pw04/MLP_from_HOG.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "X_wPegydbafH"
   },
   "source": [
    "# Computing features to train a MLP\n",
    "This notebook will guide you through the use of the `keras` package to train a multilayer perceptron for handwritten digits classification. You are going to computing the histogram of gradients from the images in the `mnist` dataset (LeCun et al. 1998)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "61oGWRyMbafJ"
   },
   "source": [
    "## Loading the packages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "id": "duI9Ar-8bafS"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: tensorflow in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (2.12.0)\n",
      "Requirement already satisfied: numpy<1.24,>=1.22 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (1.23.5)\n",
      "Requirement already satisfied: astunparse>=1.6.0 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (1.6.3)\n",
      "Requirement already satisfied: grpcio<2.0,>=1.24.3 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (1.53.0)\n",
      "Requirement already satisfied: google-pasta>=0.1.1 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (0.2.0)\n",
      "Requirement already satisfied: gast<=0.4.0,>=0.2.1 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (0.4.0)\n",
      "Requirement already satisfied: protobuf!=4.21.0,!=4.21.1,!=4.21.2,!=4.21.3,!=4.21.4,!=4.21.5,<5.0.0dev,>=3.20.3 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (4.22.1)\n",
      "Requirement already satisfied: termcolor>=1.1.0 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (2.2.0)\n",
      "Requirement already satisfied: libclang>=13.0.0 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (16.0.0)\n",
      "Requirement already satisfied: flatbuffers>=2.0 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (23.3.3)\n",
      "Requirement already satisfied: typing-extensions>=3.6.6 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (4.5.0)\n",
      "Requirement already satisfied: absl-py>=1.0.0 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (1.4.0)\n",
      "Requirement already satisfied: tensorboard<2.13,>=2.12 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (2.12.1)\n",
      "Requirement already satisfied: h5py>=2.9.0 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (3.8.0)\n",
      "Requirement already satisfied: opt-einsum>=2.3.2 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (3.3.0)\n",
      "Requirement already satisfied: packaging in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (23.1)\n",
      "Requirement already satisfied: wrapt<1.15,>=1.11.0 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (1.14.1)\n",
      "Requirement already satisfied: keras<2.13,>=2.12.0 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (2.12.0)\n",
      "Requirement already satisfied: tensorflow-estimator<2.13,>=2.12.0 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (2.12.0)\n",
      "Requirement already satisfied: tensorflow-io-gcs-filesystem>=0.23.1 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (0.32.0)\n",
      "Requirement already satisfied: setuptools in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (65.6.3)\n",
      "Requirement already satisfied: jax>=0.3.15 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (0.4.8)\n",
      "Requirement already satisfied: six>=1.12.0 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorflow) (1.16.0)\n",
      "Requirement already satisfied: wheel<1.0,>=0.23.0 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from astunparse>=1.6.0->tensorflow) (0.38.4)\n",
      "Requirement already satisfied: scipy>=1.7 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from jax>=0.3.15->tensorflow) (1.9.1)\n",
      "Requirement already satisfied: ml-dtypes>=0.0.3 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from jax>=0.3.15->tensorflow) (0.1.0)\n",
      "Requirement already satisfied: google-auth-oauthlib<1.1,>=0.5 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorboard<2.13,>=2.12->tensorflow) (1.0.0)\n",
