End-to-End Speech Recognition in Tensorflow.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# End-to-End Speech Recognition in Tensorflow"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### The dataset that will be used for this tutorial is from [Speech Recognition Challenge in Kaggle](https://www.kaggle.com/c/tensorflow-speech-recognition-challenge/data)\n",
    "##### It has 65,000 one-second long utterances of 30 short words, by thousands of different people. We’ll build a speech recognition system that understands simple spoken commands.\n",
    "##### The 10 labels we will need to predict in Test-set are 'yes', 'no', 'up', 'down', 'left', 'right', 'on', 'off', 'stop', and 'go'. Everything else should be considered either unknown or NA."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Python version: 3.7.6\n",
      "Tensorflow version: 2.2.0\n",
      "Keras version: 2.3.0-tf\n",
      "The tensorboard extension is already loaded. To reload it, use:\n",
      "  %reload_ext tensorboard\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "# LibROSA is a python package for music and audio analysis. It provides the building blocks necessary to create music information retrieval systems.\n",
    "import librosa\n",
    "from scipy.io import wavfile\n",
    "import warnings\n",
    "\n",
    "import os\n",
    "import librosa\n",
    "import IPython.display as ipd\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import platform\n",
    "import datetime\n",
    "import os\n",
    "import math\n",
    "import random\n",
    "import pathlib\n",
    "import json\n",
    "import pickle\n",
    "import re\n",
    "from tqdm import tqdm\n",
    "\n",
    "from keras.layers import Dense, Dropout, Flatten, Conv1D, Conv2D, Input, MaxPooling1D\n",
    "from keras.models import Model\n",
    "from keras.callbacks import EarlyStopping, ModelCheckpoint\n",
    "from keras import backend as K\n",
    "\n",
    "\n",
    "print('Python version:', platform.python_version())\n",
    "print('Tensorflow version:', tf.__version__)\n",
    "print('Keras version:', tf.keras.__version__)\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "%reload_ext autoreload\n",
    "%autoreload 2\n",
    "%matplotlib inline\n",
    "%load_ext tensorboard"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'C:\\\\Users\\\\bokhy\\\\Desktop\\\\Python\\\\Python-Projects\\\\Tensorflow\\\\data'"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# checking current directory\n",
    "os.path.abspath('./data')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1. Load Data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Librosa tutorial and codes \n",
    "#### https://librosa.github.io/librosa/tutorial.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x2af49eeba88>]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_audio_path = 'E:/Python/data/tensorflow-speech-recognition-challenge/train/audio/'\n",
    "samples, sample_rate = librosa.load(train_audio_path+'yes/0a7c2a8d_nohash_0.wav', sr = 16000)\n",
    "fig = plt.figure(figsize=(20, 10))\n",
    "ax = fig.add_subplot(211)\n",
    "ax.set_title('Raw wave of ' + '../input/train/audio/yes/0a7c2a8d_nohash_0.wav')\n",
    "ax.plot(np.linspace(0, sample_rate/len(samples), sample_rate), samples)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2. Audio EDA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "                <audio  controls=\"controls\" >\n",
       "                    <source src=\"data:audio/wav;base64,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\" type=\"audio/wav\" />\n",
       "                    Your browser does not support the audio element.\n",
       "                </audio>\n",
       "              "
      ],
      "text/plain": [
       "<IPython.lib.display.Audio object>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check the Sampling Rate\n",
    "ipd.Audio(samples, rate=sample_rate)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "16000\n"
