07.00-autoencolders.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# importando as bibliotecas necessárias\n",
    "\n",
    "import numpy as np                         # biblioteca para cálculos matemáticos\n",
    "import pandas as pd                        # biblioteca para manipulação de dados \n",
    "import torch                               # biblioteca para cálculos com GPUs\n",
    "import torch.nn as nn                      # biblioteca para criação de redes neurais \n",
    "import torch.nn.parallel                   # biblioteca para criação de redes neurais paralelas\n",
    "import torch.optim as optim                # biblioteca para otimização de redes neurais\n",
    "from torch.autograd import Variable        # biblioteca para criação de variáveis\n",
    "import torch.utils.data                    # biblioteca para criação de dados"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# carregando os dados\n",
    "\n",
    "# filmes\n",
    "movies = pd.read_csv('./data/movies.dat', sep='::', header=None, engine='python', encoding='latin-1')\n",
    "\n",
    "# usuários\n",
    "users = pd.read_csv('./data/users.dat', sep='::', header=None, engine='python', encoding='latin-1')\n",
    "\n",
    "# avaliações dos usuários\n",
    "ratings = pd.read_csv('./data/ratings.dat', sep='::', header=None, engine='python', encoding='latin-1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>Toy Story (1995)</td>\n",
       "      <td>Animation|Children's|Comedy</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>Jumanji (1995)</td>\n",
       "      <td>Adventure|Children's|Fantasy</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>Grumpier Old Men (1995)</td>\n",
       "      <td>Comedy|Romance</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>Waiting to Exhale (1995)</td>\n",
       "      <td>Comedy|Drama</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>Father of the Bride Part II (1995)</td>\n",
       "      <td>Comedy</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   0                                   1                             2\n",
       "0  1                    Toy Story (1995)   Animation|Children's|Comedy\n",
       "1  2                      Jumanji (1995)  Adventure|Children's|Fantasy\n",
       "2  3             Grumpier Old Men (1995)                Comedy|Romance\n",
       "3  4            Waiting to Exhale (1995)                  Comedy|Drama\n",
       "4  5  Father of the Bride Part II (1995)                        Comedy"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# visualizando os primeiros registros dos filmes\n",
    "\n",
    "movies.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>F</td>\n",
       "      <td>1</td>\n",
       "      <td>10</td>\n",
       "      <td>48067</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>M</td>\n",
       "      <td>56</td>\n",
       "      <td>16</td>\n",
       "      <td>70072</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>M</td>\n",
       "      <td>25</td>\n",
       "      <td>15</td>\n",
       "      <td>55117</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>M</td>\n",
       "      <td>45</td>\n",
       "      <td>7</td>\n",
       "      <td>02460</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>M</td>\n",
       "      <td>25</td>\n",
       "      <td>20</td>\n",
       "      <td>55455</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   0  1   2   3      4\n",
       "0  1  F   1  10  48067\n",
       "1  2  M  56  16  70072\n",
       "2  3  M  25  15  55117\n",
       "3  4  M  45   7  02460\n",
       "4  5  M  25  20  55455"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# visualizando os primeiros registros dos usuários\n",
