fig.subplots_adjust(hspace=0.4, wspace=0.4)
for i, label in enumerate(labels):
fn = files[label]
fig.add_subplot(5, 2, i+1)
plt.title(label)
data, sample_rate = librosa.load(fn)
_ = librosa.display.waveplot(data, sr= sample_rate)
plt.savefig('class_examples.png')
# Log graphic of waveforms to Comet
experiment.log_image('class_examples.png')
# Log audio files to Comet for debugging
for label in labels:
fn = files[label]
experiment.log_audio(fn, metadata = {'name': label})
audiodata = []
for index, row in df.iterrows():
fn = 'UrbanSound8K/audio/fold{}/{}'.format(row['fold'], row['slice_file_name'])
data = read_file_properties(fn)
audiodata.append(data)
# Convert to dataframe
audiodf = pd.DataFrame(audiodata, columns=['num_channels', 'sample_rate', 'bit_depth'])
# Extract MFCC features
fn = 'UrbanSound8K/audio/fold1/191431-9-0-66.wav'
librosa_audio, librosa_sample_rate = librosa.load(fn)
