net.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "confirmed-mother",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index(['valence', 'year', 'acousticness', 'artists', 'danceability',\n",
      "       'duration_ms', 'energy', 'explicit', 'id', 'instrumentalness', 'key',\n",
      "       'liveness', 'loudness', 'mode', 'name', 'popularity', 'release_date',\n",
      "       'speechiness', 'tempo'],\n",
      "      dtype='object')\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import plotly.express as px\n",
    "df = pd.read_csv('data.csv')\n",
    "print(df.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "informed-american",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/jupyter/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:516: FutureWarning:\n",
      "\n",
      "Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "\n",
      "/home/jupyter/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:517: FutureWarning:\n",
      "\n",
      "Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "\n",
      "/home/jupyter/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:518: FutureWarning:\n",
      "\n",
      "Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "\n",
      "/home/jupyter/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:519: FutureWarning:\n",
      "\n",
      "Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "\n",
      "/home/jupyter/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:520: FutureWarning:\n",
      "\n",
      "Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "\n",
      "/home/jupyter/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:525: FutureWarning:\n",
      "\n",
      "Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "\n",
      "/home/jupyter/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:541: FutureWarning:\n",
      "\n",
      "Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "\n",
      "/home/jupyter/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:542: FutureWarning:\n",
      "\n",
      "Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "\n",
      "/home/jupyter/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:543: FutureWarning:\n",
      "\n",
      "Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "\n",
      "/home/jupyter/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:544: FutureWarning:\n",
      "\n",
      "Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "\n",
      "/home/jupyter/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:545: FutureWarning:\n",
      "\n",
      "Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "\n",
      "/home/jupyter/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:550: FutureWarning:\n",
      "\n",
      "Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "%load_ext tensorboard\n",
    "import datetime\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras import layers\n",
    "log_dir = \"log/fit/\" + datetime.datetime.now().strftime(\"%Y%m%d-%H%M%S\")\n",
    "tensorboard_callback = tf.keras.callbacks.TensorBoard(log_dir=log_dir, histogram_freq=1)\n",
    "#%tensorboard --logdir logs/fit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "political-victoria",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1\n",
      "Rush\n"
     ]
    }
   ],
   "source": [
    "filtered_df = df.copy()\n",
    "filtered_df['artists'] = filtered_df['artists'].apply(lambda x: x[1:-1].replace(\"'\", \"\"))\n",
    "art = filtered_df['artists'].sample()\n",
    "print(len(art))\n",
    "print(art.values[0])#.index[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "internal-lebanon",
   "metadata": {},
   "outputs": [],
   "source": [
    "filtered_df = filtered_df.drop(columns=['id']) #we dont need the spotify id's \n",
    "# remastered_list = [x for x in filtered_df.name if \"remaster\" in x.lower()]\n",
    "# len(remastered_list)\n",
    "# filtered_df = filtered_df[~filtered_df.name.isin(remastered_list)] # all the songs without the remastered \n",
    "\n",
    "# #are there any songs with the named remaster im discriminating against?\n",
    "# #should i just get rid of everything remastered?\n",
    "\n",
    "# skits_list = [x for x in filtered_df.name if \"skit\" in x.lower()]\n",
    "# filtered_df = filtered_df[~filtered_df.name.isin(skits_list)]\n",
    "# #28 skits a few are actual songs but idt they will be impactful\n",
    "\n",
    "# demo_list = [x for x in filtered_df.name if \"demo \" in x.lower()]\n",
    "# filtered_df = filtered_df[~filtered_df.name.isin(demo_list)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "grand-frank",
   "metadata": {},
   "outputs": [],
   "source": [
    "filenames = ['classical_idx.csv','hip-hop_idx.csv','jazz_idx.csv','metal_idx.csv','reggae_idx.csv','rock_idx.csv']\n",
    "\n",
    "gen = pd.DataFrame(0, index=np.arange(filtered_df.shape[0]),columns=['genera'])\n",
    "\n",
    "#filtered_data['genera'] = \n",
    "for n in filenames:\n",
    "    gen.iloc[pd.read_csv('genera_csv/'+n, header=None).iloc[:,0]] = n[:-8]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "dense-stanley",
   "metadata": {},
   "outputs": [],
   "source": [
    "final_df = filtered_df.join(gen)\n",
    "labeled_data = final_df[final_df.genera != 0]\n",
    "unlabeled_data = final_df[final_df.genera == 0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "exotic-express",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "exotic-toyota",
   "metadata": {},
   "outputs": [],
   "source": [
    "y = labeled_data.genera\n",
    "features_list = ['valence', 'acousticness', 'danceability', 'duration_ms', 'energy', 'instrumentalness', 'key', 'loudness', 'mode', 'popularity', 'speechiness', 'tempo']\n",
    "X = labeled_data.copy()#[features_list].copy()\n",
    "Z = unlabeled_data.copy()#[features_list].copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "radio-bacteria",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "137       classical\n",
       "229            jazz\n",
       "806            jazz\n",
       "814            jazz\n",
       "1409           jazz\n",
       "            ...    \n",
       "167324       reggae\n",
       "167516       reggae\n",
       "168592         rock\n",
       "169414         jazz\n",
       "169454    classical\n",
       "Name: genera, Length: 606, dtype: object"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.preprocessing import MinMaxScaler\n",
    "to_norm_list = ['duration_ms', 'key','loudness', 'popularity', 'tempo']\n",
    "\n",
    "scaler = MinMaxScaler(feature_range=(0, 1))\n",
    "d = scaler.fit_transform(X[to_norm_list])\n",
    "scaled_df = pd.DataFrame(d, columns=to_norm_list)\n",
    "scaled_df.head()\n",
    "X.drop(to_norm_list, axis=1,inplace=True)\n",
    "X = X.reset_index(drop=True)\n",
    "X[to_norm_list] = scaled_df\n",
    "y\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "specialized-swaziland",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "7262           jazz\n",
       "15580       hip-hop\n",
       "14021       hip-hop\n",
       "10810          rock\n",
       "10045        reggae\n",
       "73087     classical\n",
       "11582          rock\n",
       "24431     classical\n",
       "48422     classical\n",
       "11610          rock\n",
       "14099          rock\n",
       "13384          rock\n",
       "13259       hip-hop\n",
       "10793        reggae\n",
       "101396       reggae\n",
       "70978          jazz\n",
       "5478      classical\n",
       "52346       hip-hop\n",
       "165064       reggae\n",
       "50918        reggae\n",
       "Name: genera, dtype: object"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y.sample(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "confident-philippines",
   "metadata": {},
   "outputs": [],
   "source": [
    "y = pd.get_dummies(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "determined-adapter",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "X_train, X_test, y_train, y_test = train_test_split(X[features_list], y,test_size=0.2)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "southern-question",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "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>valence</th>\n",
       "      <th>year</th>\n",
       "      <th>acousticness</th>\n",
       "      <th>artists</th>\n",
       "      <th>danceability</th>\n",
       "      <th>energy</th>\n",
       "      <th>explicit</th>\n",
       "      <th>instrumentalness</th>\n",
       "      <th>liveness</th>\n",
       "      <th>mode</th>\n",
       "      <th>name</th>\n",
       "      <th>release_date</th>\n",
       "      <th>speechiness</th>\n",
       "      <th>genera</th>\n",
       "      <th>duration_ms</th>\n",
       "      <th>key</th>\n",
       "      <th>loudness</th>\n",
       "      <th>popularity</th>\n",
       "      <th>tempo</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0594</td>\n",
       "      <td>1921</td>\n",
       "      <td>0.98200</td>\n",
       "      <td>Sergei Rachmaninoff, James Levine, Berliner Ph...</td>\n",
       "      <td>0.279</td>\n",
       "      <td>0.211</td>\n",
       "      <td>0</td>\n",
       "      <td>0.878000</td>\n",
       "      <td>0.6650</td>\n",
       "      <td>1</td>\n",
       "      <td>Piano Concerto No. 3 in D Minor, Op. 30: III. ...</td>\n",
       "      <td>1921</td>\n",
       "      <td>0.0366</td>\n",
       "      <td>0</td>\n",
       "      <td>0.153112</td>\n",
       "      <td>0.909091</td>\n",
       "      <td>0.624916</td>\n",
       "      <td>0.04</td>\n",
       "      <td>0.332450</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.9630</td>\n",
       "      <td>1921</td>\n",
       "      <td>0.73200</td>\n",
       "      <td>Dennis Day</td>\n",
       "      <td>0.819</td>\n",
       "      <td>0.341</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.1600</td>\n",
       "      <td>1</td>\n",
       "      <td>Clancy Lowered the Boom</td>\n",
       "      <td>1921</td>\n",
       "      <td>0.4150</td>\n",
       "      <td>0</td>\n",
       "      <td>0.032496</td>\n",
       "      <td>0.636364</td>\n",
       "      <td>0.744797</td>\n",
       "      <td>0.05</td>\n",
       "      <td>0.250243</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0394</td>\n",
       "      <td>1921</td>\n",
       "      <td>0.96100</td>\n",
       "      <td>KHP Kridhamardawa Karaton Ngayogyakarta Hadini...</td>\n",
       "      <td>0.328</td>\n",
       "      <td>0.166</td>\n",
       "      <td>0</td>\n",
       "      <td>0.913000</td>\n",
       "      <td>0.1010</td>\n",
       "      <td>1</td>\n",
       "      <td>Gati Bali</td>\n",
       "      <td>1921</td>\n",
       "      <td>0.0339</td>\n",
       "      <td>0</td>\n",
       "      <td>0.091685</td>\n",
       "      <td>0.272727</td>\n",
       "      <td>0.707071</td>\n",
       "      <td>0.05</td>\n",
       "      <td>0.453125</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.1650</td>\n",
       "      <td>1921</td>\n",
       "      <td>0.96700</td>\n",
       "      <td>Frank Parker</td>\n",
       "      <td>0.275</td>\n",
       "      <td>0.309</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000028</td>\n",
       "      <td>0.3810</td>\n",
       "      <td>1</td>\n",
       "      <td>Danny Boy</td>\n",
       "      <td>1921</td>\n",
       "      <td>0.0354</td>\n",
       "      <td>0</td>\n",
       "      <td>0.037954</td>\n",
       "      <td>0.454545</td>\n",