    "\n",
    "users.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>1193</td>\n",
       "      <td>5</td>\n",
       "      <td>978300760</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>661</td>\n",
       "      <td>3</td>\n",
       "      <td>978302109</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>914</td>\n",
       "      <td>3</td>\n",
       "      <td>978301968</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>3408</td>\n",
       "      <td>4</td>\n",
       "      <td>978300275</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>2355</td>\n",
       "      <td>5</td>\n",
       "      <td>978824291</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   0     1  2          3\n",
       "0  1  1193  5  978300760\n",
       "1  1   661  3  978302109\n",
       "2  1   914  3  978301968\n",
       "3  1  3408  4  978300275\n",
       "4  1  2355  5  978824291"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# visualizando os primeiros registros das avaliações\n",
    "\n",
    "ratings.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# carregando os dados de treino e teste\n",
    "\n",
    "training_set = pd.read_csv('./data/u1.base', sep='\\t')\n",
    "test_set = pd.read_csv('./data/u1.test', sep='\\t')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# transformando os dados em um array numpy\n",
    "\n",
    "training_set = np.array(training_set, dtype='int')\n",
    "test_set = np.array(test_set, dtype='int')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# pegando o numero máximo de usuários e filmes\n",
    "\n",
    "nb_users = int(max(max(training_set[ :, 0]), max(test_set[ :, 0])))\n",
    "nb_movies = int(max(max(training_set[ :, 1]), max(test_set[ :, 1])))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# função para transformação dos dados de treino e teste para uma lista(users) de listas(movies)\n",
    "\n",
    "def convert(data):\n",
    "    new_data = []\n",
    "    \n",
    "    for id_users in range(1, nb_users + 1):\n",
    "        id_movies = data[:, 1][data[:, 0] == id_users]\n",
    "        id_ratings = data[:, 2][data[:, 0] == id_users]\n",
    "        ratings = np.zeros(nb_movies)\n",
    "        ratings[id_movies - 1] = id_ratings\n",
    "        new_data.append(list(ratings))\n",
    "    \n",
    "    return new_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# transformando os dados de treino e teste para uma lista(users) de listas(movies)\n",
    "\n",
    "training_set = convert(training_set)\n",
    "test_set = convert(test_set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "# convertendo os dados de treino e teste para o formato torch\n",
    "\n",
    "training_set = torch.FloatTensor(training_set)\n",
    "test_set = torch.FloatTensor(test_set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "class SAE(nn.Module):\n",
    "\n",
    "    def __init__(self, ):\n",
    "        super(SAE, self).__init__()           # chamando o construtor da classe mãe\n",
    "        \n",
    "        # definindo parametros da camada de entrada\n",
    "        self.fc1 = nn.Linear(nb_movies, 20)    # 20, numeros de features da camada de entrada\n",
    "        self.fc2 = nn.Linear(20, 10)           # 10, numeros de features da camada oculta\n",
    "        self.fc3 = nn.Linear(10, 20)           # 20, numeros de features da camada de saída\n",
    "        self.fc4 = nn.Linear(20, nb_movies)    # nb_movies, numeros de features da camada de saída\n",
    "        self.activation = nn.Sigmoid()         # função de ativação\n",
    "\n",
    "    def forward(self, x):\n",
    "        \n",
    "        x = self.activation(self.fc1(x))\n",
