{"cells":[{"metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true},"cell_type":"code","source":"import numpy as np\nimport pandas as pd \nfrom scipy.io import wavfile\nimport matplotlib.pyplot as plt\nimport seaborn as sns \n%matplotlib inline\nnp.random.seed(101)\nimport IPython.display as ipd\nimport os\nfrom scipy import signal\nprint(os.listdir(\"../input\"))\nfrom tqdm import tqdm, tqdm_notebook; tqdm.pandas() # Progress bar\nfrom sklearn.metrics import label_ranking_average_precision_score\nfrom sklearn.model_selection import train_test_split\n\n# Machine Learning\nimport tensorflow as tf\nfrom keras import backend as K\nfrom keras.engine.topology import Layer\nfrom keras import initializers, regularizers, constraints, optimizers, layers\nfrom keras.layers import (Dense, Bidirectional, CuDNNLSTM,\n                          Dropout, LeakyReLU, Convolution2D, \n                          Conv2D, Conv1D)\nfrom keras.models import Sequential\nfrom keras.optimizers import Adam\nfrom keras.callbacks import EarlyStopping\n\nimport warnings; warnings.filterwarnings(\"ignore\")\n#print(os.listdir(\"../input/train_curated\"))","execution_count":1,"outputs":[{"output_type":"stream","text":"['test', 'train_noisy.csv', 'train_curated.csv', 'train_curated', 'sample_submission.csv', 'train_noisy']\n","name":"stdout"},{"output_type":"stream","text":"Using TensorFlow backend.\n","name":"stderr"}]},{"metadata":{"_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","trusted":true},"cell_type":"code","source":"train_cur = pd.read_csv(\"../input/train_curated.csv\")\ntrain_nos = pd.read_csv(\"../input/train_noisy.csv\")\ntest = pd.read_csv(\"../input/sample_submission.csv\")","execution_count":2,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"train_cur.head()","execution_count":3,"outputs":[{"output_type":"execute_result","execution_count":3,"data":{"text/plain":"          fname           labels\n0  0006ae4e.wav             Bark\n1  0019ef41.wav         Raindrop\n2  001ec0ad.wav  Finger_snapping\n3  0026c7cb.wav              Run\n4  0026f116.wav  Finger_snapping","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>fname</th>\n      <th>labels</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>0006ae4e.wav</td>\n      <td>Bark</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>0019ef41.wav</td>\n      <td>Raindrop</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>001ec0ad.wav</td>\n      <td>Finger_snapping</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>0026c7cb.wav</td>\n      <td>Run</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>0026f116.wav</td>\n      <td>Finger_snapping</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"train_nos.head()","execution_count":4,"outputs":[{"output_type":"execute_result","execution_count":4,"data":{"text/plain":"          fname                                          labels\n0  00097e21.wav                    Bathtub_(filling_or_washing)\n1  000b6cfb.wav                                      Motorcycle\n2  00116cd2.wav              Marimba_and_xylophone,Glockenspiel\n3  00127d14.wav  Water_tap_and_faucet,Sink_(filling_or_washing)\n4  0019adae.wav                                        Raindrop","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>fname</th>\n      <th>labels</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>00097e21.wav</td>\n      <td>Bathtub_(filling_or_washing)</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>000b6cfb.wav</td>\n      <td>Motorcycle</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>00116cd2.wav</td>\n      <td>Marimba_and_xylophone,Glockenspiel</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>00127d14.wav</td>\n      <td>Water_tap_and_faucet,Sink_(filling_or_washing)</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>0019adae.wav</td>\n      <td>Raindrop</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"test.head()","execution_count":5,"outputs":[{"output_type":"execute_result","execution_count":5,"data":{"text/plain":"          fname        ...          Zipper_(clothing)\n0  000ccb97.wav        ...                          0\n1  0012633b.wav        ...                          0\n2  001ed5f1.wav        ...                          0\n3  00294be0.wav        ...                          0\n4  003fde7a.wav        ...                          