       "      <td>0.793736</td>\n",
       "      <td>0.03</td>\n",
       "      <td>0.411113</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.2530</td>\n",
       "      <td>1921</td>\n",
       "      <td>0.95700</td>\n",
       "      <td>Phil Regan</td>\n",
       "      <td>0.418</td>\n",
       "      <td>0.193</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>0.2290</td>\n",
       "      <td>1</td>\n",
       "      <td>When Irish Eyes Are Smiling</td>\n",
       "      <td>1921</td>\n",
       "      <td>0.0380</td>\n",
       "      <td>0</td>\n",
       "      <td>0.029932</td>\n",
       "      <td>0.272727</td>\n",
       "      <td>0.781521</td>\n",
       "      <td>0.02</td>\n",
       "      <td>0.417503</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>170042</th>\n",
       "      <td>0.6080</td>\n",
       "      <td>2020</td>\n",
       "      <td>0.08460</td>\n",
       "      <td>Anuel AA, Daddy Yankee, KAROL G, Ozuna, J Balvin</td>\n",
       "      <td>0.786</td>\n",
       "      <td>0.808</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000289</td>\n",
       "      <td>0.0822</td>\n",
       "      <td>1</td>\n",
       "      <td>China</td>\n",
       "      <td>2020-05-29</td>\n",
       "      <td>0.0881</td>\n",
       "      <td>0</td>\n",
       "      <td>0.054943</td>\n",
       "      <td>0.636364</td>\n",
       "      <td>0.881654</td>\n",
       "      <td>0.72</td>\n",
       "      <td>0.431318</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>170043</th>\n",
       "      <td>0.7340</td>\n",
       "      <td>2020</td>\n",
       "      <td>0.20600</td>\n",
       "      <td>Ashnikko</td>\n",
       "      <td>0.717</td>\n",
       "      <td>0.753</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.1010</td>\n",
       "      <td>1</td>\n",
       "      <td>Halloweenie III: Seven Days</td>\n",
       "      <td>2020-10-23</td>\n",
       "      <td>0.0605</td>\n",
       "      <td>0</td>\n",
       "      <td>0.026961</td>\n",
       "      <td>0.636364</td>\n",
       "      <td>0.845353</td>\n",
       "      <td>0.68</td>\n",
       "      <td>0.566456</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>170044</th>\n",
       "      <td>0.6370</td>\n",
       "      <td>2020</td>\n",
       "      <td>0.10100</td>\n",
       "      <td>MAMAMOO</td>\n",
       "      <td>0.634</td>\n",
       "      <td>0.858</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000009</td>\n",
       "      <td>0.2580</td>\n",
       "      <td>0</td>\n",
       "      <td>AYA</td>\n",
       "      <td>2020-11-03</td>\n",
       "      <td>0.0809</td>\n",
       "      <td>0</td>\n",
       "      <td>0.038191</td>\n",
       "      <td>0.363636</td>\n",
       "      <td>0.904769</td>\n",
       "      <td>0.76</td>\n",
       "      <td>0.376531</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>170045</th>\n",
       "      <td>0.1950</td>\n",
       "      <td>2020</td>\n",
       "      <td>0.00998</td>\n",
       "      <td>Eminem</td>\n",
       "      <td>0.671</td>\n",
       "      <td>0.623</td>\n",
       "      <td>1</td>\n",
       "      <td>0.000008</td>\n",
       "      <td>0.6430</td>\n",
       "      <td>1</td>\n",
       "      <td>Darkness</td>\n",
       "      <td>2020-01-17</td>\n",
       "      <td>0.3080</td>\n",
       "      <td>0</td>\n",
       "      <td>0.061507</td>\n",
       "      <td>0.181818</td>\n",
       "      <td>0.827484</td>\n",
       "      <td>0.70</td>\n",
       "      <td>0.308225</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>170046</th>\n",
       "      <td>0.6420</td>\n",
       "      <td>2020</td>\n",
       "      <td>0.13200</td>\n",
       "      <td>KEVVO, J Balvin</td>\n",
       "      <td>0.856</td>\n",
       "      <td>0.721</td>\n",
       "      <td>1</td>\n",
       "      <td>0.004710</td>\n",
       "      <td>0.1820</td>\n",
       "      <td>1</td>\n",
       "      <td>Billetes Azules (with J Balvin)</td>\n",
       "      <td>2020-10-16</td>\n",
       "      <td>0.1080</td>\n",
       "      <td>0</td>\n",
       "      <td>0.034158</td>\n",
       "      <td>0.636364</td>\n",
       "      <td>0.862454</td>\n",
       "      <td>0.74</td>\n",
       "      <td>0.390096</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>170047 rows × 19 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        valence  year  acousticness  \\\n",
       "0        0.0594  1921       0.98200   \n",
       "1        0.9630  1921       0.73200   \n",
       "2        0.0394  1921       0.96100   \n",
       "3        0.1650  1921       0.96700   \n",
       "4        0.2530  1921       0.95700   \n",
       "...         ...   ...           ...   \n",
       "170042   0.6080  2020       0.08460   \n",
       "170043   0.7340  2020       0.20600   \n",
       "170044   0.6370  2020       0.10100   \n",
       "170045   0.1950  2020       0.00998   \n",
       "170046   0.6420  2020       0.13200   \n",
       "\n",
       "                                                  artists  danceability  \\\n",
       "0       Sergei Rachmaninoff, James Levine, Berliner Ph...         0.279   \n",
       "1                                              Dennis Day         0.819   \n",
       "2       KHP Kridhamardawa Karaton Ngayogyakarta Hadini...         0.328   \n",
       "3                                            Frank Parker         0.275   \n",
       "4                                              Phil Regan         0.418   \n",
       "...                                                   ...           ...   \n",
       "170042   Anuel AA, Daddy Yankee, KAROL G, Ozuna, J Balvin         0.786   \n",
       "170043                                           Ashnikko         0.717   \n",
       "170044                                            MAMAMOO         0.634   \n",
       "170045                                             Eminem         0.671   \n",
       "170046                                    KEVVO, J Balvin         0.856   \n",
       "\n",
       "        energy  explicit  instrumentalness  liveness  mode  \\\n",
       "0        0.211         0          0.878000    0.6650     1   \n",
       "1        0.341         0          0.000000    0.1600     1   \n",
       "2        0.166         0          0.913000    0.1010     1   \n",
       "3        0.309         0          0.000028    0.3810     1   \n",
       "4        0.193         0          0.000002    0.2290     1   \n",
       "...        ...       ...               ...       ...   ...   \n",
       "170042   0.808         0          0.000289    0.0822     1   \n",
       "170043   0.753         0          0.000000    0.1010     1   \n",
       "170044   0.858         0          0.000009    0.2580     0   \n",
       "170045   0.623         1          0.000008    0.6430     1   \n",
       "170046   0.721         1          0.004710    0.1820     1   \n",
       "\n",
       "                                                     name release_date  \\\n",
       "0       Piano Concerto No. 3 in D Minor, Op. 30: III. ...         1921   \n",
       "1                                 Clancy Lowered the Boom         1921   \n",
       "2                                               Gati Bali         1921   \n",
       "3                                               Danny Boy         1921   \n",
       "4                             When Irish Eyes Are Smiling         1921   \n",
       "...                                                   ...          ...   \n",
       "170042                                              China   2020-05-29   \n",
       "170043                        Halloweenie III: Seven Days   2020-10-23   \n",
       "170044                                                AYA   2020-11-03   \n",
       "170045                                           Darkness   2020-01-17   \n",
       "170046                    Billetes Azules (with J Balvin)   2020-10-16   \n",
       "\n",
       "        speechiness genera  duration_ms       key  loudness  popularity  \\\n",
       "0            0.0366      0     0.153112  0.909091  0.624916        0.04   \n",
       "1            0.4150      0     0.032496  0.636364  0.744797        0.05   \n",
       "2            0.0339      0     0.091685  0.272727  0.707071        0.05   \n",
       "3            0.0354      0     0.037954  0.454545  0.793736        0.03   \n",
       "4            0.0380      0     0.029932  0.272727  0.781521        0.02   \n",
       "...             ...    ...          ...       ...       ...         ...   \n",
       "170042       0.0881      0     0.054943  0.636364  0.881654        0.72   \n",
       "170043       0.0605      0     0.026961  0.636364  0.845353        0.68   \n",
       "170044       0.0809      0     0.038191  0.363636  0.904769        0.76   \n",
       "170045       0.3080      0     0.061507  0.181818  0.827484        0.70   \n",
       "170046       0.1080      0     0.034158  0.636364  0.862454        0.74   \n",
       "\n",
       "           tempo  \n",
       "0       0.332450  \n",
       "1       0.250243  \n",
       "2       0.453125  \n",
       "3       0.411113  \n",
       "4       0.417503  \n",
       "...          ...  \n",
       "170042  0.431318  \n",
       "170043  0.566456  \n",
       "170044  0.376531  \n",
       "170045  0.308225  \n",
       "170046  0.390096  \n",
       "\n",
       "[170047 rows x 19 columns]"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "d = scaler.fit_transform(Z[to_norm_list])\n",
    "scaled_df = pd.DataFrame(d, columns=to_norm_list)\n",
    "scaled_df.head()\n",
    "Z.drop(to_norm_list, axis=1,inplace=True)\n",
    "Z = Z.reset_index(drop=True)\n",
    "Z[to_norm_list] = scaled_df\n",
    "Z"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "id": "flush-guyana",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "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>valence</th>\n",
       "      <th>year</th>\n",
       "      <th>acousticness</th>\n",
       "      <th>artists</th>\n",
       "      <th>danceability</th>\n",
       "      <th>energy</th>\n",
       "      <th>explicit</th>\n",
       "      <th>instrumentalness</th>\n",
       "      <th>liveness</th>\n",
       "      <th>mode</th>\n",
       "      <th>name</th>\n",
       "      <th>release_date</th>\n",
       "      <th>speechiness</th>\n",
       "      <th>genera</th>\n",
       "      <th>duration_ms</th>\n",
       "      <th>key</th>\n",
       "      <th>loudness</th>\n",
       "      <th>popularity</th>\n",
       "      <th>tempo</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0594</td>\n",
       "      <td>1921</td>\n",
       "      <td>0.98200</td>\n",
       "      <td>Sergei Rachmaninoff, James Levine, Berliner Ph...</td>\n",
       "      <td>0.279</td>\n",
       "      <td>0.211</td>\n",
       "      <td>0</td>\n",
       "      <td>0.878000</td>\n",
       "      <td>0.6650</td>\n",
       "      <td>1</td>\n",
       "      <td>Piano Concerto No. 3 in D Minor, Op. 30: III. ...</td>\n",
       "      <td>1921</td>\n",
       "      <td>0.0366</td>\n",
       "      <td>0</td>\n",
       "      <td>0.153112</td>\n",
       "      <td>0.909091</td>\n",
       "      <td>0.624916</td>\n",
       "      <td>0.04</td>\n",
       "      <td>0.332450</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.9630</td>\n",
       "      <td>1921</td>\n",
       "      <td>0.73200</td>\n",
       "      <td>Dennis Day</td>\n",