     ]
    }
   ],
   "source": [
    "# check sampling rate\n",
    "print(sample_rate)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 1.5111059e-05 -9.0932073e-05  1.6909138e-04 ...  1.2670006e-03\n",
      "  5.7899451e-04  2.8395216e-04]\n"
     ]
    }
   ],
   "source": [
    "# Resample\n",
    "# We see that the sampling rate of the signal is 16000 hz. Let's resample it to 8000 hz (since most of the speech related frequencies are present in 8000z)\n",
    "samples = librosa.resample(samples, sample_rate, 8000)\n",
    "ipd.Audio(samples, rate=8000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['bed',\n",
       " 'bird',\n",
       " 'cat',\n",
       " 'dog',\n",
       " 'down',\n",
       " 'eight',\n",
       " 'five',\n",
       " 'four',\n",
       " 'go',\n",
       " 'happy',\n",
       " 'house',\n",
       " 'left',\n",
       " 'marvin',\n",
       " 'nine',\n",
       " 'no',\n",
       " 'off',\n",
       " 'on',\n",
       " 'one',\n",
       " 'right',\n",
       " 'seven',\n",
       " 'sheila',\n",
       " 'six',\n",
       " 'stop',\n",
       " 'three',\n",
       " 'tree',\n",
       " 'two',\n",
       " 'up',\n",
       " 'wow',\n",
       " 'yes',\n",
       " 'zero',\n",
       " '_background_noise_']"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "os.listdir(train_audio_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# number of recordings for each voice command\n",
    "labels=os.listdir(train_audio_path)\n",
    "\n",
    "recording=[]\n",
    "for label in labels:\n",
    "    waves = [f for f in os.listdir(train_audio_path + '/'+ label) if f.endswith('.wav')]\n",
    "    recording.append(len(waves))\n",
    "    \n",
    "#plot\n",
    "plt.figure(figsize=(30,5))\n",
    "index = np.arange(len(labels))\n",
    "plt.bar(index, recording)\n",
    "plt.xlabel('Commands', fontsize=12)\n",
    "plt.ylabel('No of recordings', fontsize=12)\n",
    "plt.xticks(index, labels, fontsize=15, rotation=60)\n",
    "plt.title('No. of recordings for each command')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "# List out lables\n",
    "labels=[\"yes\", \"no\", \"up\", \"down\", \"left\", \"right\", \"on\", \"off\", \"stop\", \"go\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([1.5000e+01, 3.0000e+01, 4.4000e+01, 1.3800e+02, 1.3600e+02,\n",
       "        1.7900e+02, 3.6600e+02, 4.3400e+02, 5.9300e+02, 2.1747e+04]),\n",
       " array([0.418 , 0.4762, 0.5344, 0.5926, 0.6508, 0.709 , 0.7672, 0.8254,\n",
       "        0.8836, 0.9418, 1.    ]),\n",
       " <a list of 10 Patch objects>)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "duration_of_recordings=[]\n",
    "for label in labels:\n",
    "    waves = [f for f in os.listdir(train_audio_path + '/'+ label) if f.endswith('.wav')]\n",
    "    for wav in waves:\n",
    "        sample_rate, samples = wavfile.read(train_audio_path + '/' + label + '/' + wav)\n",
    "        duration_of_recordings.append(float(len(samples)/sample_rate))\n",
    "    \n",
    "plt.hist(np.array(duration_of_recordings))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 3. Data Pre-processing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Duration of a few recordings is less than 1 second and the sampling rate is too high."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### (1) Resampling\n",
    "##### (2) Remove outliers (audio shorter than 1 second)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "yes\n",
      "no\n",
      "up\n",
      "down\n",
      "left\n",
      "right\n",
      "on\n",
      "off\n",
      "stop\n",
      "go\n"
     ]
    }
   ],
   "source": [
    "all_wave = []\n",
    "all_label = []\n",
    "for label in labels:\n",
    "    print(label)\n",
    "    waves = [f for f in os.listdir(train_audio_path + '/'+ label) if f.endswith('.wav')]\n",
    "    for wav in waves:\n",
    "        samples, sample_rate = librosa.load(train_audio_path + '/' + label + '/' + wav, sr = 16000)\n",