      "Requirement already satisfied: tensorboard-data-server<0.8.0,>=0.7.0 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorboard<2.13,>=2.12->tensorflow) (0.7.0)\n",
      "Requirement already satisfied: google-auth<3,>=1.6.3 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorboard<2.13,>=2.12->tensorflow) (2.17.2)\n",
      "Requirement already satisfied: markdown>=2.6.8 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorboard<2.13,>=2.12->tensorflow) (3.4.3)\n",
      "Requirement already satisfied: werkzeug>=1.0.1 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorboard<2.13,>=2.12->tensorflow) (2.2.3)\n",
      "Requirement already satisfied: tensorboard-plugin-wit>=1.6.0 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorboard<2.13,>=2.12->tensorflow) (1.8.1)\n",
      "Requirement already satisfied: requests<3,>=2.21.0 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from tensorboard<2.13,>=2.12->tensorflow) (2.28.2)\n",
      "Requirement already satisfied: pyasn1-modules>=0.2.1 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from google-auth<3,>=1.6.3->tensorboard<2.13,>=2.12->tensorflow) (0.2.8)\n",
      "Requirement already satisfied: cachetools<6.0,>=2.0.0 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from google-auth<3,>=1.6.3->tensorboard<2.13,>=2.12->tensorflow) (5.3.0)\n",
      "Requirement already satisfied: rsa<5,>=3.1.4 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from google-auth<3,>=1.6.3->tensorboard<2.13,>=2.12->tensorflow) (4.9)\n",
      "Requirement already satisfied: requests-oauthlib>=0.7.0 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from google-auth-oauthlib<1.1,>=0.5->tensorboard<2.13,>=2.12->tensorflow) (1.3.1)\n",
      "Requirement already satisfied: importlib-metadata>=4.4 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from markdown>=2.6.8->tensorboard<2.13,>=2.12->tensorflow) (6.3.0)\n",
      "Requirement already satisfied: certifi>=2017.4.17 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from requests<3,>=2.21.0->tensorboard<2.13,>=2.12->tensorflow) (2022.12.7)\n",
      "Requirement already satisfied: urllib3<1.27,>=1.21.1 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from requests<3,>=2.21.0->tensorboard<2.13,>=2.12->tensorflow) (1.26.15)\n",
      "Requirement already satisfied: idna<4,>=2.5 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from requests<3,>=2.21.0->tensorboard<2.13,>=2.12->tensorflow) (3.4)\n",
      "Requirement already satisfied: charset-normalizer<4,>=2 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from requests<3,>=2.21.0->tensorboard<2.13,>=2.12->tensorflow) (3.1.0)\n",
      "Requirement already satisfied: MarkupSafe>=2.1.1 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from werkzeug>=1.0.1->tensorboard<2.13,>=2.12->tensorflow) (2.1.2)\n",
      "Requirement already satisfied: zipp>=0.5 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from importlib-metadata>=4.4->markdown>=2.6.8->tensorboard<2.13,>=2.12->tensorflow) (3.15.0)\n",
      "Requirement already satisfied: pyasn1<0.5.0,>=0.4.6 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from pyasn1-modules>=0.2.1->google-auth<3,>=1.6.3->tensorboard<2.13,>=2.12->tensorflow) (0.4.8)\n",
      "Requirement already satisfied: oauthlib>=3.0.0 in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (from requests-oauthlib>=0.7.0->google-auth-oauthlib<1.1,>=0.5->tensorboard<2.13,>=2.12->tensorflow) (3.2.2)\n",
      "Note: you may need to restart the kernel to use updated packages.\n",
      "Requirement already satisfied: keras in /Users/shabnam/opt/anaconda3/envs/ARN_2023/lib/python3.9/site-packages (2.12.0)\n",
      "Note: you may need to restart the kernel to use updated packages.\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "from keras.datasets import mnist\n",
    "from keras.layers.core import Dense\n",
    "from keras.models import Sequential\n",
    "from keras.optimizers import RMSprop\n",
    "from keras.utils import np_utils\n",
    "from matplotlib import pyplot as pl\n",
    "from skimage import exposure\n",
    "from skimage.feature import hog\n",
    "from sklearn import metrics as me\n",
    "\n",
