/tf_words_classifier_with_cnn

this project is a simple demonstration of the convolution neural network trained to classify a 1-second audio segment of audio and output a result of ~94%

Primary LanguageJupyter Notebook

TensorFlow audio classifier Using CNN Model Architecture

this project is a simple demonstration of the convolution neural network trained to classify a 1-second audio segment of audio and output a result of 94% or so

How to run the repo

first of all, you need to install TensorFlow as :

pip install tensorflow

then run the provided notebook in the directory in the command line:

jupyter notebook

and open load_train_audio_clf.ipynb file

NOTE: as of now the actual preprocessed dataset is in my drive account (size of 600 mb)

if you wish to download here is a link Go to the Dataset

then run each cell to train and evaluate the model accuracy