    "        samples = librosa.resample(samples, sample_rate, 8000)\n",
    "        if(len(samples)== 8000) : \n",
    "            all_wave.append(samples)\n",
    "            all_label.append(label)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### (3) Convert the output labels to integer-encoded:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['down', 'go', 'left', 'no', 'off', 'on', 'right', 'stop', 'up', 'yes']\n"
     ]
    }
   ],
   "source": [
    "from sklearn.preprocessing import LabelEncoder\n",
    "le = LabelEncoder()\n",
    "y=le.fit_transform(all_label)\n",
    "classes= list(le.classes_)\n",
    "print(classes)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### (4) Convert integer-encoded to a one-hot vector since it is a multi-classification problem"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.utils import np_utils\n",
    "y=np_utils.to_categorical(y, num_classes=len(labels))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### (5) Reshape the 2D array to 3D since the input to the conv1d must be a 3D array:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[[-2.88872398e-03],\n",
       "        [-4.09192173e-03],\n",
       "        [-3.25268088e-03],\n",
       "        ...,\n",
       "        [-7.19285978e-04],\n",
       "        [-1.12521090e-03],\n",
       "        [-1.30131142e-03]],\n",
       "\n",
       "       [[ 4.03327926e-04],\n",
       "        [ 6.26417692e-04],\n",
       "        [ 3.67147470e-04],\n",
       "        ...,\n",
       "        [ 8.11780745e-04],\n",
       "        [ 4.86697943e-04],\n",
       "        [ 7.11379049e-04]],\n",
       "\n",
       "       [[-1.58481183e-04],\n",
       "        [-1.79531227e-04],\n",
       "        [-1.15519455e-04],\n",
       "        ...,\n",
       "        [ 4.31013264e-04],\n",
       "        [ 3.31727992e-04],\n",
       "        [ 3.07029666e-04]],\n",
       "\n",
       "       ...,\n",
       "\n",
       "       [[-1.52462671e-04],\n",
       "        [-2.13622799e-04],\n",
       "        [-1.75312554e-04],\n",
       "        ...,\n",
       "        [ 4.78350848e-04],\n",
       "        [ 3.12889199e-04],\n",
       "        [ 2.94630474e-04]],\n",
       "\n",
       "       [[-5.92960969e-05],\n",
       "        [-1.66434096e-04],\n",
       "        [-1.80037678e-04],\n",
       "        ...,\n",
       "        [ 1.01546226e-04],\n",
       "        [-4.89999038e-05],\n",
       "        [-1.18220858e-04]],\n",
       "\n",
       "       [[ 1.62130367e-04],\n",
       "        [ 4.54971945e-04],\n",
       "        [ 3.75219883e-04],\n",
       "        ...,\n",
       "        [ 3.78574114e-05],\n",
       "        [-1.10495501e-04],\n",
       "        [ 3.32449388e-04]]], dtype=float32)"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "all_wave = np.array(all_wave).reshape(-1,8000,1) ; all_wave"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 4. Train/Test Split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "x_tr, x_val, y_tr, y_val = train_test_split(np.array(all_wave),np.array(y),\n",
    "                                            stratify = y,\n",
    "                                            test_size = 0.2,\n",
    "                                            random_state = 623,\n",
    "                                            shuffle = True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 5. Create a Model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# We build the model in convulutional-1d. Conv1d is a convolutional neural network which performs the convolution along only one dimension."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         (None, 8000, 1)           0         \n",
      "_________________________________________________________________\n",
      "conv1d_1 (Conv1D)            (None, 7988, 8)           112       \n",
      "_________________________________________________________________\n",
