MusicGenreClassification/musicgenreclassification.py at main · ronitkapoor0106/MusicGenreClassification · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
# -*- coding: utf-8 -*-
"""MusicGenreClassification.ipynb

Automatically generated by Colaboratory.

Original file is located at
    https://colab.research.google.com/drive/12MJ9RUqALnwJW7HBtqzugjTOPGwGTYD0
"""

from google.colab import drive
drive.mount('/content/gdrive/')

from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.decomposition import PCA
from sklearn.neighbors import KNeighborsClassifier
from sklearn.model_selection import GridSearchCV
from sklearn.metrics import confusion_matrix

import librosa

from joblib import dump, load
import numpy as np
import os
import time
import pandas as pd
import sys

# global fileName = []
# # contains all the paths for each au
# def fileIntoList(genre):
#   for i in os.walk("/content/gdrive/My Drive/genres/" + genre):
#     root, dirs, files = i
#     # creating path using os
#     for name in files:
#       fullpath = os.path.join(root, name)
#       fileName.append(fullpath)

# Splitting training/testing
Data = []
Label = []
# 80 v 20
minsize = sys.maxsize


# print(i.size)
def trainTest(genre, label):
  global minsize
  fileNames = []
  for i in os.walk("/content/gdrive/My Drive/genres/" + genre):
    root, dirs, files = i
    for name in files:
      fullpath = os.path.join(root, name)
      if genre in fullpath:
        fileNames.append(fullpath)
  count = 0
  print("file name loaded", len(fileNames))

    # print("file name counted", fileNames[count])
  while (count < len(fileNames)):
    y, sr = librosa.load(fileNames[count])
    y = y.flatten('F')[:y.shape[0]]
    mfccs = librosa.feature.mfcc(y, sr=sr)
    mean_data = [np.mean(feature) for feature in mfccs]
    Data.append(mean_data)
    if (len(mean_data)) < minsize:
      minsize = len(mean_data)
    count += 1
    Label.append(label)
    # print("file name counted", fileNames[count])

  for i in range(len(Data)):
    y = Data[i]
    if len(y) > minsize:

      y = y[:minsize]
      Data[i] = y

  print(minsize)

#genreList = ["blues", "classical", "country", "disco", "hiphop", "jazz", "metal", "pop", "reggae", "rock"]
genreList = ["blues", "classical", "country", "disco", "hiphop", "jazz", "metal", "pop", "reggae", "rock"]
yVals = np.identity(len(genreList))
for i in range(len(genreList)):
  genre = genreList[i]
  label = yVals[i]
  trainTest(genre, label)


x_train, x_test, y_train, y_test = train_test_split(Data, Label)

scaler = StandardScaler()
scaler.fit(x_train)
x_train_scaled = scaler.transform(x_train)
x_test_scaled = scaler.transform(x_test)


pca = PCA().fit(x_train_scaled)

grid_params = {
    'n_neighbors': [1, 3, 7, 99],
    'weights': ['uniform', 'distance'],
    'metric': ['euclidean', 'manhattan']
}

model = GridSearchCV(KNeighborsClassifier(), grid_params, cv=5, n_jobs=-1)

model.fit(x_train_scaled, y_train)

print(f'Model Score: {model.score(x_test_scaled, y_test)}')

y, sr = librosa.load("/content/gdrive/My Drive/genres/alla-turca.au")
y = y.flatten('F')[:y.shape[0]]
mfccs = librosa.feature.mfcc(y, sr=sr)
mean_data = [np.mean(feature) for feature in mfccs]
model.predict([mean_data])

import pickle
with open("/content/gdrive/My Drive/genres/testModel", 'wb') as f:
  pickle.dump(model, f)