0\n\n[5 rows x 81 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>fname</th>\n      <th>Accelerating_and_revving_and_vroom</th>\n      <th>Accordion</th>\n      <th>Acoustic_guitar</th>\n      <th>Applause</th>\n      <th>Bark</th>\n      <th>Bass_drum</th>\n      <th>Bass_guitar</th>\n      <th>Bathtub_(filling_or_washing)</th>\n      <th>Bicycle_bell</th>\n      <th>Burping_and_eructation</th>\n      <th>Bus</th>\n      <th>Buzz</th>\n      <th>Car_passing_by</th>\n      <th>Cheering</th>\n      <th>Chewing_and_mastication</th>\n      <th>Child_speech_and_kid_speaking</th>\n      <th>Chink_and_clink</th>\n      <th>Chirp_and_tweet</th>\n      <th>Church_bell</th>\n      <th>Clapping</th>\n      <th>Computer_keyboard</th>\n      <th>Crackle</th>\n      <th>Cricket</th>\n      <th>Crowd</th>\n      <th>Cupboard_open_or_close</th>\n      <th>Cutlery_and_silverware</th>\n      <th>Dishes_and_pots_and_pans</th>\n      <th>Drawer_open_or_close</th>\n      <th>Drip</th>\n      <th>Electric_guitar</th>\n      <th>Fart</th>\n      <th>Female_singing</th>\n      <th>Female_speech_and_woman_speaking</th>\n      <th>Fill_(with_liquid)</th>\n      <th>Finger_snapping</th>\n      <th>Frying_(food)</th>\n      <th>Gasp</th>\n      <th>Glockenspiel</th>\n      <th>Gong</th>\n      <th>...</th>\n      <th>Harmonica</th>\n      <th>Hi-hat</th>\n      <th>Hiss</th>\n      <th>Keys_jangling</th>\n      <th>Knock</th>\n      <th>Male_singing</th>\n      <th>Male_speech_and_man_speaking</th>\n      <th>Marimba_and_xylophone</th>\n      <th>Mechanical_fan</th>\n      <th>Meow</th>\n      <th>Microwave_oven</th>\n      <th>Motorcycle</th>\n      <th>Printer</th>\n      <th>Purr</th>\n      <th>Race_car_and_auto_racing</th>\n      <th>Raindrop</th>\n      <th>Run</th>\n      <th>Scissors</th>\n      <th>Screaming</th>\n      <th>Shatter</th>\n      <th>Sigh</th>\n      <th>Sink_(filling_or_washing)</th>\n      <th>Skateboard</th>\n      <th>Slam</th>\n      <th>Sneeze</th>\n      <th>Squeak</th>\n      <th>Stream</th>\n      <th>Strum</th>\n      <th>Tap</th>\n      <th>Tick-tock</th>\n      <th>Toilet_flush</th>\n      <th>Traffic_noise_and_roadway_noise</th>\n      <th>Trickle_and_dribble</th>\n      <th>Walk_and_footsteps</th>\n      <th>Water_tap_and_faucet</th>\n      <th>Waves_and_surf</th>\n      <th>Whispering</th>\n      <th>Writing</th>\n      <th>Yell</th>\n      <th>Zipper_(clothing)</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>000ccb97.wav</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>...</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>0012633b.wav</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>...</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>001ed5f1.wav</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>...</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>00294be0.wav</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>...</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>003fde7a.wav</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>...</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"train_audio_cur = \"../input/train_curated/\"\nipd.Audio(train_audio_cur+\"31a0f9cc.wav\")\n","execution_count":6,"outputs":[{"output_type":"execute_result","execution_count":6,"data":{"text/plain":"<IPython.lib.display.Audio object>","text/html":"\n                <audio controls=\"controls\" >\n                    <source src=\"data:audio/x-wav;base64,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\" type=\"audio/x-wav\" />\n                    Your browser does not support the audio element.\n                </audio>\n              "},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"train_cur.loc[train_cur['fname']=='31a0f9cc.wav']","execution_count":7,"outputs":[{"output_type":"execute_result","execution_count":7,"data":{"text/plain":"            fname    labels\n945  31a0f9cc.wav  Clapping","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>fname</th>\n      <th>labels</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>945</th>\n      <td>31a0f9cc.wav</td>\n      <td>Clapping</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"sampling_rate, data = wavfile.read(train_audio_cur+\"31a0f9cc.wav\")\nprint(\"Sampling Rate: {}\".format(sampling_rate))\nprint(\"Data of the audio wave: {}\".format(data))\nprint(\"Duration