TeaPoly

CTC-OptimizedLoss

Computes the MWER (minimum WER) Loss with CTC beam search. Knowledge distillation for CTC loss.

Conformer-Athena

Dynamic Chunk Streaming and Offline Conformer based on athena-team/Athena.

speexdsp-ns-python

Python bindings of speexdsp noise suppression library

CE-OptimizedLoss

Computes the MWER (minimum WER) Loss with beam search and negative sampling strategy.

warp-ctc-crf

An extension of thu-spmi/CAT which contains a full-fledged implementation of CTC-CRF for Tensorflow.

PLCPA-ASYM-Loss

The power-law compressed phase-aware asymmetric (PLCPA-ASYM) loss

cat_tensorflow

Crf-based Asr Toolkit with TensorFlow implement

AIF-PyTorch

(NOT Official) Implementation Auto-regressive Integrate-and-Fire (AIF)

TeaPoly.github.io

matmulfreellm

Implementation for MatMul-free LM.

PercepNet

(Work In Progress) Unofficial implementation of PercepNet: A Perceptually-Motivated Approach for Low-Complexity, Real-Time Enhancement of Fullband Speech

SV-GMM

Speaker Verification using GMMs

warp-rnnt

CUDA-Warp RNN-Transducer with TensorFlow and PyTorch binding.

asr_frontend

PyTorch implementation of frontend, like PCEN (per-channel energy normalization) and Mel-Filterbank (mel-filterbank).

athena

an open-source implementation of sequence-to-sequence based speech processing engine

conv-tasnet

A PyTorch implementation of "Improving noise robust automatic speech recognition with single-channel time-domain enhancement network"

FunASR

A Fundamental End-to-End Speech Recognition Toolkit

icefall

nara_wpe

Different implementations of "Weighted Prediction Error" for speech dereverberation

NKF-AEC

Acoustic Echo Cancellation with Nerual Kalman Filtering

NSD-MS2S

CHIME-7 diarization champion system: neural speaker diarization using memory-aware multi-speaker embedding with sequence-to-sequence architecture

py-aec-unified2021

rir-configuration-generator

Generation of virtual rooms configurations.

self_attention_alignment

Deep model with built-in self-attention alignment for acoustic echo cancellation, Pytorch implement

SpectrumAugmenter

Performs data augmentation as according to the SpecAugment paper. Modified from Lingvo (TensorFlow > 1.10.0).

speechbrain

A PyTorch-based Speech Toolkit

torchdistance

Edit-distance PyTorch extension with Cpu and CUDA kernels

unified2021

A UNIFIED SPEECH ENHANCEMENT FRONT-END FOR ONLINE DEREVERBERATION, ACOUSTIC ECHO CANCELLATION, AND SOURCE SEPARATION

wenet

Production First and Production Ready End-to-End Speech Recognition Toolkit

you-get

:arrow_double_down: Dumb downloader that scrapes the web