    "        x = self.activation(self.fc2(x))\n",
    "        x = self.activation(self.fc3(x))\n",
    "        x = self.fc4(x)\n",
    "        \n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# instanciando a classe SAE\n",
    "\n",
    "sae = SAE()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# parametros da rede\n",
    "\n",
    "criteon = nn.MSELoss()\n",
    "optimizer = optim.RMSprop(sae.parameters(), lr=0.01, weight_decay=0.5)\n",
    "nb_epochs = 200"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch: 1 loss: tensor(1.7711)\n",
      "epoch: 2 loss: tensor(1.0967)\n",
      "epoch: 3 loss: tensor(1.0535)\n",
      "epoch: 4 loss: tensor(1.0382)\n",
      "epoch: 5 loss: tensor(1.0311)\n",
      "epoch: 6 loss: tensor(1.0264)\n",
      "epoch: 7 loss: tensor(1.0236)\n",
      "epoch: 8 loss: tensor(1.0220)\n",
      "epoch: 9 loss: tensor(1.0205)\n",
      "epoch: 10 loss: tensor(1.0197)\n",
      "epoch: 11 loss: tensor(1.0188)\n",
      "epoch: 12 loss: tensor(1.0186)\n",
      "epoch: 13 loss: tensor(1.0178)\n",
      "epoch: 14 loss: tensor(1.0177)\n",
      "epoch: 15 loss: tensor(1.0172)\n",
      "epoch: 16 loss: tensor(1.0171)\n",
      "epoch: 17 loss: tensor(1.0168)\n",
      "epoch: 18 loss: tensor(1.0164)\n",
      "epoch: 19 loss: tensor(1.0163)\n",
      "epoch: 20 loss: tensor(1.0163)\n",
      "epoch: 21 loss: tensor(1.0161)\n",
      "epoch: 22 loss: tensor(1.0162)\n",
      "epoch: 23 loss: tensor(1.0161)\n",
      "epoch: 24 loss: tensor(1.0159)\n",
      "epoch: 25 loss: tensor(1.0159)\n",
      "epoch: 26 loss: tensor(1.0157)\n",
      "epoch: 27 loss: tensor(1.0154)\n",
      "epoch: 28 loss: tensor(1.0150)\n",
      "epoch: 29 loss: tensor(1.0134)\n",
      "epoch: 30 loss: tensor(1.0112)\n",
      "epoch: 31 loss: tensor(1.0100)\n",
      "epoch: 32 loss: tensor(1.0076)\n",
      "epoch: 33 loss: tensor(1.0088)\n",
      "epoch: 34 loss: tensor(1.0030)\n",
      "epoch: 35 loss: tensor(1.0018)\n",
      "epoch: 36 loss: tensor(0.9997)\n",
      "epoch: 37 loss: tensor(1.0002)\n",
      "epoch: 38 loss: tensor(0.9961)\n",
      "epoch: 39 loss: tensor(0.9949)\n",
      "epoch: 40 loss: tensor(0.9918)\n",
      "epoch: 41 loss: tensor(0.9937)\n",
      "epoch: 42 loss: tensor(0.9902)\n",
      "epoch: 43 loss: tensor(0.9890)\n",
      "epoch: 44 loss: tensor(0.9835)\n",
      "epoch: 45 loss: tensor(0.9833)\n",
      "epoch: 46 loss: tensor(0.9781)\n",
      "epoch: 47 loss: tensor(0.9802)\n",
      "epoch: 48 loss: tensor(0.9758)\n",
      "epoch: 49 loss: tensor(0.9758)\n",
      "epoch: 50 loss: tensor(0.9702)\n",
      "epoch: 51 loss: tensor(0.9672)\n",
      "epoch: 52 loss: tensor(0.9687)\n",
      "epoch: 53 loss: tensor(0.9642)\n",
      "epoch: 54 loss: tensor(0.9638)\n",
      "epoch: 55 loss: tensor(0.9632)\n",
      "epoch: 56 loss: tensor(0.9611)\n",
      "epoch: 57 loss: tensor(0.9603)\n",
      "epoch: 58 loss: tensor(0.9573)\n",
      "epoch: 59 loss: tensor(0.9554)\n",
      "epoch: 60 loss: tensor(0.9539)\n",
      "epoch: 61 loss: tensor(0.9534)\n",
      "epoch: 62 loss: tensor(0.9531)\n",
      "epoch: 63 loss: tensor(0.9524)\n",
      "epoch: 64 loss: tensor(0.9497)\n",
      "epoch: 65 loss: tensor(0.9497)\n",
      "epoch: 66 loss: tensor(0.9491)\n",
      "epoch: 67 loss: tensor(0.9495)\n",
      "epoch: 68 loss: tensor(0.9478)\n",
      "epoch: 69 loss: tensor(0.9475)\n",
      "epoch: 70 loss: tensor(0.9443)\n",
      "epoch: 71 loss: tensor(0.9456)\n",
      "epoch: 72 loss: tensor(0.9431)\n",
      "epoch: 73 loss: tensor(0.9449)\n",
      "epoch: 74 loss: tensor(0.9428)\n",
      "epoch: 75 loss: tensor(0.9439)\n",
      "epoch: 76 loss: tensor(0.9419)\n",