    "from utils import plot_confusion_matrix\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "KpPmUqiRbafz"
   },
   "source": [
    "Load the `mnist` dataset and normalize in the range [0, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "id": "verKKFaFbaf6",
    "outputId": "29e842da-35fa-49b6-a7f7-f35d9e8a71b9"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "60000 train samples\n",
      "10000 test samples\n"
     ]
    }
   ],
   "source": [
    "(X_train, y_train), (X_test, y_test) = mnist.load_data()\n",
    "\n",
    "n_train, height, width = X_train.shape\n",
    "n_test, _, _ = X_test.shape\n",
    "\n",
    "X_train = X_train.reshape(n_train, height, width, 1).astype('float32')\n",
    "X_test = X_test.reshape(n_test, height, width, 1).astype('float32')\n",
    "\n",
    "X_train /= 255.0\n",
    "X_test /= 255.0\n",
    "\n",
    "n_classes = 10\n",
    "\n",
    "print(n_train, 'train samples')\n",
    "print(n_test, 'test samples')\n",
    "\n",
    "# convert class vectors to binary class matrices\n",
    "Y_train = np_utils.to_categorical(y_train, n_classes)\n",
    "Y_test = np_utils.to_categorical(y_test, n_classes)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "TazeB7LLbagE"
   },
   "source": [
    "### An example of how to compute the histogram of gradients\n",
    "Let's compute the HOG for one image in the test dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "id": "cL-609oKbagI"
   },
   "outputs": [],
   "source": [
    "n_orientations = 8\n",
    "pix_p_cell = 4\n",
    "hog_size = int(height * width * n_orientations / (pix_p_cell * pix_p_cell))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "id": "r2DA2Da_bage"
   },
   "outputs": [],
   "source": [
    "fd, hog_image = hog(X_test[0,:,:,0], orientations=n_orientations, pixels_per_cell=(pix_p_cell, pix_p_cell), cells_per_block=(1, 1), visualize=True)\n",
    "hog_image_rescaled = exposure.rescale_intensity(hog_image, in_range=(0, 0.02))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "s2Ow4uJtbago"
   },
   "source": [
    "Show the image and the corresponding gradients"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 351
    },
    "id": "83RZeljKbagu",
    "outputId": "e1507d91-5cd9-484c-e0bc-0673b046a274"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = pl.subplots(1, 2, figsize=(12, 6), sharex=True, sharey=True)\n",
    "ax1.axis('off')\n",
    "ax1.imshow(X_test[0,:,:,0], cmap=pl.get_cmap('Greys'), interpolation='nearest')\n",
    "ax1.set_title('Input image')\n",
    "ax1.set_adjustable('box')\n",
    "\n",
    "ax2.axis('off')\n",
    "ax2.imshow(hog_image_rescaled, cmap=pl.get_cmap('Greys'), interpolation='nearest')\n",
    "ax2.set_title('Histogram of Oriented Gradients')\n",
    "ax1.set_adjustable('box')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "3Yv-ep9ybag4"
   },
   "source": [
    "Compute the HOGs for the train and test datasets\n",
    "\n",
    "> Indented block\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "id": "y8Q3cdrebag6",
    "outputId": "4c21e6ce-56d2-4287-eab6-634cda1ad94b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X_train done\n",
      "X_test done\n"
     ]
    }
   ],
   "source": [
    "X_train_hog = np.zeros((X_train.shape[0], hog_size))\n",
    "X_test_hog = np.zeros((X_test.shape[0], hog_size))\n",
    "\n",
    "for i in np.arange(X_train_hog.shape[0]):\n",
    "    X_train_hog[i,:] = hog(X_train[i,:,:,0], orientations=n_orientations, pixels_per_cell=(pix_p_cell, pix_p_cell), cells_per_block=(1, 1), visualize=False)\n",
    "print('X_train done')\n",
    "\n",
    "for i in np.arange(X_test_hog.shape[0]):\n",
    "    X_test_hog[i,:] = hog(X_test[i,:,:,0], orientations=n_orientations, pixels_per_cell=(pix_p_cell, pix_p_cell), cells_per_block=(1, 1), visualize=False)\n",
    "print('X_test done')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "kos0aTYmbahU"
   },
   "source": [
    "Create the MLP"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 221
    },
    "id": "NnawtqztbahV",