of Audio file: {}\".format(len(data)/sampling_rate))","execution_count":8,"outputs":[{"output_type":"stream","text":"Sampling Rate: 44100\nData of the audio wave: [-3 -5 -4 ... -4 -8 -8]\nDuration of Audio file: 0.3134920634920635\n","name":"stdout"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"def plot_raw_wave(data):\n    plt.figure(figsize=(16,9))\n    plt.title(\"Raw Representation of Audio Wave\")\n    plt.ylabel(\"Amplitude\")\n    plt.plot(data)\nplot_raw_wave(data)","execution_count":9,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 1152x648 with 1 Axes>","image/png":"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\n"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"#https://www.kaggle.com/davids1992/audio-representation-what-it-s-all-about\ndef log_specgram(audio, sample_rate, window_size=100,\n                 step_size=10, eps=1e-10):\n    nperseg = int(round(window_size * sample_rate / 1e3))\n    noverlap = int(round(step_size * sample_rate / 1e3))\n    freqs, times, spec = signal.spectrogram(audio,\n                                    fs=sample_rate,\n                                    window='hann',\n                                    nperseg=nperseg,\n                                    noverlap=noverlap,\n                                    detrend=False)\n    return freqs, times, np.log(spec.T.astype(np.float32) + eps)\n\ndef plot_log_specgram(audio, sample_rate, window_size=20, step_size=10, eps=1e-10):\n    \n    fig = plt.figure(figsize=(16,9))\n    freqs, times, spectrogram = log_specgram(audio, sample_rate)\n    plt.imshow(spectrogram.T, aspect='auto', origin='lower', \n               extent=[times.min(), times.max(), freqs.min(), freqs.max()])\n    plt.yticks(freqs[::16])\n    plt.xticks(times[::16])\n    plt.title('Spectrogram')\n    plt.ylabel('Freqs in Hz')\n    plt.xlabel('Seconds')\n    plt.show()\nplot_log_specgram(data, sampling_rate, window_size=1000,step_size=10)","execution_count":10,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 1152x648 with 1 Axes>","image/png":"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\n"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"import librosa\nimport librosa.display","execution_count":11,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"#https://www.kaggle.com/davids1992/audio-representation-what-it-s-all-about\nS = librosa.feature.melspectrogram(data.astype(float), sr=sampling_rate, n_mels=128)\n\n# Convert to log scale (dB). We'll use the peak power (max) as reference.\nlog_S = librosa.power_to_db(S, ref=np.max)\n\nplt.figure(figsize=(16, 9))\nlibrosa.display.specshow(log_S, sr=sampling_rate, x_axis='time', y_axis='mel')\nplt.title('Mel power spectrogram ')\nplt.colorbar(format='%+02.0f dB')\nplt.tight_layout()","execution_count":12,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 1152x648 with 2 Axes>","image/png":"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\n"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"mfcc = librosa.feature.mfcc(S=log_S, n_mfcc=13)\n\n# Let's pad on the first and second deltas while we're at it\ndelta2_mfcc = librosa.feature.delta(mfcc, order=2)\n\nplt.figure(figsize=(12, 4))\nlibrosa.display.specshow(delta2_mfcc)\nplt.ylabel('MFCC coeffs')\nplt.xlabel('Time')\nplt.title('MFCC')\nplt.colorbar()\nplt.tight_layout()","execution_count":13,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 864x288 with 2 Axes>","image/png":"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\n"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"def calculate_overall_lwlrap_sklearn(truth, scores):\n    \"\"\"Calculate the overall lwlrap using sklearn.metrics.lrap.\"\"\"\n    # sklearn doesn't correctly apply weighting to samples with no labels, so just skip them.