       "      <td>0.819</td>\n",
       "      <td>0.341</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.1600</td>\n",
       "      <td>1</td>\n",
       "      <td>Clancy Lowered the Boom</td>\n",
       "      <td>1921</td>\n",
       "      <td>0.4150</td>\n",
       "      <td>0</td>\n",
       "      <td>0.032496</td>\n",
       "      <td>0.636364</td>\n",
       "      <td>0.744797</td>\n",
       "      <td>0.05</td>\n",
       "      <td>0.250243</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0394</td>\n",
       "      <td>1921</td>\n",
       "      <td>0.96100</td>\n",
       "      <td>KHP Kridhamardawa Karaton Ngayogyakarta Hadini...</td>\n",
       "      <td>0.328</td>\n",
       "      <td>0.166</td>\n",
       "      <td>0</td>\n",
       "      <td>0.913000</td>\n",
       "      <td>0.1010</td>\n",
       "      <td>1</td>\n",
       "      <td>Gati Bali</td>\n",
       "      <td>1921</td>\n",
       "      <td>0.0339</td>\n",
       "      <td>0</td>\n",
       "      <td>0.091685</td>\n",
       "      <td>0.272727</td>\n",
       "      <td>0.707071</td>\n",
       "      <td>0.05</td>\n",
       "      <td>0.453125</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.1650</td>\n",
       "      <td>1921</td>\n",
       "      <td>0.96700</td>\n",
       "      <td>Frank Parker</td>\n",
       "      <td>0.275</td>\n",
       "      <td>0.309</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000028</td>\n",
       "      <td>0.3810</td>\n",
       "      <td>1</td>\n",
       "      <td>Danny Boy</td>\n",
       "      <td>1921</td>\n",
       "      <td>0.0354</td>\n",
       "      <td>0</td>\n",
       "      <td>0.037954</td>\n",
       "      <td>0.454545</td>\n",
       "      <td>0.793736</td>\n",
       "      <td>0.03</td>\n",
       "      <td>0.411113</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.2530</td>\n",
       "      <td>1921</td>\n",
       "      <td>0.95700</td>\n",
       "      <td>Phil Regan</td>\n",
       "      <td>0.418</td>\n",
       "      <td>0.193</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>0.2290</td>\n",
       "      <td>1</td>\n",
       "      <td>When Irish Eyes Are Smiling</td>\n",
       "      <td>1921</td>\n",
       "      <td>0.0380</td>\n",
       "      <td>0</td>\n",
       "      <td>0.029932</td>\n",
       "      <td>0.272727</td>\n",
       "      <td>0.781521</td>\n",
       "      <td>0.02</td>\n",
       "      <td>0.417503</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>170648</th>\n",
       "      <td>0.6080</td>\n",
       "      <td>2020</td>\n",
       "      <td>0.08460</td>\n",
       "      <td>Anuel AA, Daddy Yankee, KAROL G, Ozuna, J Balvin</td>\n",
       "      <td>0.786</td>\n",
       "      <td>0.808</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000289</td>\n",
       "      <td>0.0822</td>\n",
       "      <td>1</td>\n",
       "      <td>China</td>\n",
       "      <td>2020-05-29</td>\n",
       "      <td>0.0881</td>\n",
       "      <td>0</td>\n",
       "      <td>0.054943</td>\n",
       "      <td>0.636364</td>\n",
       "      <td>0.881654</td>\n",
       "      <td>0.72</td>\n",
       "      <td>0.431318</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>170649</th>\n",
       "      <td>0.7340</td>\n",
       "      <td>2020</td>\n",
       "      <td>0.20600</td>\n",
       "      <td>Ashnikko</td>\n",
       "      <td>0.717</td>\n",
       "      <td>0.753</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.1010</td>\n",
       "      <td>1</td>\n",
       "      <td>Halloweenie III: Seven Days</td>\n",
       "      <td>2020-10-23</td>\n",
       "      <td>0.0605</td>\n",
       "      <td>0</td>\n",
       "      <td>0.026961</td>\n",
       "      <td>0.636364</td>\n",
       "      <td>0.845353</td>\n",
       "      <td>0.68</td>\n",
       "      <td>0.566456</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>170650</th>\n",
       "      <td>0.6370</td>\n",
       "      <td>2020</td>\n",
       "      <td>0.10100</td>\n",
       "      <td>MAMAMOO</td>\n",
       "      <td>0.634</td>\n",
       "      <td>0.858</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000009</td>\n",
       "      <td>0.2580</td>\n",
       "      <td>0</td>\n",
       "      <td>AYA</td>\n",
       "      <td>2020-11-03</td>\n",
       "      <td>0.0809</td>\n",
       "      <td>0</td>\n",
       "      <td>0.038191</td>\n",
       "      <td>0.363636</td>\n",
       "      <td>0.904769</td>\n",
       "      <td>0.76</td>\n",
       "      <td>0.376531</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>170651</th>\n",
       "      <td>0.1950</td>\n",
       "      <td>2020</td>\n",
       "      <td>0.00998</td>\n",
       "      <td>Eminem</td>\n",
       "      <td>0.671</td>\n",
       "      <td>0.623</td>\n",
       "      <td>1</td>\n",
       "      <td>0.000008</td>\n",
       "      <td>0.6430</td>\n",
       "      <td>1</td>\n",
       "      <td>Darkness</td>\n",
       "      <td>2020-01-17</td>\n",
       "      <td>0.3080</td>\n",
       "      <td>0</td>\n",
       "      <td>0.061507</td>\n",
       "      <td>0.181818</td>\n",
       "      <td>0.827484</td>\n",
       "      <td>0.70</td>\n",
       "      <td>0.308225</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>170652</th>\n",
       "      <td>0.6420</td>\n",
       "      <td>2020</td>\n",
       "      <td>0.13200</td>\n",
       "      <td>KEVVO, J Balvin</td>\n",
       "      <td>0.856</td>\n",
       "      <td>0.721</td>\n",
       "      <td>1</td>\n",
       "      <td>0.004710</td>\n",
       "      <td>0.1820</td>\n",
       "      <td>1</td>\n",
       "      <td>Billetes Azules (with J Balvin)</td>\n",
       "      <td>2020-10-16</td>\n",
       "      <td>0.1080</td>\n",
       "      <td>0</td>\n",
       "      <td>0.034158</td>\n",
       "      <td>0.636364</td>\n",
       "      <td>0.862454</td>\n",
       "      <td>0.74</td>\n",
       "      <td>0.390096</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>170653 rows × 19 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        valence  year  acousticness  \\\n",
       "0        0.0594  1921       0.98200   \n",
       "1        0.9630  1921       0.73200   \n",
       "2        0.0394  1921       0.96100   \n",
       "3        0.1650  1921       0.96700   \n",
       "4        0.2530  1921       0.95700   \n",
       "...         ...   ...           ...   \n",
       "170648   0.6080  2020       0.08460   \n",
       "170649   0.7340  2020       0.20600   \n",
       "170650   0.6370  2020       0.10100   \n",
       "170651   0.1950  2020       0.00998   \n",
       "170652   0.6420  2020       0.13200   \n",
       "\n",
       "                                                  artists  danceability  \\\n",
       "0       Sergei Rachmaninoff, James Levine, Berliner Ph...         0.279   \n",
       "1                                              Dennis Day         0.819   \n",
       "2       KHP Kridhamardawa Karaton Ngayogyakarta Hadini...         0.328   \n",
       "3                                            Frank Parker         0.275   \n",
       "4                                              Phil Regan         0.418   \n",
       "...                                                   ...           ...   \n",
       "170648   Anuel AA, Daddy Yankee, KAROL G, Ozuna, J Balvin         0.786   \n",
       "170649                                           Ashnikko         0.717   \n",
       "170650                                            MAMAMOO         0.634   \n",
       "170651                                             Eminem         0.671   \n",
       "170652                                    KEVVO, J Balvin         0.856   \n",
       "\n",
       "        energy  explicit  instrumentalness  liveness  mode  \\\n",
       "0        0.211         0          0.878000    0.6650     1   \n",
       "1        0.341         0          0.000000    0.1600     1   \n",
       "2        0.166         0          0.913000    0.1010     1   \n",
       "3        0.309         0          0.000028    0.3810     1   \n",
       "4        0.193         0          0.000002    0.2290     1   \n",
       "...        ...       ...               ...       ...   ...   \n",
       "170648   0.808         0          0.000289    0.0822     1   \n",
       "170649   0.753         0          0.000000    0.1010     1   \n",
       "170650   0.858         0          0.000009    0.2580     0   \n",
       "170651   0.623         1          0.000008    0.6430     1   \n",
       "170652   0.721         1          0.004710    0.1820     1   \n",
       "\n",
       "                                                     name release_date  \\\n",
       "0       Piano Concerto No. 3 in D Minor, Op. 30: III. ...         1921   \n",
       "1                                 Clancy Lowered the Boom         1921   \n",
       "2                                               Gati Bali         1921   \n",
       "3                                               Danny Boy         1921   \n",
       "4                             When Irish Eyes Are Smiling         1921   \n",
       "...                                                   ...          ...   \n",
       "170648                                              China   2020-05-29   \n",
       "170649                        Halloweenie III: Seven Days   2020-10-23   \n",
       "170650                                                AYA   2020-11-03   \n",
       "170651                                           Darkness   2020-01-17   \n",
       "170652                    Billetes Azules (with J Balvin)   2020-10-16   \n",
       "\n",
       "        speechiness genera  duration_ms       key  loudness  popularity  \\\n",
       "0            0.0366      0     0.153112  0.909091  0.624916        0.04   \n",
       "1            0.4150      0     0.032496  0.636364  0.744797        0.05   \n",
       "2            0.0339      0     0.091685  0.272727  0.707071        0.05   \n",
       "3            0.0354      0     0.037954  0.454545  0.793736        0.03   \n",
       "4            0.0380      0     0.029932  0.272727  0.781521        0.02   \n",
       "...             ...    ...          ...       ...       ...         ...   \n",
       "170648       0.0881      0     0.054943  0.636364  0.881654        0.72   \n",
       "170649       0.0605      0     0.026961  0.636364  0.845353        0.68   \n",
       "170650       0.0809      0     0.038191  0.363636  0.904769        0.76   \n",
       "170651       0.3080      0     0.061507  0.181818  0.827484        0.70   \n",
       "170652       0.1080      0     0.034158  0.636364  0.862454        0.74   \n",
       "\n",
       "           tempo  \n",
       "0       0.332450  \n",
       "1       0.250243  \n",
       "2       0.453125  \n",
       "3       0.411113  \n",
       "4       0.417503  \n",
       "...          ...  \n",
       "170648  0.431318  \n",
       "170649  0.566456  \n",
       "170650  0.376531  \n",
       "170651  0.308225  \n",
       "170652  0.390096  \n",
       "\n",
       "[170653 rows x 19 columns]"
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "af_df = final_df.copy()\n",
    "d = scaler.fit_transform(af_df[to_norm_list])\n",
    "scaled_df = pd.DataFrame(d, columns=to_norm_list)\n",
    "scaled_df.head()\n",
    "af_df.drop(to_norm_list, axis=1,inplace=True)\n",
    "af_df = af_df.reset_index(drop=True)\n",
    "af_df[to_norm_list] = scaled_df\n",
