MLP_autoencoder.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "a2d57698",
   "metadata": {},
   "source": [
    "# Implementación de Autoencoders\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "893438d9",
   "metadata": {},
   "source": [
    "## Librerias "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "162f4f9a",
   "metadata": {},
   "source": [
    "###### Importar librerias completas "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "b5e2783c",
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch #para realizar operaciones con tensores es la libreria basica\n",
    "import torchvision #Contiene paquetes para trabajar con datasets populares, arquitecturas e imagenes de vision computacional\n",
    "import matplotlib.pyplot as plt \n",
    "from matplotlib.pyplot      import close, figure, imshow, savefig, show, title\n",
    "from matplotlib.lines       import Line2D\n",
    "from os.path                import join\n",
    "import numpy as np\n",
    "#import winsound "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d89ee939",
   "metadata": {},
   "source": [
    "## Hardware "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "d57d8375",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "¿Esta el GPU disponible?: True\n",
      "Propiedades del dispositivo: _CudaDeviceProperties(name='NVIDIA GeForce RTX 2080 SUPER', major=7, minor=5, total_memory=7966MB, multi_processor_count=48)\n"
     ]
    }
   ],
   "source": [
    "print(\"¿Esta el GPU disponible?:\",torch.cuda.is_available())\n",
    "device = torch.cuda.current_device()\n",
    "print('Propiedades del dispositivo:',torch.cuda.get_device_properties(device))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bdc5ea25",
   "metadata": {},
   "source": [
    "## Datos"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9479fe60",
   "metadata": {},
   "source": [
    "###### Importar datos"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "05131983",
   "metadata": {},
   "outputs": [],
   "source": [
    "img_transform=torchvision.transforms.Compose([\n",
    "    torchvision.transforms.ToTensor(),\n",
    "    torchvision.transforms.Normalize([0.5],[0.5])\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "012ebfdb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tamaño del conjunto de datos 60000\n"
     ]
    }
   ],
   "source": [
    "dataset=torchvision.datasets.MNIST(root='./data', train=True, transform=img_transform, download=True)\n",
    "print('Tamaño del conjunto de datos',len(dataset))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "74668caa",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tamaño del conjunto de datos de validación 10000\n"
     ]
    }
   ],
   "source": [
    "dataset_validacion =torchvision.datasets.MNIST(root='./data', train=False, transform=img_transform, download=True)\n",
    "print('Tamaño del conjunto de datos de validación',len(dataset_validacion))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fd6c5067",
   "metadata": {},
   "source": [
    "###### Preparar datos de entrenamiento "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "0f03f7e5",
   "metadata": {},
   "outputs": [],
   "source": [
    "dataloader=torch.utils.data.DataLoader(dataset, batch_size=512, shuffle=True)\n",
    "dataloader_validacion = torch.utils.data.DataLoader(dataset_validacion, batch_size=10000, shuffle=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8142cea1",
   "metadata": {},
   "source": [
    "## Arquitectura"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "d0ea8b10",
   "metadata": {},
   "outputs": [],
   "source": [
    "class autoencoder(torch.nn.Module):\n",
    "    \n",
    "    def __init__(self):\n",
    "        super(autoencoder, self).__init__()\n",
    "        \n",
    "        self.encoder = torch.nn.Sequential(\n",
    "            \n",
    "            torch.nn.Linear(28*28, 128), \n",
    "            torch.nn.ReLU(True),\n",
    "            \n",
    "            torch.nn.Linear(128,64),\n",
    "            torch.nn.ReLU(True),\n",