\n    sample_weight = np.sum(truth > 0, axis=1)\n    nonzero_weight_sample_indices = np.flatnonzero(sample_weight > 0)\n    overall_lwlrap = label_ranking_average_precision_score(\n        truth[nonzero_weight_sample_indices, :] > 0, \n        scores[nonzero_weight_sample_indices, :], \n        sample_weight=sample_weight[nonzero_weight_sample_indices])\n    return overall_lwlrap\n","execution_count":14,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"def split_and_label(rows_label):\n    labeled_rows = []\n    for row in rows_label:\n        rows_label = row.split(\",\")\n        arr = np.zeros((80))\n        for label in rows_label:\n            idx = label_mapping[label]\n            arr[idx] = 1\n        labeled_rows.append(arr)\n    return labeled_rows ","execution_count":15,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"label_columns = test.columns[1:]\nlabel_mapping = dict((label,index) for index, label in enumerate(label_columns))\nprint(\"Total Number of Classes: {}\".format(len(label_columns)))","execution_count":16,"outputs":[{"output_type":"stream","text":"Total Number of Classes: 80\n","name":"stdout"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"train_cur_labels = split_and_label(train_cur['labels'])\ntrain_nos_labels = split_and_label(train_nos['labels'])","execution_count":17,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"for col in label_columns:\n    train_cur[col] = 0\n    train_nos[col] = 0\ntrain_cur[label_columns] = train_cur_labels\ntrain_nos[label_columns] = train_nos_labels","execution_count":18,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"train_cur.head()","execution_count":19,"outputs":[{"output_type":"execute_result","execution_count":19,"data":{"text/plain":"          fname           labels        ...          Yell  Zipper_(clothing)\n0  0006ae4e.wav             Bark        ...           0.0                0.0\n1  0019ef41.wav         Raindrop        ...           0.0                0.0\n2  001ec0ad.wav  Finger_snapping        ...           0.0                0.0\n3  0026c7cb.wav              Run        ...           0.0                0.0\n4  0026f116.wav  Finger_snapping        ...           0.0                0.0\n\n[5 rows x 82 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>fname</th>\n      <th>labels</th>\n      <th>Accelerating_and_revving_and_vroom</th>\n      <th>Accordion</th>\n      <th>Acoustic_guitar</th>\n      <th>Applause</th>\n      <th>Bark</th>\n      <th>Bass_drum</th>\n      <th>Bass_guitar</th>\n      <th>Bathtub_(filling_or_washing)</th>\n      <th>Bicycle_bell</th>\n      <th>Burping_and_eructation</th>\n      <th>Bus</th>\n      <th>Buzz</th>\n      <th>Car_passing_by</th>\n      <th>Cheering</th>\n      <th>Chewing_and_mastication</th>\n      <th>Child_speech_and_kid_speaking</th>\n      <th>Chink_and_clink</th>\n      <th>Chirp_and_tweet</th>\n      <th>Church_bell</th>\n      <th>Clapping</th>\n      <th>Computer_keyboard</th>\n      <th>Crackle</th>\n      <th>Cricket</th>\n      <th>Crowd</th>\n      <th>Cupboard_open_or_close</th>\n      <th>Cutlery_and_silverware</th>\n      <th>Dishes_and_pots_and_pans</th>\n      <th>Drawer_open_or_close</th>\n      <th>Drip</th>\n      <th>Electric_guitar</th>\n      <th>Fart</th>\n      <th>Female_singing</th>\n      <th>Female_speech_and_woman_speaking</th>\n      <th>Fill_(with_liquid)</th>\n      <th>Finger_snapping</th>\n      <th>Frying_(food)</th>\n      <th>Gasp</th>\n      <th>Glockenspiel</th>\n      <th>...</th>\n      <th>Harmonica</th>\n      <th>Hi-hat</th>\n      <th>Hiss</th>\n      <th>Keys_jangling</th>\n      <th>Knock</th>\n      <th>Male_singing</th>\n      <th>Male_speech_and_man_speaking</th>\n      <th>Marimba_and_xylophone</th>\n      <th>Mechanical_fan</th>\n      <th>Meow</th>\n      <th>Microwave_oven</th>\n      <th>Motorcycle</th>\n      <th>Printer</th>\n      <th>Purr</th>\n      <th>Race_car_and_auto_racing</th>\n      <th>Raindrop</th>\n      <th>Run</th>\n      <th>Scissors</th>\n      <th>Screaming</th>\n      <th>Shatter</th>\n      <th>Sigh</th>\n      <th>Sink_(filling_or_washing)</th>\n      <th>Skateboard</th>\n      <th>Slam</th>\n      <th>Sneeze</th>\n      <th>Squeak</th>\n      <th>Stream</th>\n      <th>Strum</th>\n      <th>Tap</th>\n      <th>Tick-tock</th>\n      <th>Toilet_flush</th>\n      <th>Traffic_noise_and_roadway_noise</th>\n      <th>Trickle_and_dribble</th>\n      <th>Walk_and_footsteps</th>\n      <th>Water_tap_and_faucet</th>\n      <th>Waves_and_surf</th>\n      <th>Whispering</th>\n      <th>Writing</th>\n      <th>Yell</th>\n      <th>Zipper_(clothing)</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>0006ae4e.wav</td>\n      <td>Bark</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>1.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>...</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>0019ef41.wav</td>\n      <td>Raindrop</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>...