    "af_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "facial-intelligence",
   "metadata": {},
   "outputs": [],
   "source": [
    "early_stop = tf.keras.callbacks.EarlyStopping(min_delta=0.01, patience=15, restore_best_weights=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "sharing-explanation",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /home/jupyter/.local/lib/python3.6/site-packages/tensorflow/python/ops/init_ops.py:1251: calling VarianceScaling.__init__ (from tensorflow.python.ops.init_ops) with dtype is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Call initializer instance with the dtype argument instead of passing it to the constructor\n",
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense (Dense)                (None, 12)                156       \n",
      "_________________________________________________________________\n",
      "batch_normalization (BatchNo (None, 12)                48        \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 300)               3900      \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 300)               0         \n",
      "_________________________________________________________________\n",
      "batch_normalization_1 (Batch (None, 300)               1200      \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 6)                 1806      \n",
      "=================================================================\n",
      "Total params: 7,110\n",
      "Trainable params: 6,486\n",
      "Non-trainable params: 624\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "\n",
    "\n",
    "model = tf.keras.Sequential()\n",
    "model.add(layers.Dense(units=12, input_shape=(len(features_list),)))\n",
    "model.add(layers.BatchNormalization())\n",
    "model.add(layers.Dense(units=300, activation='sigmoid'))\n",
    "model.add(layers.Dropout(0.2))\n",
    "model.add(layers.BatchNormalization())\n",
    "\n",
    "# model.add(layers.ReLU())\n",
    "model.add(layers.Dense(units=6, activation='softmax'))\n",
    "opt = tf.keras.optimizers.Adam(learning_rate=0.0003)\n",
    "model.compile(loss='categorical_crossentropy', optimizer=opt, metrics=['categorical_accuracy', 'mse'])\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "supreme-perry",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 484 samples, validate on 122 samples\n",
      "Epoch 1/400\n",
      "484/484 [==============================] - 0s 601us/sample - loss: 2.2912 - categorical_accuracy: 0.1860 - mean_squared_error: 0.1657 - val_loss: 1.7793 - val_categorical_accuracy: 0.1639 - val_mean_squared_error: 0.1383\n",
      "Epoch 2/400\n",
      "484/484 [==============================] - 0s 117us/sample - loss: 2.1405 - categorical_accuracy: 0.2107 - mean_squared_error: 0.1586 - val_loss: 1.7270 - val_categorical_accuracy: 0.3525 - val_mean_squared_error: 0.1351\n",
      "Epoch 3/400\n",
      "484/484 [==============================] - 0s 138us/sample - loss: 2.0943 - categorical_accuracy: 0.2707 - mean_squared_error: 0.1557 - val_loss: 1.6808 - val_categorical_accuracy: 0.3525 - val_mean_squared_error: 0.1324\n",
      "Epoch 4/400\n",
      "484/484 [==============================] - 0s 151us/sample - loss: 1.8976 - categorical_accuracy: 0.3244 - mean_squared_error: 0.1431 - val_loss: 1.6347 - val_categorical_accuracy: 0.3525 - val_mean_squared_error: 0.1298\n",
      "Epoch 5/400\n",
      "484/484 [==============================] - 0s 144us/sample - loss: 1.7974 - categorical_accuracy: 0.3099 - mean_squared_error: 0.1419 - val_loss: 1.5765 - val_categorical_accuracy: 0.3689 - val_mean_squared_error: 0.1265\n",
      "Epoch 6/400\n",
      "484/484 [==============================] - 0s 136us/sample - loss: 1.6882 - categorical_accuracy: 0.3554 - mean_squared_error: 0.1351 - val_loss: 1.5090 - val_categorical_accuracy: 0.4016 - val_mean_squared_error: 0.1225\n",
      "Epoch 7/400\n",
      "484/484 [==============================] - 0s 126us/sample - loss: 1.6802 - categorical_accuracy: 0.3926 - mean_squared_error: 0.1325 - val_loss: 1.4335 - val_categorical_accuracy: 0.4672 - val_mean_squared_error: 0.1179\n",
      "Epoch 8/400\n",
      "484/484 [==============================] - 0s 119us/sample - loss: 1.5961 - categorical_accuracy: 0.3657 - mean_squared_error: 0.1313 - val_loss: 1.3439 - val_categorical_accuracy: 0.5164 - val_mean_squared_error: 0.1118\n",
      "Epoch 9/400\n",
      "484/484 [==============================] - 0s 133us/sample - loss: 1.5428 - categorical_accuracy: 0.4256 - mean_squared_error: 0.1218 - val_loss: 1.2642 - val_categorical_accuracy: 0.5328 - val_mean_squared_error: 0.1063\n",
      "Epoch 10/400\n",
      "484/484 [==============================] - 0s 124us/sample - loss: 1.5156 - categorical_accuracy: 0.4607 - mean_squared_error: 0.1210 - val_loss: 1.1834 - val_categorical_accuracy: 0.6148 - val_mean_squared_error: 0.1005\n",
      "Epoch 11/400\n",
      "484/484 [==============================] - 0s 120us/sample - loss: 1.5220 - categorical_accuracy: 0.4298 - mean_squared_error: 0.1252 - val_loss: 1.1028 - val_categorical_accuracy: 0.6557 - val_mean_squared_error: 0.0943\n",
      "Epoch 12/400\n",
      "484/484 [==============================] - 0s 120us/sample - loss: 1.3619 - categorical_accuracy: 0.4690 - mean_squared_error: 0.1139 - val_loss: 1.0298 - val_categorical_accuracy: 0.6967 - val_mean_squared_error: 0.0886\n",
      "Epoch 13/400\n",
      "484/484 [==============================] - 0s 132us/sample - loss: 1.3151 - categorical_accuracy: 0.4835 - mean_squared_error: 0.1110 - val_loss: 0.9683 - val_categorical_accuracy: 0.7213 - val_mean_squared_error: 0.0839\n",
      "Epoch 14/400\n",
      "484/484 [==============================] - 0s 138us/sample - loss: 1.3370 - categorical_accuracy: 0.4897 - mean_squared_error: 0.1123 - val_loss: 0.9149 - val_categorical_accuracy: 0.7377 - val_mean_squared_error: 0.0796\n",
      "Epoch 15/400\n",
      "484/484 [==============================] - 0s 119us/sample - loss: 1.3725 - categorical_accuracy: 0.5021 - mean_squared_error: 0.1139 - val_loss: 0.8610 - val_categorical_accuracy: 0.7541 - val_mean_squared_error: 0.0751\n",
      "Epoch 16/400\n",
      "484/484 [==============================] - 0s 118us/sample - loss: 1.3646 - categorical_accuracy: 0.4835 - mean_squared_error: 0.1129 - val_loss: 0.8209 - val_categorical_accuracy: 0.7623 - val_mean_squared_error: 0.0719\n",
      "Epoch 17/400\n",
      "484/484 [==============================] - 0s 132us/sample - loss: 1.3113 - categorical_accuracy: 0.4793 - mean_squared_error: 0.1149 - val_loss: 0.7783 - val_categorical_accuracy: 0.7869 - val_mean_squared_error: 0.0681\n",
      "Epoch 18/400\n",
      "484/484 [==============================] - 0s 128us/sample - loss: 1.2639 - categorical_accuracy: 0.4876 - mean_squared_error: 0.1111 - val_loss: 0.7478 - val_categorical_accuracy: 0.7869 - val_mean_squared_error: 0.0656\n",
      "Epoch 19/400\n",
      "484/484 [==============================] - 0s 138us/sample - loss: 1.2203 - categorical_accuracy: 0.5372 - mean_squared_error: 0.1033 - val_loss: 0.7227 - val_categorical_accuracy: 0.7869 - val_mean_squared_error: 0.0637\n",
      "Epoch 20/400\n",
      "484/484 [==============================] - 0s 131us/sample - loss: 1.2040 - categorical_accuracy: 0.5537 - mean_squared_error: 0.1031 - val_loss: 0.7020 - val_categorical_accuracy: 0.7869 - val_mean_squared_error: 0.0623\n",
      "Epoch 21/400\n",
      "484/484 [==============================] - 0s 123us/sample - loss: 1.1880 - categorical_accuracy: 0.5537 - mean_squared_error: 0.1003 - val_loss: 0.6844 - val_categorical_accuracy: 0.8033 - val_mean_squared_error: 0.0606\n",
      "Epoch 22/400\n",
      "484/484 [==============================] - 0s 121us/sample - loss: 1.2130 - categorical_accuracy: 0.5289 - mean_squared_error: 0.1048 - val_loss: 0.6699 - val_categorical_accuracy: 0.8033 - val_mean_squared_error: 0.0596\n",
      "Epoch 23/400\n",
      "484/484 [==============================] - 0s 132us/sample - loss: 1.1092 - categorical_accuracy: 0.5496 - mean_squared_error: 0.0969 - val_loss: 0.6492 - val_categorical_accuracy: 0.8115 - val_mean_squared_error: 0.0577\n",
      "Epoch 24/400\n",
      "484/484 [==============================] - 0s 135us/sample - loss: 1.0932 - categorical_accuracy: 0.5826 - mean_squared_error: 0.0968 - val_loss: 0.6363 - val_categorical_accuracy: 0.8115 - val_mean_squared_error: 0.0567\n",
      "Epoch 25/400\n",
      "484/484 [==============================] - 0s 135us/sample - loss: 1.0428 - categorical_accuracy: 0.5826 - mean_squared_error: 0.0932 - val_loss: 0.6290 - val_categorical_accuracy: 0.8279 - val_mean_squared_error: 0.0563\n",
      "Epoch 26/400\n",
      "484/484 [==============================] - 0s 124us/sample - loss: 1.1095 - categorical_accuracy: 0.5744 - mean_squared_error: 0.0972 - val_loss: 0.6188 - val_categorical_accuracy: 0.8279 - val_mean_squared_error: 0.0556\n",
      "Epoch 27/400\n",
      "484/484 [==============================] - 0s 126us/sample - loss: 1.0406 - categorical_accuracy: 0.5868 - mean_squared_error: 0.0914 - val_loss: 0.6095 - val_categorical_accuracy: 0.8279 - val_mean_squared_error: 0.0550\n",
      "Epoch 28/400\n",
      "484/484 [==============================] - 0s 140us/sample - loss: 1.0775 - categorical_accuracy: 0.5723 - mean_squared_error: 0.0959 - val_loss: 0.5954 - val_categorical_accuracy: 0.8197 - val_mean_squared_error: 0.0535\n",
      "Epoch 29/400\n",
      "484/484 [==============================] - 0s 127us/sample - loss: 0.9855 - categorical_accuracy: 0.5909 - mean_squared_error: 0.0875 - val_loss: 0.5871 - val_categorical_accuracy: 0.8197 - val_mean_squared_error: 0.0528\n",
      "Epoch 30/400\n",
      "484/484 [==============================] - 0s 136us/sample - loss: 1.0354 - categorical_accuracy: 0.5950 - mean_squared_error: 0.0921 - val_loss: 0.5758 - val_categorical_accuracy: 0.8197 - val_mean_squared_error: 0.0519\n",
      "Epoch 31/400\n",
      "484/484 [==============================] - 0s 132us/sample - loss: 1.0199 - categorical_accuracy: 0.5579 - mean_squared_error: 0.0931 - val_loss: 0.5693 - val_categorical_accuracy: 0.8279 - val_mean_squared_error: 0.0516\n",
      "Epoch 32/400\n",
      "484/484 [==============================] - 0s 128us/sample - loss: 0.9305 - categorical_accuracy: 0.6033 - mean_squared_error: 0.0854 - val_loss: 0.5635 - val_categorical_accuracy: 0.8279 - val_mean_squared_error: 0.0512\n",
      "Epoch 33/400\n",
      "484/484 [==============================] - 0s 129us/sample - loss: 0.9344 - categorical_accuracy: 0.6219 - mean_squared_error: 0.0859 - val_loss: 0.5601 - val_categorical_accuracy: 0.8033 - val_mean_squared_error: 0.0513\n",
      "Epoch 34/400\n",
      "484/484 [==============================] - 0s 135us/sample - loss: 0.9846 - categorical_accuracy: 0.5888 - mean_squared_error: 0.0905 - val_loss: 0.5471 - val_categorical_accuracy: 0.8197 - val_mean_squared_error: 0.0500\n",
      "Epoch 35/400\n",
      "484/484 [==============================] - 0s 130us/sample - loss: 0.9675 - categorical_accuracy: 0.5992 - mean_squared_error: 0.0902 - val_loss: 0.5372 - val_categorical_accuracy: 0.8443 - val_mean_squared_error: 0.0489\n",
      "Epoch 36/400\n",
      "484/484 [==============================] - 0s 133us/sample - loss: 1.0049 - categorical_accuracy: 0.6054 - mean_squared_error: 0.0909 - val_loss: 0.5298 - val_categorical_accuracy: 0.8279 - val_mean_squared_error: 0.0482\n",
      "Epoch 37/400\n",
