Name: Tensorflow music prediction
Rating: 3.4 (9141 reviews)
Author: victordalet

In this article, I show how to use tensorflow to predict a style of music.
In my example, I compare techno and classical music.

You can find the code on my github :
https://github.com/victordalet/sound_to_partition

I - Dataset

For the first step, you need to create one dataset forlder and inside add one folder for music style, for example i add one techno folder and classic folder in which in put my wav soung.

II - Train

I create a train file, with the arguments max_epochs to be completed.

Modify the classes in the constructor that correspond to your directory in the dataset folder.

In the loading and processing method, I retrieve the wav file from a different directory and obtain the spectogram.

For training purposes, I use the Keras convolutions and model.

import os
import sys
from typing import List

import librosa
import numpy as np
from tensorflow.keras.layers import Input, Conv2D, MaxPooling2D, Flatten, Dense
from tensorflow.keras.models import Model
from tensorflow.keras.optimizers import Adam
from sklearn.model_selection import train_test_split
from tensorflow.keras.utils import to_categorical
from tensorflow.image import resize



class Train:

    def __init__(self):
        self.X_train = None
        self.X_test = None
        self.y_train = None
        self.y_test = None
        self.data_dir: str = 'dataset'
        self.classes: List[str] = ['techno','classic']
        self.max_epochs: int = int(sys.argv[1])

    @staticmethod
    def load_and_preprocess_data(data_dir, classes, target_shape=(128, 128)):
        data = []
        labels = []

        for i, class_name in enumerate(classes):
            class_dir = os.path.join(data_dir, class_name)
            for filename in os.listdir(class_dir):
                if filename.endswith('.wav'):
                    file_path = os.path.join(class_dir, filename)
                    audio_data, sample_rate = librosa.load(file_path, sr=None)
                    mel_spectrogram = librosa.feature.melspectrogram(y=audio_data, sr=sample_rate)
                    mel_spectrogram = resize(np.expand_dims(mel_spectrogram, axis=-1), target_shape)
                    data.append(mel_spectrogram)
                    labels.append(i)

        return np.array(data), np.array(labels)

    def create_model(self):
        data, labels = self.load_and_preprocess_data(self.data_dir, self.classes)
        labels = to_categorical(labels, num_classes=len(self.classes))  # Convert labels to one-hot encoding
        self.X_train, self.X_test, self.y_train, self.y_test = train_test_split(data, labels, test_size=0.2,
                                                                                random_state=42)

        input_shape = self.X_train[0].shape
        input_layer = Input(shape=input_shape)
        x = Conv2D(32, (3, 3), activation='relu')(input_layer)
        x = MaxPooling2D((2, 2))(x)
        x = Conv2D(64, (3, 3), activation='relu')(x)
        x = MaxPooling2D((2, 2))(x)
        x = Flatten()(x)
        x = Dense(64, activation='relu')(x)
        output_layer = Dense(len(self.classes), activation='softmax')(x)
        self.model = Model(input_layer, output_layer)

        self.model.compile(optimizer=Adam(learning_rate=0.001), loss='categorical_crossentropy', metrics=['accuracy'])

    def train_model(self):
        self.model.fit(self.X_train, self.y_train, epochs=self.max_epochs, batch_size=32,
                       validation_data=(self.X_test, self.y_test))
        test_accuracy = self.model.evaluate(self.X_test, self.y_test, verbose=0)
        print(test_accuracy[1])

    def save_model(self):
        self.model.save('weight.h5')


if __name__ == '__main__':
    train = Train()
    train.create_model()
    train.train_model()
    train.save_model()

III - Test

To test and use the model, I've created this class to retrieve the weight and predict the style of the music.

Don't forget to add the right classes to the constructor.

from typing import List

import librosa
import numpy as np
from tensorflow.keras.models import load_model
from tensorflow.image import resize
import tensorflow as tf



class Test:

    def __init__(self, audio_file_path: str):
        self.model = load_model('weight.h5')
        self.target_shape = (128, 128)
        self.classes: List[str] = ['techno','classic']
        self.audio_file_path: str = audio_file_path

    def test_audio(self, file_path, model):
        audio_data, sample_rate = librosa.load(file_path, sr=None)
        mel_spectrogram = librosa.feature.melspectrogram(y=audio_data, sr=sample_rate)
        mel_spectrogram = resize(np.expand_dims(mel_spectrogram, axis=-1), self.target_shape)
        mel_spectrogram = tf.reshape(mel_spectrogram, (1,) + self.target_shape + (1,))

        predictions = model.predict(mel_spectrogram)

        class_probabilities = predictions[0]

        predicted_class_index = np.argmax(class_probabilities)

        return class_probabilities, predicted_class_index

    def test(self):
        class_probabilities, predicted_class_index = self.test_audio(self.audio_file_path, self.model)

        for i, class_label in enumerate(self.classes):
            probability = class_probabilities[i]
            print(f'Class: {class_label}, Probability: {probability:.4f}')

        predicted_class = self.classes[predicted_class_index]
        accuracy = class_probabilities[predicted_class_index]
        print(f'The audio is classified as: {predicted_class}')
        print(f'Accuracy: {accuracy:.4f}')