    "            \n",
    "            torch.nn.Linear(64,12), \n",
    "            torch.nn.ReLU(True), \n",
    "            \n",
    "            torch.nn.Linear(12,1))\n",
    "        \n",
    "        self.decoder = torch.nn.Sequential(\n",
    "            torch.nn.Linear(1,12),\n",
    "            torch.nn.ReLU(True),\n",
    "            \n",
    "            torch.nn.Linear(12,64),\n",
    "            torch.nn.ReLU(True),\n",
    "            \n",
    "            torch.nn.Linear(64, 128),\n",
    "            torch.nn.ReLU(True), \n",
    "            \n",
    "            torch.nn.Linear(128, 28*28), \n",
    "            torch.nn.Tanh())\n",
    "        self.encode = True\n",
    "        self.decode = True\n",
    "\n",
    "    def forward(self, x):\n",
    "        \n",
    "        x = self.encoder(x)\n",
    "        x =self.decoder(x)\n",
    "        return x"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d48f02e",
   "metadata": {},
   "source": [
    "## Entrenamiento "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "014e8b95",
   "metadata": {},
   "outputs": [],
   "source": [
    "model=autoencoder().cuda()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "id": "6227294a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Entrenamiento\n",
    "# Entradas: Dataset, modelo, funciondeperdida, optimizador\n",
    "# Salidas: Error\n",
    "def entrena(dataloader, model, loss_fn, optimizer):\n",
    "    size = len(dataloader.dataset)\n",
    "    for batch, (X,y) in enumerate(dataloader):\n",
    "        img = X\n",
    "        img = img.view(img.size(0), -1)\n",
    "        img = torch.autograd.Variable(img).cuda() \n",
    "\n",
    "        #=============forward==================\n",
    "        output = model(img)\n",
    "        loss = loss_fn(output, img)\n",
    "\n",
    "        #=============backward=================\n",
    "        optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        #if batch % 20 == 0:\n",
    "        #    loss, current = loss.item(), (batch + 1) * len(X)\n",
    "        #    print(f\"loss: {loss:>7f}  [{current:>5d}/{size:>5d}]\")\n",
    "    return loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "id": "7e7200f1",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Validación\n",
    "# Entradas: Dataset, modelo, funciondeperdida\n",
    "# Salidas: Error_promedio\n",
    "def valida(dataloader, model, loss_fn):\n",
    "    num_batches = len(dataloader)\n",
    "    loss = 0.0\n",
    "\n",
    "    with torch.no_grad():\n",
    "        for data in dataloader:\n",
    "            img, _ = data\n",
    "            img = img.view(img.size(0), -1)\n",
    "            img = torch.autograd.Variable(img).cuda() \n",
    "            prediccion = model(img)\n",
    "            test_loss = loss_fn(prediccion, img)\n",
    "            loss += test_loss.item()\n",
    "    \n",
    "    loss /= num_batches\n",
    "    #print(\"Error prueba: El error promedio es: {}\".format(loss))\n",
    "    return loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "id": "2c14b462",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Época 1\n",
      "-------------------------------\n",
      "loss min: 0.8901017308235168\n",
      "Numero de epoca: 1 Funcion de perdida: 0.89010173\n",
      "Época 2\n",
      "-------------------------------\n",
      "loss min: 0.8472817540168762\n",
      "Final:\n",
      "\n",
      "Numero de epoca: 2 Funcion de perdida: tensor(0.8473, device='cuda:0', grad_fn=<MseLossBackward0>)\n"
     ]
    }
   ],
   "source": [
    "num_epochs=2 #Numero de epocas\n",
    "criterion=torch.nn.MSELoss() #Metrica error\n",
    "optimizer=torch.optim.Adam(  #Optimizador\n",
    "    model.parameters(), \n",
    "    lr=1e-5, \n",
    "    weight_decay=1e-5)\n",
    "\n",
    "loss_ant = 10.0\n",
    "losses = {\"entrenamiento\":[], \"validación\":[]}\n",
    "for epoch in range(num_epochs):\n",
    "    \n",
    "    print(f\"Época {epoch+1}\\n-------------------------------\")\n",
