech signal without running the decoding phase by calling the.encode_batch,