mnist_test_with_own_input.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# This code is based from the following:\n",
    "# https://github.com/makeyourownneuralnetwork/makeyourownneuralnetwork/blob/master/part3_neural_network_mnist_and_own_single_image.ipynb\n",
    "# Author: Tariq Rashid\n",
    "# License: GNU GENERAL PUBLIC LICENSE\n",
    "\n",
    "import pickle\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import imageio"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading image...\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Raw estimates:\n",
      "[[ 0.24972299 -0.03186571  0.06568367  0.2332831   0.31352327  0.33646345\n",
      "  -0.13489472  0.27846039  0.14665539 -0.32803075]]\n",
      "The network says it's a 5\n"
     ]
    }
   ],
   "source": [
    "# Test the neural network with our own images\n",
    "\n",
    "# Load image data from png file into an array\n",
    "print (\"Loading image...\")\n",
    "img_array = imageio.imread('sample_input.png', as_gray=True)\n",
    "    \n",
    "# Reshape from 28x28 to list of 784 values, then invert values\n",
    "img_data  = 255 - img_array.reshape(784)\n",
    "    \n",
    "# Scale data to range from 0.01 to 1.0\n",
    "img_data = (img_data / 255.0 * 0.99) + 0.01\n",
    "\n",
    "# Plot image\n",
    "plt.title(f\"Showing input image\")\n",
    "plt.imshow(img_data.reshape(28,28), cmap='gray')\n",
    "plt.show()\n",
    "\n",
    "# Perform prediction on own image\n",
    "model = pickle.load(open('models/mnist_net.p', 'rb'))\n",
    "print('Raw estimates:')\n",
    "predictions = model.forward(img_data)\n",
    "print(predictions)\n",
    "\n",
    "clean_predictions = predictions.argmax()\n",
    "print(f\"The network says it's a {clean_predictions}\")"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "0f3556a7753b14587b2ccf5606ed10d03c0f0411da4e1e635276bea95ffe7800"
  },
  "kernelspec": {
   "display_name": "Python 3.9.6 64-bit ('base': conda)",
   "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": "3.9.6"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}