      "484/484 [==============================] - 0s 140us/sample - loss: 0.9530 - categorical_accuracy: 0.6198 - mean_squared_error: 0.0857 - val_loss: 0.5258 - val_categorical_accuracy: 0.8279 - val_mean_squared_error: 0.0478\n",
      "Epoch 38/400\n",
      "484/484 [==============================] - 0s 130us/sample - loss: 0.9434 - categorical_accuracy: 0.6219 - mean_squared_error: 0.0856 - val_loss: 0.5217 - val_categorical_accuracy: 0.8279 - val_mean_squared_error: 0.0475\n",
      "Epoch 39/400\n",
      "484/484 [==============================] - 0s 133us/sample - loss: 0.9291 - categorical_accuracy: 0.6426 - mean_squared_error: 0.0825 - val_loss: 0.5180 - val_categorical_accuracy: 0.8443 - val_mean_squared_error: 0.0472\n",
      "Epoch 40/400\n",
      "484/484 [==============================] - 0s 139us/sample - loss: 0.8898 - categorical_accuracy: 0.6302 - mean_squared_error: 0.0824 - val_loss: 0.5136 - val_categorical_accuracy: 0.8525 - val_mean_squared_error: 0.0467\n",
      "Epoch 41/400\n",
      "484/484 [==============================] - 0s 120us/sample - loss: 0.8667 - categorical_accuracy: 0.6281 - mean_squared_error: 0.0807 - val_loss: 0.5078 - val_categorical_accuracy: 0.8279 - val_mean_squared_error: 0.0463\n",
      "Epoch 42/400\n",
      "484/484 [==============================] - 0s 114us/sample - loss: 0.8908 - categorical_accuracy: 0.6219 - mean_squared_error: 0.0814 - val_loss: 0.5076 - val_categorical_accuracy: 0.8443 - val_mean_squared_error: 0.0464\n",
      "Epoch 43/400\n",
      "484/484 [==============================] - 0s 126us/sample - loss: 0.8774 - categorical_accuracy: 0.6281 - mean_squared_error: 0.0814 - val_loss: 0.4993 - val_categorical_accuracy: 0.8361 - val_mean_squared_error: 0.0454\n",
      "Epoch 44/400\n",
      "484/484 [==============================] - 0s 128us/sample - loss: 0.8751 - categorical_accuracy: 0.6674 - mean_squared_error: 0.0792 - val_loss: 0.4954 - val_categorical_accuracy: 0.8361 - val_mean_squared_error: 0.0452\n",
      "Epoch 45/400\n",
      "484/484 [==============================] - 0s 130us/sample - loss: 0.8594 - categorical_accuracy: 0.6426 - mean_squared_error: 0.0794 - val_loss: 0.4948 - val_categorical_accuracy: 0.8279 - val_mean_squared_error: 0.0453\n",
      "Epoch 46/400\n",
      "484/484 [==============================] - 0s 139us/sample - loss: 0.8685 - categorical_accuracy: 0.6488 - mean_squared_error: 0.0819 - val_loss: 0.4960 - val_categorical_accuracy: 0.8197 - val_mean_squared_error: 0.0457\n",
      "Epoch 47/400\n",
      "484/484 [==============================] - 0s 132us/sample - loss: 0.8353 - categorical_accuracy: 0.6240 - mean_squared_error: 0.0798 - val_loss: 0.4930 - val_categorical_accuracy: 0.8279 - val_mean_squared_error: 0.0456\n",
      "Epoch 48/400\n",
      "484/484 [==============================] - 0s 120us/sample - loss: 0.8067 - categorical_accuracy: 0.6653 - mean_squared_error: 0.0741 - val_loss: 0.4842 - val_categorical_accuracy: 0.8361 - val_mean_squared_error: 0.0446\n",
      "Epoch 49/400\n",
      "484/484 [==============================] - 0s 121us/sample - loss: 0.8445 - categorical_accuracy: 0.6260 - mean_squared_error: 0.0814 - val_loss: 0.4821 - val_categorical_accuracy: 0.8279 - val_mean_squared_error: 0.0445\n",
      "Epoch 50/400\n",
      "484/484 [==============================] - 0s 116us/sample - loss: 0.8636 - categorical_accuracy: 0.6529 - mean_squared_error: 0.0795 - val_loss: 0.4826 - val_categorical_accuracy: 0.8197 - val_mean_squared_error: 0.0447\n",
      "Epoch 51/400\n",
      "484/484 [==============================] - 0s 119us/sample - loss: 0.7847 - categorical_accuracy: 0.6818 - mean_squared_error: 0.0708 - val_loss: 0.4761 - val_categorical_accuracy: 0.8279 - val_mean_squared_error: 0.0439\n",
      "Epoch 52/400\n",
      "484/484 [==============================] - 0s 132us/sample - loss: 0.7518 - categorical_accuracy: 0.7004 - mean_squared_error: 0.0686 - val_loss: 0.4729 - val_categorical_accuracy: 0.8361 - val_mean_squared_error: 0.0434\n",
      "Epoch 53/400\n",
      "484/484 [==============================] - 0s 121us/sample - loss: 0.7970 - categorical_accuracy: 0.6591 - mean_squared_error: 0.0747 - val_loss: 0.4698 - val_categorical_accuracy: 0.8197 - val_mean_squared_error: 0.0431\n",
      "Epoch 54/400\n",
      "484/484 [==============================] - 0s 112us/sample - loss: 0.7652 - categorical_accuracy: 0.6880 - mean_squared_error: 0.0695 - val_loss: 0.4694 - val_categorical_accuracy: 0.8197 - val_mean_squared_error: 0.0432\n",
      "Epoch 55/400\n",
      "484/484 [==============================] - 0s 128us/sample - loss: 0.8035 - categorical_accuracy: 0.6405 - mean_squared_error: 0.0769 - val_loss: 0.4666 - val_categorical_accuracy: 0.8361 - val_mean_squared_error: 0.0429\n",
      "Epoch 56/400\n",
      "484/484 [==============================] - 0s 128us/sample - loss: 0.7728 - categorical_accuracy: 0.6860 - mean_squared_error: 0.0706 - val_loss: 0.4640 - val_categorical_accuracy: 0.8361 - val_mean_squared_error: 0.0425\n",
      "Epoch 57/400\n",
      "484/484 [==============================] - 0s 133us/sample - loss: 0.8035 - categorical_accuracy: 0.6818 - mean_squared_error: 0.0739 - val_loss: 0.4616 - val_categorical_accuracy: 0.8361 - val_mean_squared_error: 0.0423\n",
      "Epoch 58/400\n",
      "484/484 [==============================] - 0s 125us/sample - loss: 0.8066 - categorical_accuracy: 0.6591 - mean_squared_error: 0.0764 - val_loss: 0.4601 - val_categorical_accuracy: 0.8443 - val_mean_squared_error: 0.0422\n",
      "Epoch 59/400\n",
      "484/484 [==============================] - 0s 129us/sample - loss: 0.8279 - categorical_accuracy: 0.6529 - mean_squared_error: 0.0781 - val_loss: 0.4583 - val_categorical_accuracy: 0.8607 - val_mean_squared_error: 0.0420\n",
      "Epoch 60/400\n",
      "484/484 [==============================] - 0s 128us/sample - loss: 0.7336 - categorical_accuracy: 0.7004 - mean_squared_error: 0.0684 - val_loss: 0.4579 - val_categorical_accuracy: 0.8525 - val_mean_squared_error: 0.0421\n",
      "Epoch 61/400\n",
      "484/484 [==============================] - 0s 129us/sample - loss: 0.8263 - categorical_accuracy: 0.6488 - mean_squared_error: 0.0776 - val_loss: 0.4585 - val_categorical_accuracy: 0.8525 - val_mean_squared_error: 0.0421\n",
      "Epoch 62/400\n",
      "484/484 [==============================] - 0s 123us/sample - loss: 0.8062 - categorical_accuracy: 0.6942 - mean_squared_error: 0.0730 - val_loss: 0.4572 - val_categorical_accuracy: 0.8361 - val_mean_squared_error: 0.0419\n",
      "Epoch 63/400\n",
      "484/484 [==============================] - 0s 118us/sample - loss: 0.7165 - categorical_accuracy: 0.7025 - mean_squared_error: 0.0679 - val_loss: 0.4516 - val_categorical_accuracy: 0.8525 - val_mean_squared_error: 0.0413\n",
      "Epoch 64/400\n",
      "484/484 [==============================] - 0s 125us/sample - loss: 0.7773 - categorical_accuracy: 0.6777 - mean_squared_error: 0.0742 - val_loss: 0.4529 - val_categorical_accuracy: 0.8443 - val_mean_squared_error: 0.0417\n",
      "Epoch 65/400\n",
      "484/484 [==============================] - 0s 137us/sample - loss: 0.8035 - categorical_accuracy: 0.6715 - mean_squared_error: 0.0757 - val_loss: 0.4503 - val_categorical_accuracy: 0.8443 - val_mean_squared_error: 0.0414\n",
      "Epoch 66/400\n",
      "484/484 [==============================] - 0s 120us/sample - loss: 0.7549 - categorical_accuracy: 0.6777 - mean_squared_error: 0.0715 - val_loss: 0.4517 - val_categorical_accuracy: 0.8443 - val_mean_squared_error: 0.0416\n",
      "Epoch 67/400\n",
      "484/484 [==============================] - 0s 134us/sample - loss: 0.7786 - categorical_accuracy: 0.7066 - mean_squared_error: 0.0700 - val_loss: 0.4492 - val_categorical_accuracy: 0.8361 - val_mean_squared_error: 0.0413\n",
      "Epoch 68/400\n",
      "484/484 [==============================] - 0s 129us/sample - loss: 0.7444 - categorical_accuracy: 0.6818 - mean_squared_error: 0.0711 - val_loss: 0.4463 - val_categorical_accuracy: 0.8279 - val_mean_squared_error: 0.0410\n",
      "Epoch 69/400\n",
      "484/484 [==============================] - 0s 137us/sample - loss: 0.7125 - categorical_accuracy: 0.7335 - mean_squared_error: 0.0654 - val_loss: 0.4432 - val_categorical_accuracy: 0.8361 - val_mean_squared_error: 0.0405\n",
      "Epoch 70/400\n",
      "484/484 [==============================] - 0s 121us/sample - loss: 0.7707 - categorical_accuracy: 0.6798 - mean_squared_error: 0.0741 - val_loss: 0.4428 - val_categorical_accuracy: 0.8279 - val_mean_squared_error: 0.0406\n",
      "Epoch 71/400\n",
      "484/484 [==============================] - 0s 131us/sample - loss: 0.7407 - categorical_accuracy: 0.7211 - mean_squared_error: 0.0671 - val_loss: 0.4444 - val_categorical_accuracy: 0.8279 - val_mean_squared_error: 0.0408\n",
      "Epoch 72/400\n",
      "484/484 [==============================] - 0s 134us/sample - loss: 0.7688 - categorical_accuracy: 0.6736 - mean_squared_error: 0.0712 - val_loss: 0.4431 - val_categorical_accuracy: 0.8361 - val_mean_squared_error: 0.0407\n",
      "Epoch 73/400\n",
      "484/484 [==============================] - 0s 142us/sample - loss: 0.7420 - categorical_accuracy: 0.7045 - mean_squared_error: 0.0697 - val_loss: 0.4385 - val_categorical_accuracy: 0.8279 - val_mean_squared_error: 0.0402\n",
      "Epoch 74/400\n",
      "484/484 [==============================] - 0s 128us/sample - loss: 0.7030 - categorical_accuracy: 0.7417 - mean_squared_error: 0.0637 - val_loss: 0.4323 - val_categorical_accuracy: 0.8361 - val_mean_squared_error: 0.0395\n",
      "Epoch 75/400\n",
      "484/484 [==============================] - 0s 124us/sample - loss: 0.7015 - categorical_accuracy: 0.7149 - mean_squared_error: 0.0655 - val_loss: 0.4300 - val_categorical_accuracy: 0.8525 - val_mean_squared_error: 0.0393\n",
      "Epoch 76/400\n",
      "484/484 [==============================] - 0s 126us/sample - loss: 0.7174 - categorical_accuracy: 0.7066 - mean_squared_error: 0.0677 - val_loss: 0.4310 - val_categorical_accuracy: 0.8607 - val_mean_squared_error: 0.0394\n",
      "Epoch 77/400\n",
      "484/484 [==============================] - 0s 133us/sample - loss: 0.7325 - categorical_accuracy: 0.7004 - mean_squared_error: 0.0696 - val_loss: 0.4336 - val_categorical_accuracy: 0.8607 - val_mean_squared_error: 0.0399\n",
      "Epoch 78/400\n",
      "484/484 [==============================] - 0s 124us/sample - loss: 0.7576 - categorical_accuracy: 0.7107 - mean_squared_error: 0.0679 - val_loss: 0.4338 - val_categorical_accuracy: 0.8443 - val_mean_squared_error: 0.0397\n",
      "Epoch 79/400\n",
      "484/484 [==============================] - 0s 130us/sample - loss: 0.7401 - categorical_accuracy: 0.7066 - mean_squared_error: 0.0684 - val_loss: 0.4342 - val_categorical_accuracy: 0.8443 - val_mean_squared_error: 0.0398\n",
      "Epoch 80/400\n",
      "484/484 [==============================] - 0s 131us/sample - loss: 0.7302 - categorical_accuracy: 0.6798 - mean_squared_error: 0.0701 - val_loss: 0.4370 - val_categorical_accuracy: 0.8607 - val_mean_squared_error: 0.0401\n",
      "Epoch 81/400\n",
      "484/484 [==============================] - 0s 133us/sample - loss: 0.7366 - categorical_accuracy: 0.7087 - mean_squared_error: 0.0682 - val_loss: 0.4328 - val_categorical_accuracy: 0.8607 - val_mean_squared_error: 0.0396\n",
      "Epoch 82/400\n",
