redes-neuronales/Cuadernos/Pythorch_04_transformaciones.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "846d2760",
   "metadata": {},
   "source": [
    "<figure>\n",
    "<img src=\"../Imagenes/logo-final-ap.png\"  width=\"80\" height=\"80\" align=\"left\"/> \n",
    "</figure>\n",
    "\n",
    "# <span style=\"color:#4361EE\"><left>Aprendizaje Profundo</left></span>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "36b1714f",
   "metadata": {},
   "source": [
    "# <span style=\"color:red\"><center>Diplomado en Inteligencia Artificial y Aprendizaje Profundo</center></span>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ahead-abuse",
   "metadata": {
    "tags": []
   },
   "source": [
    "# <span style=\"color:green\"><center>Transforms</center></span>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5118173d",
   "metadata": {},
   "source": [
    "<figure>\n",
    "<center>\n",
    "<img src=\"https://raw.githubusercontent.com/AprendizajeProfundo/Alejandria/main/Archivos_Generales/Imagenes/Robot.jpg\" width=\"600\" height=\"400\" align=\"center\" /> \n",
    "</center>   \n",
    "</figure>\n",
    "<center>\n",
    "\n",
    "Fuente: [Pexels Alex Knight](https://www.pexels.com/es-es/foto/foto-de-alto-angulo-del-robot-2599244/)\n",
    "\n",
    "</center>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0550d999",
   "metadata": {},
   "source": [
    "## <span style=\"color:#4361EE\">Profesores</span>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "80440db5",
   "metadata": {},
   "source": [
    "1. Alvaro  Montenegro, PhD, ammontenegrod@unal.edu.co\n",
    "1. Camilo José Torres Jiménez, Msc, cjtorresj@unal.edu.co\n",
    "1. Daniel  Montenegro, Msc, dextronomo@gmail.com "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0c9cdf3a",
   "metadata": {},
   "source": [
    "## <span style=\"color:#4361EE\">Asesora Medios y Marketing digital</span>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fa341c44",
   "metadata": {},
   "source": [
    "4. Maria del Pilar Montenegro, pmontenegro88@gmail.com\n",
    "5. Jessica López Mejía, jelopezme@unal.edu.co\n",
    "6. Venus Puertas, vpuertasg@unal.edu.co"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "70ad8889",
   "metadata": {},
   "source": [
    "## <span style=\"color:#4361EE\">Jefe Jurídica</span>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "77105787",
   "metadata": {},
   "source": [
    "7. Paula Andrea Guzmán, guzmancruz.paula@gmail.com"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "825dd85b",
   "metadata": {},
   "source": [
    "## <span style=\"color:#4361EE\">Coordinador Jurídico</span>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8d28332e",
   "metadata": {},
   "source": [
    "8. David Fuentes, fuentesd065@gmail.com"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a76deacf",
   "metadata": {},
   "source": [
    "## <span style=\"color:#4361EE\">Desarrolladores Principales</span>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9209c7b2",
   "metadata": {},
   "source": [
    "9. Dairo Moreno, damoralesj@unal.edu.co\n",
    "10. Joan Castro, jocastroc@unal.edu.co\n",
    "11. Bryan Riveros, briveros@unal.edu.co\n",
    "12. Rosmer Vargas, rovargasc@unal.edu.co"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a6f31cfb",
   "metadata": {},
   "source": [
    "## <span style=\"color:#4361EE\">Expertos en Bases de Datos</span>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6730dc42",
   "metadata": {},
   "source": [
    "13. Giovvani Barrera, udgiovanni@gmail.com\n",
    "14. Camilo Chitivo, cchitivo@unal.edu.co"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "unlimited-testament",
   "metadata": {},
   "source": [
    "## <span style=\"color:#4361EE\">Referencias</span> "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "every-powell",
   "metadata": {},
   "source": [
    "1. Basado en los [tutoriales de Pytorch](https://pytorch.org/tutorials/)\n",
    "1. [Deep learning for coders with FastAI and Pytorch](http://library.lol/main/F13E85845AE48D9FD7488FE7630A9FD3)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "practical-dress",
   "metadata": {},
   "source": [
    "## <span style=\"color:#4361EE\">Contenido</span>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "amateur-embassy",
   "metadata": {},
   "source": [
    "* [Introducción](#Introducción)\n",
    "* [Carga de un Dataset](#Carga-de-un-Dataset)\n",
    "* [Transformaciones](#Transformaciones)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "coordinate-bowling",
   "metadata": {},
   "source": [
    "## <span style=\"color:#4361EE\">Introducción</span>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "received-breast",
   "metadata": {},
   "source": [
    "Los datos no siempre vienen en su forma procesada final que se requiere para entrenar algoritmos de aprendizaje automático. Usamos transformaciones para realizar alguna manipulación de los datos y hacerlos aptos para el entrenamiento.\n",
    "\n",
    "Todos los conjuntos de datos de TorchVision tienen dos parámetros *transform* para modificar las características y *target_transform* para modificar las etiquetas- que aceptan llamadas que contienen la lógica de transformación. El módulo *torchvision.transforms* ofrece varias transformaciones de uso común listas para usar."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "related-ladder",
   "metadata": {},
   "source": [
    "## <span style=\"color:#4361EE\">Carga de un Dataset</span>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "bright-tuition",
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "from torchvision import datasets\n",
    "from torchvision.transforms import ToTensor, Lambda"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "owned-danger",
   "metadata": {},
   "source": [
    "## <span style=\"color:#4361EE\">Transformaciones</span>\n",
    "\n",
    " "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "abstract-engine",
   "metadata": {},
   "source": [
    "### <span style=\"color:#4CC9F0\">ToTensor()</span>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "distinguished-belgium",
   "metadata": {},
   "source": [
    "`ToTensor` convierte una imagen PIL o NumPy ndarray en un FloatTensor. y escala los valores de intensidad de píxeles de la imagen en el rango [0., 1.]."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "committed-butterfly",
   "metadata": {},
   "source": [
    "### <span style=\"color:#4CC9F0\">Transformaciones Lambda</span>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "faced-composite",
   "metadata": {},
   "source": [
    "Las transformaciones `Lambda` aplican cualquier función *lambda* definida por el usuario. Aquí, definimos una función para convertir el número entero en un tensor codificado en caliente. Primero crea un tensor cero de tamaño 10 (el número de etiquetas en nuestro conjunto de datos) y llama a `scatter_`, que asigna un valor = 1 en el índice indicado por la etiqueta `y`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "distinct-customs",
   "metadata": {},
   "outputs": [],
   "source": [
    "ds = datasets.FashionMNIST(\n",
    "    root=\"data\",\n",
    "    train=True,\n",
    "    download=True,\n",
    "    transform=ToTensor(),\n",
    "    target_transform=Lambda(lambda y: torch.zeros(10, dtype=torch.float).scatter_(0, torch.tensor(y), value=1))\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "cf0877c3-2c68-42a2-82cb-5525e49e3502",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 9 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "figure = plt.figure(figsize=(8,8))\n",
    "cols, rows = 3,3\n",
    "for i in range(1, cols*rows+1):\n",
    "    sample_idx = torch.randint(len(ds), size =(1,)).item()\n",
    "    img, label = ds[sample_idx]\n",
    "    figure.add_subplot(rows, cols, i)\n",
    "    plt.title(label, fontsize=7)\n",
    "    plt.axis('off')\n",
    "    plt.imshow(img.squeeze(), cmap='gray')\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.16"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}