      "epoch: 77 loss: tensor(0.9422)\n",
      "epoch: 78 loss: tensor(0.9402)\n",
      "epoch: 79 loss: tensor(0.9410)\n",
      "epoch: 80 loss: tensor(0.9388)\n",
      "epoch: 81 loss: tensor(0.9397)\n",
      "epoch: 82 loss: tensor(0.9379)\n",
      "epoch: 83 loss: tensor(0.9393)\n",
      "epoch: 84 loss: tensor(0.9373)\n",
      "epoch: 85 loss: tensor(0.9379)\n",
      "epoch: 86 loss: tensor(0.9363)\n",
      "epoch: 87 loss: tensor(0.9367)\n",
      "epoch: 88 loss: tensor(0.9355)\n",
      "epoch: 89 loss: tensor(0.9355)\n",
      "epoch: 90 loss: tensor(0.9346)\n",
      "epoch: 91 loss: tensor(0.9348)\n",
      "epoch: 92 loss: tensor(0.9335)\n",
      "epoch: 93 loss: tensor(0.9341)\n",
      "epoch: 94 loss: tensor(0.9337)\n",
      "epoch: 95 loss: tensor(0.9338)\n",
      "epoch: 96 loss: tensor(0.9330)\n",
      "epoch: 97 loss: tensor(0.9330)\n",
      "epoch: 98 loss: tensor(0.9325)\n",
      "epoch: 99 loss: tensor(0.9322)\n",
      "epoch: 100 loss: tensor(0.9315)\n",
      "epoch: 101 loss: tensor(0.9314)\n",
      "epoch: 102 loss: tensor(0.9309)\n",
      "epoch: 103 loss: tensor(0.9315)\n",
      "epoch: 104 loss: tensor(0.9304)\n",
      "epoch: 105 loss: tensor(0.9302)\n",
      "epoch: 106 loss: tensor(0.9297)\n",
      "epoch: 107 loss: tensor(0.9295)\n",
      "epoch: 108 loss: tensor(0.9290)\n",
      "epoch: 109 loss: tensor(0.9293)\n",
      "epoch: 110 loss: tensor(0.9289)\n",
      "epoch: 111 loss: tensor(0.9287)\n",
      "epoch: 112 loss: tensor(0.9285)\n",
      "epoch: 113 loss: tensor(0.9284)\n",
      "epoch: 114 loss: tensor(0.9278)\n",
      "epoch: 115 loss: tensor(0.9277)\n",
      "epoch: 116 loss: tensor(0.9270)\n",
      "epoch: 117 loss: tensor(0.9271)\n",
      "epoch: 118 loss: tensor(0.9267)\n",
      "epoch: 119 loss: tensor(0.9270)\n",
      "epoch: 120 loss: tensor(0.9261)\n",
      "epoch: 121 loss: tensor(0.9264)\n",
      "epoch: 122 loss: tensor(0.9260)\n",
      "epoch: 123 loss: tensor(0.9260)\n",
      "epoch: 124 loss: tensor(0.9255)\n",
      "epoch: 125 loss: tensor(0.9255)\n",
      "epoch: 126 loss: tensor(0.9251)\n",
      "epoch: 127 loss: tensor(0.9247)\n",
      "epoch: 128 loss: tensor(0.9245)\n",
      "epoch: 129 loss: tensor(0.9245)\n",
      "epoch: 130 loss: tensor(0.9243)\n",
      "epoch: 131 loss: tensor(0.9238)\n",
      "epoch: 132 loss: tensor(0.9233)\n",
      "epoch: 133 loss: tensor(0.9234)\n",
      "epoch: 134 loss: tensor(0.9231)\n",
      "epoch: 135 loss: tensor(0.9231)\n",
      "epoch: 136 loss: tensor(0.9225)\n",
      "epoch: 137 loss: tensor(0.9227)\n",
      "epoch: 138 loss: tensor(0.9225)\n",
      "epoch: 139 loss: tensor(0.9225)\n",
      "epoch: 140 loss: tensor(0.9217)\n",
      "epoch: 141 loss: tensor(0.9218)\n",
      "epoch: 142 loss: tensor(0.9213)\n",
      "epoch: 143 loss: tensor(0.9213)\n",
      "epoch: 144 loss: tensor(0.9209)\n",
      "epoch: 145 loss: tensor(0.9209)\n",
      "epoch: 146 loss: tensor(0.9206)\n",
      "epoch: 147 loss: tensor(0.9205)\n",
      "epoch: 148 loss: tensor(0.9200)\n",
      "epoch: 149 loss: tensor(0.9200)\n",
      "epoch: 150 loss: tensor(0.9195)\n",
      "epoch: 151 loss: tensor(0.9195)\n",
      "epoch: 152 loss: tensor(0.9193)\n",
      "epoch: 153 loss: tensor(0.9193)\n",
      "epoch: 154 loss: tensor(0.9190)\n",
      "epoch: 155 loss: tensor(0.9188)\n",
      "epoch: 156 loss: tensor(0.9185)\n",
      "epoch: 157 loss: tensor(0.9183)\n",
      "epoch: 158 loss: tensor(0.9181)\n",
      "epoch: 159 loss: tensor(0.9180)\n",
      "epoch: 160 loss: tensor(0.9176)\n",
      "epoch: 161 loss: tensor(0.9174)\n",
      "epoch: 162 loss: tensor(0.9174)\n",
      "epoch: 163 loss: tensor(0.9172)\n",