      "484/484 [==============================] - 0s 131us/sample - loss: 0.7417 - categorical_accuracy: 0.7149 - mean_squared_error: 0.0682 - val_loss: 0.4327 - val_categorical_accuracy: 0.8443 - val_mean_squared_error: 0.0396\n",
      "Epoch 83/400\n",
      "484/484 [==============================] - 0s 134us/sample - loss: 0.7108 - categorical_accuracy: 0.7087 - mean_squared_error: 0.0688 - val_loss: 0.4336 - val_categorical_accuracy: 0.8361 - val_mean_squared_error: 0.0397\n",
      "Epoch 84/400\n",
      "484/484 [==============================] - 0s 121us/sample - loss: 0.7064 - categorical_accuracy: 0.7087 - mean_squared_error: 0.0672 - val_loss: 0.4313 - val_categorical_accuracy: 0.8443 - val_mean_squared_error: 0.0394\n",
      "Epoch 85/400\n",
      "484/484 [==============================] - 0s 128us/sample - loss: 0.7178 - categorical_accuracy: 0.7025 - mean_squared_error: 0.0665 - val_loss: 0.4307 - val_categorical_accuracy: 0.8525 - val_mean_squared_error: 0.0394\n",
      "Epoch 86/400\n",
      "484/484 [==============================] - 0s 131us/sample - loss: 0.7319 - categorical_accuracy: 0.6983 - mean_squared_error: 0.0681 - val_loss: 0.4274 - val_categorical_accuracy: 0.8525 - val_mean_squared_error: 0.0391\n",
      "Epoch 87/400\n",
      "484/484 [==============================] - 0s 132us/sample - loss: 0.6588 - categorical_accuracy: 0.7355 - mean_squared_error: 0.0630 - val_loss: 0.4271 - val_categorical_accuracy: 0.8361 - val_mean_squared_error: 0.0393\n",
      "Epoch 88/400\n",
      "484/484 [==============================] - 0s 117us/sample - loss: 0.7766 - categorical_accuracy: 0.6901 - mean_squared_error: 0.0711 - val_loss: 0.4288 - val_categorical_accuracy: 0.8443 - val_mean_squared_error: 0.0396\n",
      "Epoch 89/400\n",
      "484/484 [==============================] - 0s 118us/sample - loss: 0.7300 - categorical_accuracy: 0.7314 - mean_squared_error: 0.0651 - val_loss: 0.4285 - val_categorical_accuracy: 0.8525 - val_mean_squared_error: 0.0394\n",
      "Epoch 90/400\n",
      "484/484 [==============================] - 0s 121us/sample - loss: 0.6706 - categorical_accuracy: 0.6942 - mean_squared_error: 0.0650 - val_loss: 0.4332 - val_categorical_accuracy: 0.8443 - val_mean_squared_error: 0.0399\n",
      "Epoch 91/400\n",
      "484/484 [==============================] - 0s 132us/sample - loss: 0.7143 - categorical_accuracy: 0.7169 - mean_squared_error: 0.0663 - val_loss: 0.4348 - val_categorical_accuracy: 0.8443 - val_mean_squared_error: 0.0402\n",
      "Epoch 92/400\n",
      "484/484 [==============================] - 0s 114us/sample - loss: 0.7278 - categorical_accuracy: 0.7190 - mean_squared_error: 0.0665 - val_loss: 0.4376 - val_categorical_accuracy: 0.8443 - val_mean_squared_error: 0.0404\n",
      "Epoch 93/400\n",
      "484/484 [==============================] - 0s 116us/sample - loss: 0.6524 - categorical_accuracy: 0.7252 - mean_squared_error: 0.0626 - val_loss: 0.4362 - val_categorical_accuracy: 0.8361 - val_mean_squared_error: 0.0401\n",
      "Epoch 94/400\n",
      "484/484 [==============================] - 0s 126us/sample - loss: 0.7139 - categorical_accuracy: 0.7149 - mean_squared_error: 0.0665 - val_loss: 0.4345 - val_categorical_accuracy: 0.8279 - val_mean_squared_error: 0.0401\n",
      "Epoch 95/400\n",
      "484/484 [==============================] - 0s 133us/sample - loss: 0.6908 - categorical_accuracy: 0.7211 - mean_squared_error: 0.0635 - val_loss: 0.4333 - val_categorical_accuracy: 0.8361 - val_mean_squared_error: 0.0399\n",
      "Epoch 96/400\n",
      "484/484 [==============================] - 0s 128us/sample - loss: 0.7061 - categorical_accuracy: 0.7231 - mean_squared_error: 0.0658 - val_loss: 0.4327 - val_categorical_accuracy: 0.8443 - val_mean_squared_error: 0.0396\n",
      "Epoch 97/400\n",
      "484/484 [==============================] - 0s 114us/sample - loss: 0.7037 - categorical_accuracy: 0.7045 - mean_squared_error: 0.0670 - val_loss: 0.4304 - val_categorical_accuracy: 0.8443 - val_mean_squared_error: 0.0393\n",
      "Epoch 98/400\n",
      "484/484 [==============================] - 0s 114us/sample - loss: 0.6590 - categorical_accuracy: 0.7293 - mean_squared_error: 0.0638 - val_loss: 0.4312 - val_categorical_accuracy: 0.8525 - val_mean_squared_error: 0.0392\n",
      "Epoch 99/400\n",
      "484/484 [==============================] - 0s 115us/sample - loss: 0.6509 - categorical_accuracy: 0.7190 - mean_squared_error: 0.0620 - val_loss: 0.4332 - val_categorical_accuracy: 0.8525 - val_mean_squared_error: 0.0394\n",
      "Epoch 100/400\n",
      "484/484 [==============================] - 0s 115us/sample - loss: 0.6368 - categorical_accuracy: 0.7397 - mean_squared_error: 0.0611 - val_loss: 0.4296 - val_categorical_accuracy: 0.8525 - val_mean_squared_error: 0.0391\n",
      "Epoch 101/400\n",
      "484/484 [==============================] - 0s 208us/sample - loss: 0.6449 - categorical_accuracy: 0.7397 - mean_squared_error: 0.0603 - val_loss: 0.4282 - val_categorical_accuracy: 0.8443 - val_mean_squared_error: 0.0390\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(X_train, y_train, validation_data=(X_test, y_test), epochs=400, batch_size=20, callbacks=[tensorboard_callback, early_stop])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "pharmaceutical-edward",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "herbal-parker",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Minimum Validation Loss: 0.4271\n"
     ]
    },
    {
     "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": [
    "history_df = pd.DataFrame(history.history)\n",
    "history_df.loc[0:, ['loss', 'val_loss']].plot()\n",
    "print(\"Minimum Validation Loss: {:0.4f}\".format(history_df['val_loss'].min()));"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "annual-laundry",
   "metadata": {},
   "outputs": [
    {
     "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": [
    "from matplotlib import pyplot as plt\n",
    "plt.plot(history.history['mean_squared_error'])\n",
    "plt.plot(history.history['val_mean_squared_error'])\n",
    "plt.title('MSE')\n",
    "plt.ylabel('accuracy')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'val'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "fixed-strategy",
   "metadata": {},
   "outputs": [
    {
     "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": [
    "from matplotlib import pyplot as plt\n",
    "plt.plot(history.history['categorical_accuracy'])\n",
    "plt.plot(history.history['val_categorical_accuracy'])\n",
    "plt.title('model accuracy')\n",
    "plt.ylabel('accuracy')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'val'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "indie-absorption",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "russian-onion",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'loss': [2.2912292332688637,\n",
       "  2.1404912200841038,\n",
       "  2.0943352210620216,\n",
       "  1.8976128830397425,\n",
       "  1.7973781971892049,\n",
       "  1.688168748351168,\n",
       "  1.680202636836974,\n",
       "  1.5960642789021011,\n",
       "  1.5428474880447072,\n",
       "  1.5156352510137006,\n",
       "  1.522028888552642,\n",
       "  1.3618978656027927,\n",
       "  1.315145037390969,\n",
       "  1.337025570967966,\n",
       "  1.372529037235197,\n",
       "  1.3646193466895868,\n",
       "  1.3112599583696727,\n",
       "  1.2639093921204243,\n",
       "  1.2203325412490151,\n",
       "  1.2040263572015053,\n",
       "  1.1880327024735695,\n",
       "  1.2130317338242018,\n",
       "  1.1092497061106785,\n",
       "  1.0932107168780871,\n",
       "  1.0427935135758613,\n",
       "  1.1094993540078155,\n",
       "  1.0405948403453038,\n",
       "  1.0775153489152263,\n",
       "  0.9854581543236725,\n",
       "  1.0354462669900626,\n",
       "  1.0199157945380724,\n",
       "  0.9304981822809897,\n",
       "  0.9344309647221211,\n",
       "  0.9845870408144864,\n",
       "  0.9675349455234433,\n",
       "  1.004885857016587,\n",
       "  0.9529582725083532,\n",
       "  0.9434027322067702,\n",
       "  0.9290995627395378,\n",
       "  0.889772822664789,\n",
       "  0.8667356514733685,\n",
       "  0.8908105128067584,\n",
       "  0.8773999366878478,\n",
       "  0.8750589076152517,\n",
       "  0.8593963847672643,\n",
       "  0.8685320600990422,\n",
       "  0.8353416764539136,\n",
       "  0.8067082186375768,\n",
       "  0.8444791130783144,\n",
       "  0.8636316297468075,\n",
       "  0.784714921200571,\n",
       "  0.7517838415281832,\n",
       "  0.7969798206790419,\n",
       "  0.7652499030444248,\n",
       "  0.8034744494217486,\n",
       "  0.7727868552050314,\n",
       "  0.8035402445753744,\n",
       "  0.8066107492801572,\n",
       "  0.8278805019934316,\n",
       "  0.7336392151422737,\n",
       "  0.8263470020668566,\n",
       "  0.8062209871189654,\n",
       "  0.7164808009281631,\n",
       "  0.7773278294023404,\n",
       "  0.8035495446240606,\n",
       "  0.7548963567934746,\n",
       "  0.778572465516319,\n",
       "  0.744383293242494,\n",
       "  0.712542909728594,\n",
       "  0.7706907790554457,\n",
       "  0.7407111500905565,\n",
       "  0.7688398198647932,\n",
       "  0.741965213467267,\n",
       "  0.7030449504694662,\n",
       "  0.7014851148956078,\n",
       "  0.7174429669360484,\n",
       "  0.732535030231003,\n",
       "  0.75760052100686,\n",
       "  0.7400885842555811,\n",
       "  0.7301500769686108,\n",
       "  0.7365588486687211,\n",
       "  0.7416968579627266,\n",
       "  0.7107971687454823,\n",
       "  0.7064130449098004,\n",
       "  0.7178080884878301,\n",
       "  0.7318545295433565,\n",
       "  0.658845224163749,\n",
       "  0.7765784593653088,\n",
       "  0.7299910278852321,\n",
       "  0.6705740879882466,\n",
       "  0.7143486422940719,\n",
       "  0.7277510892753759,\n",
       "  0.6524399755414853,\n",
       "  0.7138590694459017,\n",
       "  0.6908445651373587,\n",
       "  0.7061296538873152,\n",
       "  0.7036798318062932,\n",
       "  0.658988779980289,\n",
       "  0.6508902047783875,\n",
       "  0.6368496408147261,\n",
       "  0.6448736373058035],\n",
       " 'categorical_accuracy': [0.18595041,\n",
       "  0.2107438,\n",
       "  0.27066115,\n",
       "  0.32438016,\n",
       "  0.30991736,\n",
       "  0.3553719,\n",
       "  0.39256197,\n",
       "  0.36570248,\n",
       "  0.42561984,\n",
       "  0.4607438,\n",
       "  0.42975205,\n",
       "  0.46900827,\n",
       "  0.48347107,\n",
       "  0.4896694,\n",
       "  0.50206614,\n",
       "  0.48347107,\n",
       "  0.47933885,\n",
       "  0.4876033,\n",
       "  0.5371901,\n",
       "  0.553719,\n",
       "  0.553719,\n",
       "  0.5289256,\n",
       "  0.5495868,\n",
       "  0.58264464,\n",
       "  0.58264464,\n",
       "  0.57438016,\n",
       "  0.58677685,\n",
       "  0.572314,\n",
       "  0.59090906,\n",
       "  0.59504133,\n",
       "  0.55785125,\n",
       "  0.60330576,\n",
       "  0.6219008,\n",
       "  0.588843,\n",
       "  0.59917355,\n",
       "  0.6053719,\n",
       "  0.6198347,\n",
       "  0.6219008,\n",
       "  0.642562,\n",
       "  0.6301653,\n",
       "  0.6280992,\n",
       "  0.6219008,\n",
       "  0.6280992,\n",
       "  0.66735536,\n",
       "  0.642562,\n",
       "  0.6487603,\n",
       "  0.62396693,\n",
       "  0.6652893,\n",
       "  0.62603307,\n",
       "  0.6528926,\n",
       "  0.6818182,\n",
       "  0.7004132,\n",
       "  0.65909094,\n",
       "  0.68801653,\n",