      "max_pooling1d_1 (MaxPooling1 (None, 2662, 8)           0         \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 2662, 8)           0         \n",
      "_________________________________________________________________\n",
      "conv1d_2 (Conv1D)            (None, 2652, 16)          1424      \n",
      "_________________________________________________________________\n",
      "max_pooling1d_2 (MaxPooling1 (None, 884, 16)           0         \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 884, 16)           0         \n",
      "_________________________________________________________________\n",
      "conv1d_3 (Conv1D)            (None, 876, 32)           4640      \n",
      "_________________________________________________________________\n",
      "max_pooling1d_3 (MaxPooling1 (None, 292, 32)           0         \n",
      "_________________________________________________________________\n",
      "dropout_3 (Dropout)          (None, 292, 32)           0         \n",
      "_________________________________________________________________\n",
      "conv1d_4 (Conv1D)            (None, 286, 64)           14400     \n",
      "_________________________________________________________________\n",
      "max_pooling1d_4 (MaxPooling1 (None, 95, 64)            0         \n",
      "_________________________________________________________________\n",
      "dropout_4 (Dropout)          (None, 95, 64)            0         \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 6080)              0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 256)               1556736   \n",
      "_________________________________________________________________\n",
      "dropout_5 (Dropout)          (None, 256)               0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 128)               32896     \n",
      "_________________________________________________________________\n",
      "dropout_6 (Dropout)          (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 10)                1290      \n",
      "=================================================================\n",
      "Total params: 1,611,498\n",
      "Trainable params: 1,611,498\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "inputs = Input(shape=(8000,1))\n",
    "\n",
    "#First Conv1D layer\n",
    "conv = Conv1D(8,13, padding='valid', activation='relu', strides=1)(inputs)\n",
    "conv = MaxPooling1D(3)(conv)\n",
    "conv = Dropout(0.3)(conv)\n",
    "\n",
    "#Second Conv1D layer\n",
    "conv = Conv1D(16, 11, padding='valid', activation='relu', strides=1)(conv)\n",
    "conv = MaxPooling1D(3)(conv)\n",
    "conv = Dropout(0.3)(conv)\n",
    "\n",
    "#Third Conv1D layer\n",
    "conv = Conv1D(32, 9, padding='valid', activation='relu', strides=1)(conv)\n",
    "conv = MaxPooling1D(3)(conv)\n",
    "conv = Dropout(0.3)(conv)\n",
    "\n",
    "#Fourth Conv1D layer\n",
    "conv = Conv1D(64, 7, padding='valid', activation='relu', strides=1)(conv)\n",
    "conv = MaxPooling1D(3)(conv)\n",
    "conv = Dropout(0.3)(conv)\n",
    "\n",
    "#Flatten layer\n",
    "conv = Flatten()(conv)\n",
    "\n",
    "#Dense Layer 1\n",
    "conv = Dense(256, activation='relu')(conv)\n",
    "conv = Dropout(0.3)(conv)\n",
    "\n",
    "#Dense Layer 2\n",
    "conv = Dense(128, activation='relu')(conv)\n",
    "conv = Dropout(0.3)(conv)\n",
    "\n",
    "outputs = Dense(len(labels), activation='softmax')(conv)\n",
    "\n",
    "model = Model(inputs, outputs)\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 6. Train a model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Loss function\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer='adam',\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "early_stopping = EarlyStopping(monitor='val_loss', \n",
    "                   mode='min', \n",
    "                   verbose=1, \n",
    "                   patience=10, \n",
    "                   min_delta=0.0001) \n",