</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>1.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>001ec0ad.wav</td>\n      <td>Finger_snapping</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>1.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>...</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>0026c7cb.wav</td>\n      <td>Run</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>...</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>1.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>0026f116.wav</td>\n      <td>Finger_snapping</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>1.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>...</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"train_nos.head()","execution_count":20,"outputs":[{"output_type":"execute_result","execution_count":20,"data":{"text/plain":"          fname        ...        Zipper_(clothing)\n0  00097e21.wav        ...                      0.0\n1  000b6cfb.wav        ...                      0.0\n2  00116cd2.wav        ...                      0.0\n3  00127d14.wav        ...                      0.0\n4  0019adae.wav        ...                      0.0\n\n[5 rows x 82 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>fname</th>\n      <th>labels</th>\n      <th>Accelerating_and_revving_and_vroom</th>\n      <th>Accordion</th>\n      <th>Acoustic_guitar</th>\n      <th>Applause</th>\n      <th>Bark</th>\n      <th>Bass_drum</th>\n      <th>Bass_guitar</th>\n      <th>Bathtub_(filling_or_washing)</th>\n      <th>Bicycle_bell</th>\n      <th>Burping_and_eructation</th>\n      <th>Bus</th>\n      <th>Buzz</th>\n      <th>Car_passing_by</th>\n      <th>Cheering</th>\n      <th>Chewing_and_mastication</th>\n      <th>Child_speech_and_kid_speaking</th>\n      <th>Chink_and_clink</th>\n      <th>Chirp_and_tweet</th>\n      <th>Church_bell</th>\n      <th>Clapping</th>\n      <th>Computer_keyboard</th>\n      <th>Crackle</th>\n      <th>Cricket</th>\n      <th>Crowd</th>\n      <th>Cupboard_open_or_close</th>\n      <th>Cutlery_and_silverware</th>\n      <th>Dishes_and_pots_and_pans</th>\n      <th>Drawer_open_or_close</th>\n      <th>Drip</th>\n      <th>Electric_guitar</th>\n      <th>Fart</th>\n      <th>Female_singing</th>\n      <th>Female_speech_and_woman_speaking</th>\n      <th>Fill_(with_liquid)</th>\n      <th>Finger_snapping</th>\n      <th>Frying_(food)</th>\n      <th>Gasp</th>\n      <th>Glockenspiel</th>\n      <th>...</th>\n      <th>Harmonica</th>\n      <th>Hi-hat</th>\n      <th>Hiss</th>\n      <th>Keys_jangling</th>\n      <th>Knock</th>\n      <th>Male_singing</th>\n      <th>Male_speech_and_man_speaking</th>\n      <th>Marimba_and_xylophone</th>\n      <th>Mechanical_fan</th>\n      <th>Meow</th>\n      <th>Microwave_oven</th>\n      <th>Motorcycle</th>\n      <th>Printer</th>\n      <th>Purr</th>\n      <th>Race_car_and_auto_racing</th>\n      <th>Raindrop</th>\n      <th>Run</th>\n      <th>Scissors</th>\n      <th>Screaming</th>\n      <th>Shatter</th>\n      <th>Sigh</th>\n      <th>Sink_(filling_or_washing)</th>\n      <th>Skateboard</th>\n      <th>Slam</th>\n      <th>Sneeze</th>\n      <th>Squeak</th>\n      <th>Stream</th>\n      <th>Strum</th>\n      <th>Tap</th>\n      <th>Tick-tock</th>\n      <th>Toilet_flush</th>\n      <th>Traffic_noise_and_roadway_noise</th>\n      <th>Trickle_and_dribble</th>\n      <th>Walk_and_footsteps</th>\n      <th>Water_tap_and_faucet</th>\n      <th>Waves_and_surf</th>\n      <th>Whispering</th>\n      <th>Writing</th>\n      <th>Yell</th>\n      <th>Zipper_(clothing)</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>00097e21.wav</td>\n      <td>Bathtub_(filling_or_washing)</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>1.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>...</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>000b6cfb.wav</td>\n      <td>Motorcycle</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>...</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>1.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>00116cd2.wav</td>\n      <td>Marimba_and_xylophone,Glockenspiel</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>1.0</td>\n      <td>...</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>1.