基于PyTorch的通用语音工具包, 支持语音识别、语音前端、说话人识别、情感识别、关键词识别、口语理解等多种语音任务和数据集 Development toolkits for multiple speech tasks including ASR, Speech Front-end, Speaker Recognition, Keyword Spotting, Spoken Language Understanding... and corresponding recipes, built upon PyTorch.
噪声数据, 可用于语音增强、有噪语音分离等任务(配合wsj0-2mix、LibriMix中的语音数据) noise data for tasks including Speech Enhancement, Speech Separation under noisy environment...(normally used with speech audio from wsj0-2mix or LibriMix)
基于PyTorch Lightning的深度学习语音前端工具包, 支持语音增强/语音分离/多模态/多通道等语音前端任务和数据集 Toolkits for deep-learning-based speech front-end development, built upon PyTorch Lightning, suports for various recipes and tasks including Speech Enhancement, Speech Separation, Multi-modal, Mulit-channel...
支持VoxCeleb、CN-Celeb等数据集, 支持TDNN、ResNetSE34、ECAPA-TDNN等网络 Support various recipes(VoxCeleb/CN-Celeb) and network architectures(TDNN/ResNet34/ECAPA-TDNN)
支持VoxCeleb, 支持VGGVox、ResNetSE34等网络, 支持多分类、度量学习等训练方式 Support VoxCeleb, support networks including VGGVox and ResNetSE34, support training methods including multi-class classification and metric learning