    "\n",
    "checkpoint = ModelCheckpoint('best_model.hdf5', \n",
    "                     monitor='val_acc', \n",
    "                     verbose=True, \n",
    "                     save_best_only=True, \n",
    "                     mode='max')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 17049 samples, validate on 4263 samples\n",
      "Epoch 1/100\n",
      "17049/17049 [==============================] - 75s 4ms/step - loss: 2.0043 - accuracy: 0.2424 - val_loss: 1.6191 - val_accuracy: 0.4356\n",
      "Epoch 2/100\n",
      "17049/17049 [==============================] - 67s 4ms/step - loss: 1.3507 - accuracy: 0.5117 - val_loss: 1.0047 - val_accuracy: 0.6542\n",
      "Epoch 3/100\n",
      "17049/17049 [==============================] - 71s 4ms/step - loss: 1.0682 - accuracy: 0.6206 - val_loss: 0.9290 - val_accuracy: 0.6866\n",
      "Epoch 4/100\n",
      "17049/17049 [==============================] - 73s 4ms/step - loss: 0.9245 - accuracy: 0.6765 - val_loss: 0.8162 - val_accuracy: 0.7225\n",
      "Epoch 5/100\n",
      "17049/17049 [==============================] - 70s 4ms/step - loss: 0.8274 - accuracy: 0.7097 - val_loss: 0.6851 - val_accuracy: 0.7652\n",
      "Epoch 6/100\n",
      "17049/17049 [==============================] - 69s 4ms/step - loss: 0.7629 - accuracy: 0.7352 - val_loss: 0.7226 - val_accuracy: 0.7478\n",
      "Epoch 7/100\n",
      "17049/17049 [==============================] - 72s 4ms/step - loss: 0.7095 - accuracy: 0.7517 - val_loss: 0.6777 - val_accuracy: 0.7694\n",
      "Epoch 8/100\n",
      "17049/17049 [==============================] - 75s 4ms/step - loss: 0.6562 - accuracy: 0.7687 - val_loss: 0.6253 - val_accuracy: 0.7882\n",
      "Epoch 9/100\n",
      "17049/17049 [==============================] - 77s 5ms/step - loss: 0.6380 - accuracy: 0.7821 - val_loss: 0.6038 - val_accuracy: 0.7917\n",
      "Epoch 10/100\n",
      "17049/17049 [==============================] - 75s 4ms/step - loss: 0.5777 - accuracy: 0.7972 - val_loss: 0.5867 - val_accuracy: 0.8023\n",
      "Epoch 11/100\n",
      "17049/17049 [==============================] - 74s 4ms/step - loss: 0.5590 - accuracy: 0.8004 - val_loss: 0.5850 - val_accuracy: 0.8001\n",
      "Epoch 12/100\n",
      "17049/17049 [==============================] - 72s 4ms/step - loss: 0.5457 - accuracy: 0.8112 - val_loss: 0.5264 - val_accuracy: 0.8222\n",
      "Epoch 13/100\n",
      "17049/17049 [==============================] - 75s 4ms/step - loss: 0.5130 - accuracy: 0.8229 - val_loss: 0.5060 - val_accuracy: 0.8297\n",
      "Epoch 14/100\n",
      "17049/17049 [==============================] - 77s 5ms/step - loss: 0.4978 - accuracy: 0.8267 - val_loss: 0.5215 - val_accuracy: 0.8213\n",
      "Epoch 15/100\n",
      "17049/17049 [==============================] - 72s 4ms/step - loss: 0.4817 - accuracy: 0.8316 - val_loss: 0.5157 - val_accuracy: 0.8248\n",
      "Epoch 16/100\n",
      "17049/17049 [==============================] - 81s 5ms/step - loss: 0.4732 - accuracy: 0.8325 - val_loss: 0.5418 - val_accuracy: 0.8166\n",
      "Epoch 17/100\n",
      "17049/17049 [==============================] - 72s 4ms/step - loss: 0.4553 - accuracy: 0.8412 - val_loss: 0.5106 - val_accuracy: 0.8264\n",
      "Epoch 18/100\n",
      "17049/17049 [==============================] - 72s 4ms/step - loss: 0.4360 - accuracy: 0.8474 - val_loss: 0.5167 - val_accuracy: 0.8283\n",
      "Epoch 19/100\n",
      "17049/17049 [==============================] - 77s 5ms/step - loss: 0.4191 - accuracy: 0.8525 - val_loss: 0.4758 - val_accuracy: 0.8426\n",
      "Epoch 20/100\n",
      "17049/17049 [==============================] - 78s 5ms/step - loss: 0.4035 - accuracy: 0.8620 - val_loss: 0.5099 - val_accuracy: 0.8327\n",
      "Epoch 21/100\n",
      "17049/17049 [==============================] - 72s 4ms/step - loss: 0.4214 - accuracy: 0.8559 - val_loss: 0.5123 - val_accuracy: 0.8269\n",
      "Epoch 22/100\n",