       "  0.6404959,\n",
       "  0.6859504,\n",
       "  0.6818182,\n",
       "  0.65909094,\n",
       "  0.6528926,\n",
       "  0.7004132,\n",
       "  0.6487603,\n",
       "  0.6942149,\n",
       "  0.70247936,\n",
       "  0.677686,\n",
       "  0.6714876,\n",
       "  0.677686,\n",
       "  0.7066116,\n",
       "  0.6818182,\n",
       "  0.7334711,\n",
       "  0.67975205,\n",
       "  0.7210744,\n",
       "  0.6735537,\n",
       "  0.70454544,\n",
       "  0.7417355,\n",
       "  0.71487606,\n",
       "  0.7066116,\n",
       "  0.7004132,\n",
       "  0.7107438,\n",
       "  0.7066116,\n",
       "  0.67975205,\n",
       "  0.7086777,\n",
       "  0.71487606,\n",
       "  0.7086777,\n",
       "  0.7086777,\n",
       "  0.70247936,\n",
       "  0.6983471,\n",
       "  0.7355372,\n",
       "  0.69008267,\n",
       "  0.73140496,\n",
       "  0.6942149,\n",
       "  0.71694213,\n",
       "  0.71900827,\n",
       "  0.7252066,\n",
       "  0.71487606,\n",
       "  0.7210744,\n",
       "  0.7231405,\n",
       "  0.70454544,\n",
       "  0.7293388,\n",
       "  0.71900827,\n",
       "  0.73966944,\n",
       "  0.73966944],\n",
       " 'mean_squared_error': [0.16573393,\n",
       "  0.1585617,\n",
       "  0.15574253,\n",
       "  0.14310147,\n",
       "  0.14188309,\n",
       "  0.13511342,\n",
       "  0.13247544,\n",
       "  0.13131827,\n",
       "  0.12182754,\n",
       "  0.121013016,\n",
       "  0.12524039,\n",
       "  0.11389687,\n",
       "  0.11101796,\n",
       "  0.11227931,\n",
       "  0.11387067,\n",
       "  0.11292897,\n",
       "  0.114936575,\n",
       "  0.111112185,\n",
       "  0.103300855,\n",
       "  0.10310099,\n",
       "  0.10028489,\n",
       "  0.10477016,\n",
       "  0.09686814,\n",
       "  0.096755184,\n",
       "  0.09316218,\n",
       "  0.09720273,\n",
       "  0.091431044,\n",
       "  0.09587847,\n",
       "  0.08754962,\n",
       "  0.09214955,\n",
       "  0.09313909,\n",
       "  0.08537201,\n",
       "  0.08588627,\n",
       "  0.090479985,\n",
       "  0.09022616,\n",
       "  0.09094847,\n",
       "  0.08570308,\n",
       "  0.085568264,\n",
       "  0.08252639,\n",
       "  0.08242971,\n",
       "  0.080658995,\n",
       "  0.08144919,\n",
       "  0.08143684,\n",
       "  0.07915113,\n",
       "  0.079374395,\n",
       "  0.08186226,\n",
       "  0.07977787,\n",
       "  0.074062094,\n",
       "  0.081354335,\n",
       "  0.07950503,\n",
       "  0.07083749,\n",
       "  0.068614565,\n",
       "  0.07465135,\n",
       "  0.06950826,\n",
       "  0.076856315,\n",
       "  0.07057605,\n",
       "  0.073875174,\n",
       "  0.07635471,\n",
       "  0.07809374,\n",
       "  0.06836563,\n",
       "  0.07757598,\n",
       "  0.072959624,\n",
       "  0.067920014,\n",
       "  0.07421881,\n",
       "  0.07572001,\n",
       "  0.07150515,\n",
       "  0.06996168,\n",
       "  0.071086414,\n",
       "  0.065369695,\n",
       "  0.07408146,\n",
       "  0.06705909,\n",
       "  0.07120595,\n",
       "  0.069746315,\n",
       "  0.0637376,\n",
       "  0.06552189,\n",
       "  0.06771755,\n",
       "  0.06955641,\n",
       "  0.06792294,\n",
       "  0.06844753,\n",
       "  0.070130944,\n",
       "  0.06821532,\n",
       "  0.0682039,\n",
       "  0.06877607,\n",
       "  0.06715684,\n",
       "  0.066484235,\n",
       "  0.06809059,\n",
       "  0.06297304,\n",
       "  0.07112283,\n",
       "  0.06506878,\n",
       "  0.0650384,\n",
       "  0.0663266,\n",
       "  0.06647373,\n",
       "  0.062584095,\n",
       "  0.06654201,\n",
       "  0.06346958,\n",
       "  0.06577913,\n",
       "  0.06702162,\n",
       "  0.06378819,\n",
       "  0.062007505,\n",
       "  0.061095167,\n",
       "  0.060328502],\n",
       " 'val_loss': [1.7793375234134863,\n",
       "  1.7270356553499815,\n",
       "  1.6807994803444284,\n",
       "  1.6346779455904101,\n",
       "  1.5764793900192762,\n",
       "  1.5090078330430827,\n",
       "  1.4334860786062773,\n",
       "  1.3438869968789522,\n",
       "  1.2642243944230627,\n",
       "  1.1834128196122216,\n",
       "  1.1028089787139268,\n",
       "  1.0297915153816097,\n",
       "  0.9682935886695737,\n",
       "  0.9149435094145478,\n",
       "  0.8609869656015615,\n",
       "  0.8208620069456882,\n",
       "  0.778310203161396,\n",
       "  0.7478046202268757,\n",
       "  0.7226879635795218,\n",
       "  0.7019719911403344,\n",
       "  0.6844228203179407,\n",
       "  0.6699070822997172,\n",
       "  0.6492072236342509,\n",
       "  0.6362706579145838,\n",
       "  0.6289952641627827,\n",
       "  0.6188178267635283,\n",
       "  0.6095113265709798,\n",
       "  0.595380142086842,\n",
       "  0.5870812877279813,\n",
       "  0.5757987030216905,\n",
       "  0.5693338606201235,\n",
       "  0.5634631778373093,\n",
       "  0.5600570430521106,\n",
       "  0.5471436132173069,\n",
       "  0.537188889550381,\n",
       "  0.5298209688702568,\n",
       "  0.5258149514432813,\n",
       "  0.5216607165629746,\n",
       "  0.51795993745327,\n",
       "  0.513568790232549,\n",
       "  0.5078206404310758,\n",
       "  0.5075500763830592,\n",
       "  0.4992670074838107,\n",
       "  0.4954342402395655,\n",
       "  0.49477336543505307,\n",
       "  0.4959562503900684,\n",
       "  0.4930002747989092,\n",
       "  0.4842207202168762,\n",
       "  0.48209226278008005,\n",
       "  0.4825645763365949,\n",
       "  0.47613558857167354,\n",
       "  0.47291642659511723,\n",
       "  0.46976649248209157,\n",
       "  0.4694408188100721,\n",
       "  0.46657401176749685,\n",
       "  0.46403480944086295,\n",
       "  0.4615796369607331,\n",
       "  0.46012695842101925,\n",
       "  0.4582645981038203,\n",
       "  0.45791915205658457,\n",
       "  0.45847029642003484,\n",
       "  0.4571509583563101,\n",
       "  0.4516142479708937,\n",
       "  0.4529005523587837,\n",
       "  0.45029240162646184,\n",
       "  0.45172870696568096,\n",
       "  0.44916489427206946,\n",
       "  0.4462607141400947,\n",
       "  0.4431607996831175,\n",
       "  0.44280123417494727,\n",
       "  0.4443628006294125,\n",
       "  0.4430585967712715,\n",
       "  0.4385431105973291,\n",
       "  0.4322839732297131,\n",
       "  0.4300091457171518,\n",
       "  0.4309737506459971,\n",
       "  0.43356394963186295,\n",
       "  0.4338467663428823,\n",
       "  0.4342154831182761,\n",
       "  0.4369679928803053,\n",
       "  0.43275695544530135,\n",
       "  0.4326986020705739,\n",
       "  0.43359278973008764,\n",
       "  0.4313278217784694,\n",
       "  0.43066832814060274,\n",
       "  0.4274347341451489,\n",
       "  0.42708602694214365,\n",
       "  0.4288202314103236,\n",
       "  0.428489408532127,\n",
       "  0.4332127952184833,\n",
       "  0.4347849791167212,\n",
       "  0.43761241582573435,\n",
       "  0.4362472129649803,\n",
       "  0.43447832983048235,\n",
       "  0.43331543490534924,\n",
       "  0.4326787483496744,\n",
       "  0.4304361871031464,\n",
       "  0.43118724910939327,\n",
       "  0.4331842758616463,\n",
       "  0.4295755883709329,\n",
       "  0.4282153737349588],\n",
       " 'val_categorical_accuracy': [0.16393442,\n",
       "  0.352459,\n",
       "  0.352459,\n",
       "  0.352459,\n",
       "  0.36885247,\n",
       "  0.40163934,\n",
       "  0.46721312,\n",
       "  0.5163934,\n",
       "  0.5327869,\n",
       "  0.6147541,\n",
       "  0.6557377,\n",
       "  0.6967213,\n",
       "  0.72131145,\n",
       "  0.73770493,\n",
       "  0.75409836,\n",
       "  0.76229507,\n",
       "  0.78688526,\n",
       "  0.78688526,\n",
       "  0.78688526,\n",
       "  0.78688526,\n",
       "  0.8032787,\n",
       "  0.8032787,\n",
       "  0.8114754,\n",
       "  0.8114754,\n",
       "  0.8278689,\n",
       "  0.8278689,\n",
       "  0.8278689,\n",
       "  0.8196721,\n",
       "  0.8196721,\n",
       "  0.8196721,\n",
       "  0.8278689,\n",
       "  0.8278689,\n",
       "  0.8032787,\n",
       "  0.8196721,\n",
       "  0.8442623,\n",
       "  0.8278689,\n",
       "  0.8278689,\n",
       "  0.8278689,\n",
       "  0.8442623,\n",
       "  0.852459,\n",
       "  0.8278689,\n",
       "  0.8442623,\n",
       "  0.8360656,\n",
       "  0.8360656,\n",
       "  0.8278689,\n",
       "  0.8196721,\n",
       "  0.8278689,\n",
       "  0.8360656,\n",
       "  0.8278689,\n",
       "  0.8196721,\n",
       "  0.8278689,\n",
       "  0.8360656,\n",
       "  0.8196721,\n",
       "  0.8196721,\n",
       "  0.8360656,\n",
       "  0.8360656,\n",
       "  0.8360656,\n",
       "  0.8442623,\n",
       "  0.8606557,\n",
       "  0.852459,\n",
       "  0.852459,\n",
       "  0.8360656,\n",
       "  0.852459,\n",
       "  0.8442623,\n",
       "  0.8442623,\n",
       "  0.8442623,\n",
       "  0.8360656,\n",
       "  0.8278689,\n",
       "  0.8360656,\n",
       "  0.8278689,\n",
       "  0.8278689,\n",
       "  0.8360656,\n",
       "  0.8278689,\n",
       "  0.8360656,\n",
       "  0.852459,\n",
       "  0.8606557,\n",
       "  0.8606557,\n",
       "  0.8442623,\n",
       "  0.8442623,\n",
       "  0.8606557,\n",
       "  0.8606557,\n",
       "  0.8442623,\n",
       "  0.8360656,\n",
       "  0.8442623,\n",
       "  0.852459,\n",
       "  0.852459,\n",
       "  0.8360656,\n",
       "  0.8442623,\n",
       "  0.852459,\n",
       "  0.8442623,\n",
       "  0.8442623,\n",
       "  0.8442623,\n",
       "  0.8360656,\n",
       "  0.8278689,\n",
       "  0.8360656,\n",
       "  0.8442623,\n",
       "  0.8442623,\n",
       "  0.852459,\n",
       "  0.852459,\n",
       "  0.852459,\n",
       "  0.8442623],\n",
       " 'val_mean_squared_error': [0.1383414,\n",
       "  0.13511777,\n",
       "  0.13243052,\n",
       "  0.12980737,\n",
       "  0.12645546,\n",
       "  0.12245093,\n",
       "  0.11787124,\n",
       "  0.111848995,\n",
       "  0.10630392,\n",
       "  0.10047931,\n",
       "  0.09426251,\n",
       "  0.08856426,\n",
       "  0.083946325,\n",
       "  0.079560176,\n",
       "  0.07506372,\n",
       "  0.071870595,\n",
       "  0.06808979,\n",
       "  0.065556064,\n",
       "  0.06368709,\n",
       "  0.062252473,\n",
       "  0.060630247,\n",
       "  0.059631947,\n",
       "  0.05770937,\n",
       "  0.056665376,\n",
       "  0.056305226,\n",
       "  0.055602957,\n",
       "  0.05495674,\n",
       "  0.05347139,\n",
       "  0.05280452,\n",
       "  0.051930174,\n",
       "  0.051554278,\n",
       "  0.051224045,\n",
       "  0.051333774,\n",
       "  0.050017677,\n",
       "  0.048874304,\n",
       "  0.048169993,\n",
       "  0.04783627,\n",
       "  0.047467805,\n",
       "  0.047194976,\n",
       "  0.04672078,\n",
       "  0.046290964,\n",
       "  0.046354022,\n",
       "  0.045411836,\n",
       "  0.045222532,\n",
       "  0.045326553,\n",
       "  0.04571528,\n",
       "  0.045632966,\n",
       "  0.044553578,\n",
       "  0.044512648,\n",
       "  0.044723384,\n",
       "  0.04386666,\n",
       "  0.0433847,\n",
       "  0.04309207,\n",
       "  0.04315035,\n",
       "  0.042863827,\n",
       "  0.042532098,\n",
       "  0.042254984,\n",
       "  0.042249292,\n",
       "  0.041993093,\n",
       "  0.042065322,\n",
       "  0.0421124,\n",
       "  0.041888893,\n",
       "  0.041253198,\n",
       "  0.04171152,\n",
       "  0.0413655,\n",