    "    loss_entrena = entrena(dataloader, model, criterion, optimizer)\n",
    "    loss_valida = valida(dataloader_validacion, model, criterion)\n",
    "    #============saveb_weights==============\n",
    "    losses[\"entrenamiento\"].append(loss_entrena.item())\n",
    "    losses[\"validación\"].append(loss_valida)\n",
    "    \n",
    "    if loss_entrena.item() < loss_ant:\n",
    "        torch.save(model.state_dict(), 'weights_entrenamiento_autoencMNISTprueba.pth')\n",
    "        loss_ant = loss_entrena.item()\n",
    "        mejor = epoch\n",
    "        print(\"loss min: {}\".format(loss_ant))\n",
    "\n",
    "    #================log========================    \n",
    "    if epoch % 10 == 0:\n",
    "        loss_value=loss_entrena.cpu().detach().numpy()\n",
    "        print(\"Numero de epoca:\",epoch+1,'Funcion de perdida:',loss_value)\n",
    "\n",
    "print('Final:\\n')        \n",
    "print(\"Numero de epoca:\",epoch+1,'Funcion de perdida:',loss_entrena)\n",
    "#winsound.Beep(440,1000)    "
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "948b86f4",
   "metadata": {},
   "source": [
    "#### Ploteo de errores durante el entrenamiento"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "id": "cb154081",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot losses\n",
    "plt.figure(figsize=(10,8))\n",
    "plt.semilogy(losses['entrenamiento'], label='Entrenamiento')\n",
    "plt.semilogy(losses['validación'], label='Validación')\n",
    "plt.xlabel('Época')\n",
    "plt.ylabel('Error promedio')\n",
    "#plt.grid()\n",
    "plt.legend()\n",
    "plt.title('Error entrenamiento autoencoder')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "346741b1",
   "metadata": {},
   "outputs": [],
   "source": [
    "torch.save({'epoch': epoch,\n",
    "        'state_dict': model.state_dict(),\n",
    "        'optimizer_state_dict': optimizer.state_dict(),\n",
    "        'loss': loss}, \n",
    "        'MNIST_autoencoder.pth')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fa74ae71",
   "metadata": {},
   "source": [
    "## Resultados "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "8839d0ed",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "autoencoder(\n",
       "  (encoder): Sequential(\n",
       "    (0): Linear(in_features=784, out_features=128, bias=True)\n",
       "    (1): ReLU(inplace=True)\n",
       "    (2): Linear(in_features=128, out_features=64, bias=True)\n",
       "    (3): ReLU(inplace=True)\n",
       "    (4): Linear(in_features=64, out_features=12, bias=True)\n",
       "    (5): ReLU(inplace=True)\n",
       "    (6): Linear(in_features=12, out_features=1, bias=True)\n",
       "  )\n",
       "  (decoder): Sequential(\n",
       "    (0): Linear(in_features=1, out_features=12, bias=True)\n",
       "    (1): ReLU(inplace=True)\n",
       "    (2): Linear(in_features=12, out_features=64, bias=True)\n",
       "    (3): ReLU(inplace=True)\n",
       "    (4): Linear(in_features=64, out_features=128, bias=True)\n",
       "    (5): ReLU(inplace=True)\n",
       "    (6): Linear(in_features=128, out_features=784, bias=True)\n",
       "    (7): Tanh()\n",
       "  )\n",
       ")"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "##GPU\n",
    "#model_test=autoencoder().cuda()\n",
    "\n",
    "##CPU\n",
    "model_test=autoencoder()\n",
    "checkpoint_test=torch.load('weights_entrenamiento_autoencMNIST.pth')\n",
    "model_test.load_state_dict(torch.load('weights_entrenamiento_autoencMNIST.pth'))\n",
    "model_test.eval()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "249e7b53",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Salida Codificador')"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "input_, _ = dataset[582] \n",
    "#procesamiento dato de entrada\n",
    "input_test = input_.view(input_.size(0), -1)\n",
    "output_test=model_test(input_test)\n",
    "code_output=model_test.encoder(input_test)\n",
    "#procesamiento dato de salida de autoencoder\n",
    "output_test=output_test.reshape(28,28)\n",
    "output_test=output_test.detach().numpy() \n",