      "17049/17049 [==============================] - 76s 4ms/step - loss: 0.3912 - accuracy: 0.8652 - val_loss: 0.5626 - val_accuracy: 0.8231\n",
      "Epoch 23/100\n",
      "17049/17049 [==============================] - 149s 9ms/step - loss: 0.3948 - accuracy: 0.8649 - val_loss: 0.4939 - val_accuracy: 0.8337\n",
      "Epoch 24/100\n",
      "17049/17049 [==============================] - 74s 4ms/step - loss: 0.3874 - accuracy: 0.8669 - val_loss: 0.5121 - val_accuracy: 0.8349\n",
      "Epoch 25/100\n",
      "17049/17049 [==============================] - 76s 4ms/step - loss: 0.3759 - accuracy: 0.8705 - val_loss: 0.4809 - val_accuracy: 0.8412\n",
      "Epoch 26/100\n",
      "17049/17049 [==============================] - 92s 5ms/step - loss: 0.3574 - accuracy: 0.8750 - val_loss: 0.4866 - val_accuracy: 0.8417\n",
      "Epoch 27/100\n",
      "17049/17049 [==============================] - 86s 5ms/step - loss: 0.3315 - accuracy: 0.8812 - val_loss: 0.4649 - val_accuracy: 0.8454\n",
      "Epoch 28/100\n",
      "17049/17049 [==============================] - 82s 5ms/step - loss: 0.3541 - accuracy: 0.8781 - val_loss: 0.5363 - val_accuracy: 0.8194\n",
      "Epoch 29/100\n",
      "17049/17049 [==============================] - 82s 5ms/step - loss: 0.3454 - accuracy: 0.8825 - val_loss: 0.4884 - val_accuracy: 0.8386\n",
      "Epoch 30/100\n",
      "17049/17049 [==============================] - 81s 5ms/step - loss: 0.3360 - accuracy: 0.8805 - val_loss: 0.4569 - val_accuracy: 0.8506\n",
      "Epoch 31/100\n",
      "17049/17049 [==============================] - 94s 6ms/step - loss: 0.3283 - accuracy: 0.8870 - val_loss: 0.5097 - val_accuracy: 0.8353\n",
      "Epoch 32/100\n",
      "17049/17049 [==============================] - 80s 5ms/step - loss: 0.3267 - accuracy: 0.8883 - val_loss: 0.5006 - val_accuracy: 0.8379\n",
      "Epoch 33/100\n",
      "17049/17049 [==============================] - 127s 7ms/step - loss: 0.3127 - accuracy: 0.8945 - val_loss: 0.4658 - val_accuracy: 0.8466\n",
      "Epoch 34/100\n",
      "17049/17049 [==============================] - 76s 4ms/step - loss: 0.3191 - accuracy: 0.8904 - val_loss: 0.4903 - val_accuracy: 0.8424\n",
      "Epoch 35/100\n",
      "17049/17049 [==============================] - 87s 5ms/step - loss: 0.3070 - accuracy: 0.8954 - val_loss: 0.4839 - val_accuracy: 0.8421\n",
      "Epoch 36/100\n",
      "17049/17049 [==============================] - 100s 6ms/step - loss: 0.2987 - accuracy: 0.8972 - val_loss: 0.5035 - val_accuracy: 0.8407\n",
      "Epoch 37/100\n",
      "17049/17049 [==============================] - 89s 5ms/step - loss: 0.2916 - accuracy: 0.8994 - val_loss: 0.5572 - val_accuracy: 0.8266\n",
      "Epoch 38/100\n",
      "17049/17049 [==============================] - 106s 6ms/step - loss: 0.3011 - accuracy: 0.8979 - val_loss: 0.5048 - val_accuracy: 0.8414\n",
      "Epoch 39/100\n",
      "17049/17049 [==============================] - 76s 4ms/step - loss: 0.2905 - accuracy: 0.9018 - val_loss: 0.4989 - val_accuracy: 0.8442\n",
      "Epoch 40/100\n",
      "17049/17049 [==============================] - 82s 5ms/step - loss: 0.2857 - accuracy: 0.9015 - val_loss: 0.5567 - val_accuracy: 0.8227\n",
      "Epoch 00040: early stopping\n"
     ]
    }
   ],
   "source": [
    "model_history=model.fit(x_tr, y_tr,\n",
    "                  epochs=100, \n",
    "                  callbacks=[early_stopping,checkpoint], \n",
    "                  batch_size=32, \n",
    "                  validation_data=(x_val,y_val))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "def render_training_history(model_history):\n",
    "    loss = model_history.history['loss']\n",
    "    val_loss = model_history.history['val_loss']\n",
    "\n",
    "    accuracy = model_history.history['accuracy']\n",
    "    val_accuracy = model_history.history['val_accuracy']\n",
    "\n",
    "    plt.figure(figsize=(14, 4))\n",
    "\n",
    "    plt.subplot(1, 2, 1)\n",
    "    plt.title('Loss')\n",