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>00127d14.wav</td>\n      <td>Water_tap_and_faucet,Sink_(filling_or_washing)</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>...</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>1.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>1.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>0019adae.wav</td>\n      <td>Raindrop</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>...</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>1.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"input_shape = (890,128)\nn_classes = 80\nn_epochs = 500\nhop_length = 347\nfmin = 20\nfmax = sampling_rate // 2\nn_mels = 128\nn_fft = n_mels*20\nopt = Adam(0.003, beta_1=0.75, beta_2=0.85, amsgrad=True)\nsampling_rate = 44100\nduration = 7\nsamples = sampling_rate*duration\n","execution_count":21,"outputs":[{"output_type":"stream","text":"WARNING:tensorflow:From /opt/conda/lib/python3.6/site-packages/tensorflow/python/framework/op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\nInstructions for updating:\nColocations handled automatically by placer.\n","name":"stdout"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"def read_audio(pathname, trim_long_data):\n    y, sr = librosa.load(pathname, sr=sampling_rate)\n    # trim silence\n    if 0 < len(y): # workaround: 0 length causes error\n        y, _ = librosa.effects.trim(y) # trim, top_db=default(60)\n    # make it unified length to conf.samples\n    if len(y) > samples: # long enough\n        if trim_long_data:\n            y = y[0:0+samples]\n    else: # pad blank\n        padding = samples - len(y)    # add padding at both ends\n        offset = padding // 2\n        y = np.pad(y, (offset, samples - len(y) - offset), 'constant')\n    return y\n\ndef audio_to_melspectrogram(audio):\n    spectrogram = librosa.feature.melspectrogram(audio, \n                                                 sr=sampling_rate,\n                                                 n_mels=n_mels,\n                                                 hop_length=hop_length,\n                                                 n_fft=n_fft,\n                                                 fmin=fmin,\n                                                 fmax=fmax)\n    spectrogram = librosa.power_to_db(spectrogram)\n    spectrogram = spectrogram.astype(np.float32)\n    return spectrogram\n\ndef read_as_melspectrogram(pathname, trim_long_data, debug_display=False):\n    x = read_audio(pathname, trim_long_data)\n    mels = audio_to_melspectrogram(x)\n    if debug_display:\n        IPython.display.display(IPython.display.Audio(x, rate=sampling_rate))\n        show_melspectrogram(mels)\n    return mels\n\ndef convert_wav_to_image(df, source):\n    X = []\n    for i, row in tqdm_notebook(df.iterrows()):\n        try:\n            x = read_as_melspectrogram(f'{source[0]}/{str(row.fname)}', trim_long_data=True)\n        except:\n            x = read_as_melspectrogram(f'{source[1]}/{str(row.fname)}', trim_long_data=True)\n\n        #x_color = mono_to_color(x)\n        X.append(x.transpose())\n        #df.loc[i, 'length'] = x.shape[1]\n    return X","execution_count":22,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"train = pd.concat([train_cur[:4000],train_nos[:2000]])\ntrain_curated_path = '../input/train_curated/'\ntrain_noisy_path = '../input/train_noisy/'\ntest_path = '../input/test/'\ntrain.shape","execution_count":23,"outputs":[{"output_type":"execute_result","execution_count":23,"data":{"text/plain":"(6000, 82)"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"%%time\n#X = np.array(convert_wav_to_image(train_cur,source=[train_curated_path]))\n#X = np.array(convert_wav_to_image(train_nos, source=[train_noisy_path]))\nX = np.array(convert_wav_to_image(train, source=[train_curated_path, train_noisy_path]))","execution_count":null,"outputs":[{"output_type":"display_data","data":{"text/plain":"HBox(children=(IntProgress(value=1, bar_style='info', max=1), HTML(value='')))","application/vnd.jupyter.widget-view+json":{"version_major":2,"version_minor":0,"model_id":"201cf3fae85f476eaa6bd621c6cc9e54"}},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"Y = train[label_columns].values","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"#\nclass Attention(Layer):\n    def __init__(self, step_dim,\n                 W_regularizer=None, b_regularizer=None,\n                 W_constraint=None, b_constraint=None,\n                 bias=True, **kwargs):\n        self.supports_masking = True\n        self.init = initializers.get('glorot_uniform')\n        self.W_regularizer = regularizers.get(W_regularizer)\n        self.b_regularizer = regularizers.get(b_regularizer)\n        self.W_constraint = constraints.get(W_constraint)\n        self.b_constraint = constraints.get(b_constraint)\n        self.bias = bias\n        self.step_dim = step_dim\n        self.features_dim = 0\n        super(Attention, self).