The SpeechBrain project aims to build a novel speech toolkit fully based on PyTorch. With SpeechBrain users can easily create speech processing systems, ranging from speech recognition (both HMM/DNN and end-to-end), speaker recognition, speech enhancement, speech separation, multi-microphone speech processing, and many others.

Voice Activity Detection based on Deep Learning & TensorFlow

End-to-End speech recognition implementation base on TensorFlow (CTC, Attention, and MTL training)

Pytorch and TensorFlow data loaders for several audio datasets

Python implementation of pre-processing for End-to-End speech recognition

SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition

End-to-End Speech Recognition Using Tensorflow

Final project for the Speaker Recognition course on Udemy, 机器之心, 深蓝学院 and 语音之家

End-to-End Speech Recognition using Neural Networks.

Quartznet implementation on pytorch [https://arxiv.org/abs/1910.10261]

Automatic speech recognition using neural networks

Pytorch implementation of conformer with with training script for end-to-end speech recognition on the LibriSpeech dataset.

Speech Recognition Using Tensorflow

Baselines for the Zero-Resources Speech Challenge using VisuallyGrounded Models of Spoken Language, 2021 edition

Some approaches based on deep learning to build the acoustic model for an end-to-end automatic speech recognition (ASR) pipeline.

A web-app/library for transcribing speech

Few-shot learning experiments mostly on speaker recognition.

Scripts to generate the reverberant LibriCHiME-5 dataset.

Gender Classification of the speaker from LibriSpeech Dataset

Replication of Jasper speech-to-text network using Intel optimized TensorFlow.