Easy-to-use Speech Toolkit including Self-Supervised Learning model, SOTA/Streaming ASR with punctuation, Streaming TTS with text frontend, Speaker Verification System, End-to-End Speech Translation and Keyword Spotting. Won NAACL2022 Best Demo Award.

A collection of resources to make a smart speaker

building blocks to create voice interface applications

Few-shot Keyword Spotting in Any Language and Multilingual Spoken Word Corpus

Wav2Keyword is keyword spotting(KWS) based on Wav2Vec 2.0. This model shows state-of-the-art in Speech commands dataset V1 and V2.

In this repository, we explore using a hybrid system consisting of a Convolutional Neural Network and a Support Vector Machine for Keyword Spotting task.

Cortex M KWS example with Tengine Lite.

🦁 Nala is an agile open-source voice assistant framework (20+ actions).

Tensorflow implementation of "Small-Footprint Keyword Spotting with Multi-Scale Temporal Convolution"(INTERSPEECH 2020)

Keyword Search Recipe for Subword ASR

Speech command recognition with capsule network & various NNs / KWS on Google Speech Command Dataset.

Keyword spotting for audio with attention (KWS model for audio)

Models and training scripts for "LSTMs for Keyword Spotting with ReRAM-based Compute-In-Memory Architectures" (ISCAS 2021).

Recognition of 38 speech commands in russian. Based on Yandex Cup 2021 ML Challenge: ASR

一个讯飞智能语音平台 MSC 的第三方 Python SDK，支持语音唤醒、语音识别、语音合成、语音评测等功能。A third-party Python SDK for a iFLYTEK MSC. Using for ASR, TSS, KWS.

The project is a concoction of research (audio signal processing, keyword spotting, ASR), development (audio data processing, deep neural network training, evaluation) and deployment (building model artifacts, web app development, docker, cloud PaaS) by integrating CI/CD pipelines with automated tests and releases.

Keyword spotting for XMC4000 boards, based on the ML-KWS-for-MCU from ARM

A KWS model trained on SpeechCommands dataset, written in PyTorch.

Attention Implementation for KWS