__init__(**kwargs)\n\n    def build(self, input_shape):\n        assert len(input_shape) == 3\n\n        self.W = self.add_weight((input_shape[-1],),\n                                 initializer=self.init,\n                                 name='{}_W'.format(self.name),\n                                 regularizer=self.W_regularizer,\n                                 constraint=self.W_constraint)\n        self.features_dim = input_shape[-1]\n\n        if self.bias:\n            self.b = self.add_weight((input_shape[1],),\n                                     initializer='zero',\n                                     name='{}_b'.format(self.name),\n                                     regularizer=self.b_regularizer,\n                                     constraint=self.b_constraint)\n        else:\n            self.b = None\n        self.built = True\n\n    def compute_mask(self, input, input_mask=None):\n        return None\n\n    def call(self, x, mask=None):\n        features_dim = self.features_dim\n        step_dim = self.step_dim\n\n        eij = K.reshape(K.dot(K.reshape(x, (-1, features_dim)),\n                        K.reshape(self.W, (features_dim, 1))), (-1, step_dim))\n        if self.bias:\n            eij += self.b\n        eij = K.tanh(eij)\n        a = K.exp(eij)\n        if mask is not None:\n            a *= K.cast(mask, K.floatx())\n        a /= K.cast(K.sum(a, axis=1, keepdims=True) + K.epsilon(), K.floatx())\n        a = K.expand_dims(a)\n        weighted_input = x * a\n        return K.sum(weighted_input, axis=1)\n\n    def compute_output_shape(self, input_shape):\n        return input_shape[0],  self.features_dim","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"model = Sequential()\nmodel.add(Bidirectional(CuDNNLSTM(128,return_sequences=True),input_shape=input_shape))\nmodel.add(Bidirectional(CuDNNLSTM(128,return_sequences=True),input_shape=input_shape))\nmodel.add(Bidirectional(CuDNNLSTM(128,return_sequences=True),input_shape=input_shape))\nmodel.add(Attention(input_shape[0]))\nmodel.add(Dropout(0.2))\nmodel.add(Dense(256,activation='relu'))\nmodel.add(Dropout(0.2))\nmodel.add(Dense(128,activation='relu'))\nmodel.add(Dropout(0.2))\nmodel.add(Dense(n_classes,activation='softmax'))\nmodel.compile(loss='categorical_crossentropy',optimizer=opt,metrics=['acc'])\nmodel.summary()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"es = EarlyStopping(monitor='val_acc', mode='max', verbose=2, patience=10)\nx_train, x_val, y_train, y_val = train_test_split(X, Y, test_size=0.1, random_state=101)\nhistory = model.fit(np.array(x_train),\n          y_train,\n          batch_size=512,\n          epochs=500,\n          validation_data=(np.array(x_val), y_val),\n          callbacks = [es]\n                   )","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"y_train_pred = model.predict(np.array(x_train))\ny_val_pred = model.predict(np.array(x_val))\ntrain_lwlrap = calculate_overall_lwlrap_sklearn(y_train, y_train_pred)\nval_lwlrap = calculate_overall_lwlrap_sklearn(y_val, y_val_pred)\n\n# Check training and validation LWLRAP score\nprint('Training LWLRAP : {}'.format(round(train_lwlrap,4)))\nprint('Validation LWLRAP : {}'.format(round(val_lwlrap,4)))","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"X_test = np.array(convert_wav_to_image(test, [test_path]))\npredictions = model.predict(np.array(X_test))\ntest[label_columns] = predictions\ntest.to_csv('submission.csv', index=False)","execution_count":null,"outputs":[]}],"metadata":{"kernelspec":{"display_name":"Python 3","language":"python","name":"python3"},"language_info":{"name":"python","version":"3.6.4","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat":4,"nbformat_minor":1}