friday27 / KeywordSpotting
Licence: other
Train a 4-layer Convolutional Neural Network to detect trigger word
Stars: ✭ 49
Programming Languages
python
139335 projects - #7 most used programming language
Projects that are alternatives of or similar to KeywordSpotting
keras tfrecord
Extending Keras to support tfrecord dataset
Stars: ✭ 61 (+24.49%)
Mutual labels: tfrecords
keyword-transformer
Official implementation of the Keyword Transformer: https://arxiv.org/abs/2104.00769
Stars: ✭ 76 (+55.1%)
Mutual labels: keyword-spotting
awesome-keyword-spotting
This repository is a curated list of awesome Speech Keyword Spotting (Wake-Up Word Detection).
Stars: ✭ 150 (+206.12%)
Mutual labels: keyword-spotting
tensorflow object detection helper tool
tensorflow object detection api helper tool ( custom object detection )
Stars: ✭ 30 (-38.78%)
Mutual labels: tfrecords
multilingual kws
Few-shot Keyword Spotting in Any Language and Multilingual Spoken Word Corpus
Stars: ✭ 122 (+148.98%)
Mutual labels: keyword-spotting
Porcupine
On-device wake word detection powered by deep learning.
Stars: ✭ 2,606 (+5218.37%)
Mutual labels: keyword-spotting
Speech-Command-Recognition-with-Capsule-Network
Speech command recognition with capsule network & various NNs / KWS on Google Speech Command Dataset.
Stars: ✭ 20 (-59.18%)
Mutual labels: keyword-spotting
Multi-Hotword Spotting
Won't it be cool to build a speech assistant like Alexa or Siri yourself without voice API and network connection?
Stars: ✭ 31 (-36.73%)
Mutual labels: keyword-spotting
Tensorflow template application
TensorFlow template application for deep learning
Stars: ✭ 1,851 (+3677.55%)
Mutual labels: tfrecords
TensorFlow-Input-Pipeline
TensorFlow Input Pipeline Examples based on multi-thread and FIFOQueue
Stars: ✭ 54 (+10.2%)
Mutual labels: tfrecords
EasySparse
Sparse learning in TensorFlow using data acquired from Spark.
Stars: ✭ 21 (-57.14%)
Mutual labels: tfrecords
# KeywordSpotting Keyword Spotting using Convolutional Neural Network Reference: Convolutional Recurrent Neural Networks for Small-Footprint Keyword Spotting https://arxiv.org/abs/1703.05390 stream.py - record .wav files in 3 seconds - save in "data/name" - the format of filename: 20171231_NAME_0000_0.wav (the last number = 0, without keyword) 20171231_NAME_0012_1.wav (the last number = 1, with keyword) create_catalog.py - read all .wav in data/ and create 2 catalogs of train.csv and validaiotn.csv mfcc_tfrecord.py - convert each .wav file in the certain catalog and save as .tfrecords - enhance training efficiency train_nontfrecord.py - train model with raw .wav files - save training result in graph/ - save model in ckpt/ train_tfrecord.py - train model with .tfrecords - save training result in graph/ - save model in ckpt/ inference.py - edit line 21-22 to your own checkpoint(.meta) path - continueously create(record) a test.wav(3 seconds) and do inference
Note that the project description data, including the texts, logos, images, and/or trademarks,
for each open source project belongs to its rightful owner.
If you wish to add or remove any projects, please contact us at [email protected].