    "    plt.xlabel('Epoch')\n",
    "    plt.ylabel('Loss')\n",
    "    plt.plot(loss, label='Training set')\n",
    "    plt.plot(val_loss, label='Test set', linestyle='--')\n",
    "    plt.legend()\n",
    "    plt.grid(linestyle='--', linewidth=1, alpha=0.5)\n",
    "\n",
    "    plt.subplot(1, 2, 2)\n",
    "    plt.title('Accuracy')\n",
    "    plt.xlabel('Epoch')\n",
    "    plt.ylabel('Accuracy')\n",
    "    plt.ylim([0, 1])\n",
    "    plt.plot(accuracy, label='Training set')\n",
    "    plt.plot(val_accuracy, label='Test set', linestyle='--')\n",
    "    plt.legend()\n",
    "    plt.grid(linestyle='--', linewidth=1, alpha=0.5)\n",
    "\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "render_training_history(model_history)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 7. Loading the best model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import load_model\n",
    "# Load the model for later use\n",
    "# model=load_model('best_model.hdf5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "def predict(audio):\n",
    "    prob=model.predict(audio.reshape(1,8000,1))\n",
    "    index=np.argmax(prob[0])\n",
    "    return classes[index]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 8. Evaluate and Predict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Audio: off\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "\n",
       "                <audio  controls=\"controls\" >\n",
       "                    <source src=\"data:audio/wav;base64,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\" type=\"audio/wav\" />\n",
       "                    Your browser does not support the audio element.\n",
       "                </audio>\n",
       "              "
      ],
      "text/plain": [
       "<IPython.lib.display.Audio object>"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index=random.randint(0,len(x_val)-1)\n",
    "samples=x_val[index].ravel()\n",
    "print(\"Audio:\",classes[np.argmax(y_val[index])])\n",
    "ipd.Audio(samples, rate=8000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Text: off\n"
     ]
    }
   ],
   "source": [
    "print(\"Text:\",predict(samples))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> The best part is yet to come! Here is a script that prompts a user to record voice commands. Record your own voice commands and test it on the model:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sounddevice as sd\n",
    "import soundfile as sf\n",
    "\n",
    "samplerate = 16000  \n",
    "duration = 1 # seconds\n",
    "filename = 'yes.wav'\n",
    "print(\"start\")\n",
    "mydata = sd.rec(int(samplerate * duration), samplerate=samplerate,\n",
    "    channels=1, blocking=True)\n",
    "print(\"end\")\n",
    "sd.wait()\n",
    "sf.write(filename, mydata, samplerate)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 9. Use my own voice record and convert it to text:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "os.listdir('../input/voice-commands/prateek_voice_v2')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "filepath='../input/voice-commands/prateek_voice_v2'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#reading the voice commands\n",
    "samples, sample_rate = librosa.load(filepath + '/' + 'stop.wav', sr = 16000)\n",
    "samples = librosa.resample(samples, sample_rate, 8000)\n",
    "ipd.Audio(samples,rate=8000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#converting voice commands to text\n",
    "predict(samples)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Reference\n",
    "#### https://www.assemblyai.com/blog/end-to-end-speech-recognition-pytorch\n",
    "#### https://www.kaggle.com/c/tensorflow-speech-recognition-challenge/overview "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Helpful Article \n",
    "> https://www.analyticsvidhya.com/blog/2019/07/learn-build-first-speech-to-text-model-python/"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}