    "#procesamiento de dato salida del codificador\n",
    "img_code=code_output.reshape(1,-1)\n",
    "img_code=img_code.detach().numpy()\n",
    "\n",
    "\n",
    "#Visualización\n",
    "plt.subplot(1, 3, 1)\n",
    "plt.imshow(input_.numpy()[0],cmap='gray')\n",
    "plt.title(\"Entrada\")\n",
    "\n",
    "\n",
    "plt.subplot(1, 3, 3)\n",
    "plt.imshow(output_test,cmap='gray')\n",
    "plt.title(\"Salida Decodificador\")\n",
    "\n",
    "plt.subplot(1, 3, 2)\n",
    "plt.imshow(img_code,cmap='gray')\n",
    "plt.title(\"Salida Codificador\")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "76e7146f",
   "metadata": {},
   "source": [
    "#### Generación de imagenes con el espacio latente\n",
    "\n",
    "Generaremos nuevas imágenes utilizando el espacio latente, para ello del conjunto de datos de validación se calculará una distribución normal con la media y una desviación estandar de los datos condificados por el _encoder_. El resultado se ingresará al _decoder_, que será el encargado de reconstruir las imágenes según el valor ingresado."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "72347a46",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x850 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def show_image(img):\n",
    "    npimg = img.numpy()\n",
    "    plt.imshow(np.transpose(npimg, (1, 2, 0)))\n",
    "\n",
    "model.eval()\n",
    "\n",
    "with torch.no_grad():\n",
    "    \n",
    "    # calcula la media y la desviación estándar del código latente, generado tomando imágenes de prueba como entradas \n",
    "    dataiter = iter(dataloader_validacion)\n",
    "    images, labels = next(dataiter)\n",
    "    images = images.view(images.size(0), -1)\n",
    "    latent = model_test.encoder(images)\n",
    "    latent = latent.cpu()\n",
    "    mean = latent.mean(dim=0)\n",
    "    std = (latent - mean).pow(2).mean(dim=0).sqrt()\n",
    "\n",
    "    # muestreo de vectores latentes de la distribución normal\n",
    "    latent = torch.randn(512, 1)*std + mean\n",
    "\n",
    "    # Reconstrucción con valores al azar del vector latente\n",
    "    img_recon = model_test.decoder(latent)\n",
    "    img_recon = img_recon.reshape(14336,28)\n",
    "    img_recon = img_recon.cpu()\n",
    "    \n",
    "    # visualizacion de 5 valores resultantes\n",
    "    fig, ax = plt.subplots(figsize=(20, 8.5))\n",
    "    show_image(torchvision.utils.make_grid(img_recon[:140],10,5))\n",
    "    plt.show()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "09e85ae1",
   "metadata": {},
   "source": [
    "## Graficos de espacio latente\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "fb665203",
   "metadata": {},
   "outputs": [],
   "source": [
    "datos = {\"codificado\":[], \"etiqueta\":[]}\n",
    "dataiter = iter(dataloader_validacion)\n",
    "images, labels = next(dataiter)\n",
    "images = images.view(images.size(0), -1)\n",
    "encoded_features = model_test.encoder(images)\n",
    "for i in range (len(encoded_features)):\n",
    "\n",
    "    datos['codificado'].append(encoded_features[i].item())\n",
    "    datos['etiqueta'].append(labels[i])\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "fd9a2e73",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot\n",
    "fig, ax = plt.subplots()\n",
    "\n",
    "ax.scatter(datos[\"codificado\"], datos[\"etiqueta\"], marker = '.')\n",
    "\n",
    "ax.set(xlabel='Salida del codificador', ylabel='Valores codificados', \n",
    "       xlim=(min (datos['codificado'])-1, max(datos['codificado'])+1), \n",
    "       xticks=np.arange(min (datos['codificado']), max(datos['codificado']), 10),\n",
    "       ylim=(-1, 10), yticks=np.arange(0,10))\n",
    "\n",
    "plt.title(\"Plot espacio latente\")\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "pytorch",
   "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": "3.9.15"
  },
  "vscode": {
   "interpreter": {
    "hash": "41bac1db570d670af06c178c7b2ba4ebdd2ebe54b9e7ab5cd42c7c62dc5e55f2"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}