      "epoch: 164 loss: tensor(0.9169)\n",
      "epoch: 165 loss: tensor(0.9169)\n",
      "epoch: 166 loss: tensor(0.9168)\n",
      "epoch: 167 loss: tensor(0.9166)\n",
      "epoch: 168 loss: tensor(0.9164)\n",
      "epoch: 169 loss: tensor(0.9164)\n",
      "epoch: 170 loss: tensor(0.9161)\n",
      "epoch: 171 loss: tensor(0.9159)\n",
      "epoch: 172 loss: tensor(0.9157)\n",
      "epoch: 173 loss: tensor(0.9153)\n",
      "epoch: 174 loss: tensor(0.9157)\n",
      "epoch: 175 loss: tensor(0.9154)\n",
      "epoch: 176 loss: tensor(0.9154)\n",
      "epoch: 177 loss: tensor(0.9152)\n",
      "epoch: 178 loss: tensor(0.9148)\n",
      "epoch: 179 loss: tensor(0.9147)\n",
      "epoch: 180 loss: tensor(0.9145)\n",
      "epoch: 181 loss: tensor(0.9144)\n",
      "epoch: 182 loss: tensor(0.9145)\n",
      "epoch: 183 loss: tensor(0.9141)\n",
      "epoch: 184 loss: tensor(0.9140)\n",
      "epoch: 185 loss: tensor(0.9139)\n",
      "epoch: 186 loss: tensor(0.9139)\n",
      "epoch: 187 loss: tensor(0.9138)\n",
      "epoch: 188 loss: tensor(0.9137)\n",
      "epoch: 189 loss: tensor(0.9136)\n",
      "epoch: 190 loss: tensor(0.9135)\n",
      "epoch: 191 loss: tensor(0.9132)\n",
      "epoch: 192 loss: tensor(0.9130)\n",
      "epoch: 193 loss: tensor(0.9128)\n",
      "epoch: 194 loss: tensor(0.9129)\n",
      "epoch: 195 loss: tensor(0.9125)\n",
      "epoch: 196 loss: tensor(0.9126)\n",
      "epoch: 197 loss: tensor(0.9125)\n",
      "epoch: 198 loss: tensor(0.9125)\n",
      "epoch: 199 loss: tensor(0.9123)\n",
      "epoch: 200 loss: tensor(0.9124)\n"
     ]
    }
   ],
   "source": [
    "# treinando a rede\n",
    "\n",
    "for epoch in range(1, nb_epochs +1):\n",
    "    \n",
    "    train_loss = 0\n",
    "    s = 0.\n",
    "\n",
    "    for id_user in range(nb_users):\n",
    "        \n",
    "        input = Variable(training_set[id_user]).unsqueeze(0)\n",
    "        target = input.clone()\n",
    "\n",
    "        if torch.sum(target.data > 0) > 0:\n",
    "        \n",
    "            output = sae(input)\n",
    "            target.require_grad = False\n",
    "            output[target == 0] = 0\n",
    "            loss = criteon(output, target)\n",
    "            mean_corrector = nb_movies/float(torch.sum(target.data > 0) + 1e-10)\n",
    "            loss.backward()\n",
    "            train_loss += np.sqrt(loss.data*mean_corrector)\n",
    "            s += 1.\n",
    "            optimizer.step()\n",
    "\n",
    "    print('epoch: ' + str(epoch) + ' loss: ' + str(train_loss/s))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "test loss: tensor(0.9507)\n"
     ]
    }
   ],
   "source": [
    "# testando o modelo\n",
    "\n",
    "test_loss = 0\n",
    "s = 0.\n",
    "\n",
    "for id_users in range(nb_users):\n",
    "    \n",
    "    input = Variable(training_set[id_users]).unsqueeze(0)\n",
    "    target = Variable(test_set[id_users]).unsqueeze(0)\n",
    "\n",
    "    if torch.sum(target.data > 0) > 0:\n",
    "        \n",
    "        output = sae(input)\n",
    "        target.require_grad = False\n",
    "        output[target == 0] = 0\n",
    "        loss = criteon(output, target)\n",
    "        mean_corrector = nb_movies/float(torch.sum(target.data > 0) + 1e-10)\n",
    "        test_loss += np.sqrt(loss.data*mean_corrector)\n",
    "        s += 1.\n",
    "\n",
    "print('test loss: ' + str(test_loss/s))"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "ad2bdc8ecc057115af97d19610ffacc2b4e99fae6737bb82f5d7fb13d2f2c186"
  },
  "kernelspec": {
   "display_name": "Python 3.8.8 ('base')",
   "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": "3.8.8"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}