       "  0.041580353,\n",
       "  0.04134116,\n",
       "  0.040977772,\n",
       "  0.040528093,\n",
       "  0.040599767,\n",
       "  0.04080716,\n",
       "  0.04070114,\n",
       "  0.0401651,\n",
       "  0.03951787,\n",
       "  0.039338183,\n",
       "  0.039394714,\n",
       "  0.03985859,\n",
       "  0.03973978,\n",
       "  0.03979317,\n",
       "  0.040139604,\n",
       "  0.039562102,\n",
       "  0.039634403,\n",
       "  0.039667483,\n",
       "  0.039404277,\n",
       "  0.039398342,\n",
       "  0.03908727,\n",
       "  0.039277136,\n",
       "  0.039567202,\n",
       "  0.039356172,\n",
       "  0.03990245,\n",
       "  0.040204704,\n",
       "  0.040446766,\n",
       "  0.040103067,\n",
       "  0.04009316,\n",
       "  0.039949887,\n",
       "  0.03962318,\n",
       "  0.039264865,\n",
       "  0.03917838,\n",
       "  0.03939954,\n",
       "  0.03909254,\n",
       "  0.038999237]}"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "history.history"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "retired-delicious",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "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>classical</th>\n",
       "      <th>hip-hop</th>\n",
       "      <th>jazz</th>\n",
       "      <th>metal</th>\n",
       "      <th>reggae</th>\n",
       "      <th>rock</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>131782</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>53486</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10430</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7262</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48422</th>\n",
       "      <td>1</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>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6658</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12432</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35343</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58817</th>\n",
       "      <td>1</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>46868</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>122 rows × 6 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        classical  hip-hop  jazz  metal  reggae  rock\n",
       "131782          0        0     0      0       1     0\n",
       "53486           0        1     0      0       0     0\n",
       "10430           0        0     0      0       0     1\n",
       "7262            0        0     1      0       0     0\n",
       "48422           1        0     0      0       0     0\n",
       "...           ...      ...   ...    ...     ...   ...\n",
       "6658            0        0     1      0       0     0\n",
       "12432           0        0     0      1       0     0\n",
       "35343           0        1     0      0       0     0\n",
       "58817           1        0     0      0       0     0\n",
       "46868           0        0     0      0       1     0\n",
       "\n",
       "[122 rows x 6 columns]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "polished-amsterdam",
   "metadata": {},
   "outputs": [],
   "source": [
    "# np.round(model.predict(X_test), 3)\n",
    "pred = pd.DataFrame(model.predict(X_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "interracial-glenn",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0      4\n",
       "1      1\n",
       "2      5\n",
       "3      2\n",
       "4      0\n",
       "      ..\n",
       "117    4\n",
       "118    3\n",
       "119    1\n",
       "120    0\n",
       "121    4\n",
       "Length: 122, dtype: int64"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred.idxmax(axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "thorough-merit",
   "metadata": {},
   "outputs": [],
   "source": [
    "pred_labels = pd.Series([y.columns[x] for x in pred.idxmax(axis=1)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "martial-malpractice",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0         reggae\n",
       "1        hip-hop\n",
       "2           rock\n",
       "3           jazz\n",
       "4      classical\n",
       "         ...    \n",
       "117         jazz\n",
       "118        metal\n",
       "119      hip-hop\n",
       "120    classical\n",
       "121       reggae\n",
       "Length: 122, dtype: object"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_labels = y_test.idxmax(axis=1)\n",
    "y_labels = y_labels.reset_index(drop=True)\n",
    "y_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "id": "burning-output",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6           jazz\n",
       "9         reggae\n",
       "10          jazz\n",
       "13          jazz\n",
       "19          rock\n",
       "24     classical\n",
       "31        reggae\n",
       "33          jazz\n",
       "35         metal\n",
       "38          rock\n",
       "45          rock\n",
       "58       hip-hop\n",
       "62        reggae\n",
       "75          jazz\n",
       "86       hip-hop\n",
       "91         metal\n",
       "111         jazz\n",
       "117         jazz\n",
       "dtype: object"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_labels[y_labels != pred_labels]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "nearby-savings",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index(['classical', 'hip-hop', 'jazz', 'metal', 'reggae', 'rock'], dtype='object')\n"
     ]
    },
    {
     "data": {
      "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.010</td>\n",
       "      <td>0.018</td>\n",
       "      <td>0.249</td>\n",
       "      <td>0.020</td>\n",
       "      <td>0.072</td>\n",
       "      <td>0.630</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.002</td>\n",
       "      <td>0.020</td>\n",
       "      <td>0.229</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.319</td>\n",
       "      <td>0.430</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.058</td>\n",
       "      <td>0.004</td>\n",
       "      <td>0.379</td>\n",
       "      <td>0.056</td>\n",
       "      <td>0.019</td>\n",
       "      <td>0.485</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.110</td>\n",
       "      <td>0.185</td>\n",
       "      <td>0.110</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.595</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0.000</td>\n",
       "      <td>0.689</td>\n",
       "      <td>0.007</td>\n",
       "      <td>0.026</td>\n",
       "      <td>0.043</td>\n",
       "      <td>0.234</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>0.386</td>\n",
       "      <td>0.001</td>\n",
       "      <td>0.557</td>\n",
       "      <td>0.003</td>\n",
       "      <td>0.009</td>\n",
       "      <td>0.045</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>0.141</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.086</td>\n",
       "      <td>0.036</td>\n",
       "      <td>0.046</td>\n",
       "      <td>0.691</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>0.898</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.101</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.001</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>0.000</td>\n",
       "      <td>0.015</td>\n",
       "      <td>0.002</td>\n",
       "      <td>0.448</td>\n",
       "      <td>0.004</td>\n",
       "      <td>0.530</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>0.007</td>\n",
       "      <td>0.347</td>\n",
       "      <td>0.310</td>\n",
       "      <td>0.010</td>\n",
       "      <td>0.174</td>\n",
       "      <td>0.152</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45</th>\n",
       "      <td>0.001</td>\n",
       "      <td>0.013</td>\n",
       "      <td>0.001</td>\n",
       "      <td>0.807</td>\n",
       "      <td>0.004</td>\n",
       "      <td>0.174</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58</th>\n",
       "      <td>0.000</td>\n",
       "      <td>0.332</td>\n",
       "      <td>0.002</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.665</td>\n",
       "      <td>0.002</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>62</th>\n",
       "      <td>0.000</td>\n",
       "      <td>0.536</td>\n",
       "      <td>0.001</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.457</td>\n",
       "      <td>0.006</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75</th>\n",
       "      <td>0.694</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.302</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.004</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86</th>\n",
       "      <td>0.000</td>\n",
       "      <td>0.104</td>\n",
       "      <td>0.006</td>\n",
       "      <td>0.012</td>\n",
       "      <td>0.204</td>\n",
       "      <td>0.674</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>91</th>\n",
       "      <td>0.000</td>\n",
       "      <td>0.002</td>\n",
       "      <td>0.001</td>\n",
       "      <td>0.091</td>\n",
       "      <td>0.002</td>\n",
       "      <td>0.904</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>111</th>\n",
       "      <td>0.624</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.370</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.003</td>\n",
       "      <td>0.003</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>117</th>\n",
       "      <td>0.026</td>\n",
       "      <td>0.017</td>\n",
       "      <td>0.468</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.488</td>\n",
       "      <td>0.001</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         0      1      2      3      4      5\n",
       "6    0.010  0.018  0.249  0.020  0.072  0.630\n",
       "9    0.002  0.020  0.229  0.000  0.319  0.430\n",
       "10   0.058  0.004  0.379  0.056  0.019  0.485\n",
       "13   0.110  0.185  0.110  0.000  0.595  0.000\n",
       "19   0.000  0.689  0.007  0.026  0.043  0.234\n",
       "24   0.386  0.001  0.557  0.003  0.009  0.045\n",
       "31   0.141  0.000  0.086  0.036  0.046  0.691\n",
       "33   0.898  0.000  0.101  0.000  0.001  0.000\n",
       "35   0.000  0.015  0.002  0.448  0.004  0.530\n",
       "38   0.007  0.347  0.310  0.010  0.174  0.152\n",
       "45   0.001  0.013  0.001  0.807  0.004  0.174\n",
       "58   0.000  0.332  0.002  0.000  0.665  0.002\n",
       "62   0.000  0.536  0.001  0.000  0.457  0.006\n",
       "75   0.694  0.000  0.302  0.000  0.004  0.000\n",
       "86   0.000  0.104  0.006  0.012  0.204  0.674\n",
       "91   0.000  0.002  0.001  0.091  0.002  0.904\n",
       "111  0.624  0.000  0.370  0.000  0.003  0.003\n",
       "117  0.026  0.017  0.468  0.000  0.488  0.001"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(y.columns)\n",
    "np.round(pred.iloc[y_labels[y_labels != pred_labels].index, :], 3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "id": "featured-natural",
   "metadata": {},
   "outputs": [],
   "source": [
    "af_pred = model.predict(af_df[features_list])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "id": "smart-selection",
   "metadata": {},
   "outputs": [],
   "source": [
    "to_concat = pd.DataFrame(np.round(af_pred, 3), columns=y.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "id": "fourth-allocation",
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.concat([final_df, to_concat], axis=1).to_csv('net_preds.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "muslim-competition",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}