    "outputId": "c99078e7-4754-41c7-dbe9-e953931c5f6c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " dense (Dense)               (None, 200)               78600     \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 10)                2010      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 80,610\n",
      "Trainable params: 80,610\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model = Sequential()\n",
    "model.add(Dense(200, input_shape=(hog_size,), activation='relu'))\n",
    "#model.add(Dropout(0.5))\n",
    "model.add(Dense(n_classes, activation='softmax'))\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "EGeviXzJbahY"
   },
   "source": [
    "Define some constants and train the MLP"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 374
    },
    "id": "laKFY8vcbahZ",
    "outputId": "5849045b-3563-4afb-edfa-5658aac89ec8"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "422/422 [==============================] - 1s 2ms/step - loss: 0.2383 - accuracy: 0.9436 - val_loss: 0.0868 - val_accuracy: 0.9743\n",
      "Epoch 2/10\n",
      "422/422 [==============================] - 1s 2ms/step - loss: 0.0942 - accuracy: 0.9703 - val_loss: 0.0754 - val_accuracy: 0.9755\n",
      "Epoch 3/10\n",
      "422/422 [==============================] - 1s 1ms/step - loss: 0.0781 - accuracy: 0.9749 - val_loss: 0.0705 - val_accuracy: 0.9778\n",
      "Epoch 4/10\n",
      "422/422 [==============================] - 1s 2ms/step - loss: 0.0684 - accuracy: 0.9778 - val_loss: 0.0647 - val_accuracy: 0.9793\n",
      "Epoch 5/10\n",
      "422/422 [==============================] - 1s 1ms/step - loss: 0.0600 - accuracy: 0.9808 - val_loss: 0.0655 - val_accuracy: 0.9792\n",
      "Epoch 6/10\n",
      "422/422 [==============================] - 1s 1ms/step - loss: 0.0531 - accuracy: 0.9832 - val_loss: 0.0637 - val_accuracy: 0.9795\n",
      "Epoch 7/10\n",
      "422/422 [==============================] - 1s 2ms/step - loss: 0.0472 - accuracy: 0.9854 - val_loss: 0.0656 - val_accuracy: 0.9797\n",
      "Epoch 8/10\n",
      "422/422 [==============================] - 1s 2ms/step - loss: 0.0417 - accuracy: 0.9869 - val_loss: 0.0648 - val_accuracy: 0.9807\n",
      "Epoch 9/10\n",
      "422/422 [==============================] - 1s 2ms/step - loss: 0.0370 - accuracy: 0.9886 - val_loss: 0.0582 - val_accuracy: 0.9823\n",
      "Epoch 10/10\n",
      "422/422 [==============================] - 1s 2ms/step - loss: 0.0325 - accuracy: 0.9903 - val_loss: 0.0620 - val_accuracy: 0.9820\n"
     ]
    }
   ],
   "source": [
    "batch_size = 128\n",
    "n_epoch = 10\n",
    "\n",
    "model.compile(loss='categorical_crossentropy', optimizer=RMSprop(), metrics=['accuracy'])\n",
    "\n",
    "history = model.fit(X_train_hog, Y_train,\n",
    "                    batch_size=batch_size, epochs=n_epoch,\n",
    "                    verbose=1, validation_split=0.1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "aezTZuCMbahf"
   },
   "source": [
    "Show the performance of the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 299
    },
    "id": "u6E66I52bahg",
    "outputId": "52139b93-823a-4fca-de59-a64f77f3b0ab"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test score: 0.06543819606304169\n",
      "Test accuracy: 0.9800000190734863\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pl.plot(history.history['loss'], label='Training')\n",
    "pl.plot(history.history['val_loss'], label='Testing')\n",
    "pl.legend()\n",
    "pl.grid()\n",
    "\n",
    "score = model.evaluate(X_test_hog, Y_test, verbose=0)\n",
    "print('Test score:', score[0])\n",
    "print('Test accuracy:', score[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "axdxLHYMbahu"
   },
   "source": [
    "Confusion matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 187
    },
    "id": "Mhm1wjNfbahv",
    "outputId": "8366db2a-2162-4c03-d1d3-de0a4114c3cb"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 0s 653us/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[ 969,    0,    1,    0,    0,    2,    5,    1,    1,    1],\n",
       "       [   1, 1128,    1,    1,    0,    2,    1,    0,    1,    0],\n",
       "       [   2,    5, 1010,    1,    3,    0,    1,    3,    7,    0],\n",
       "       [   0,    1,    5,  968,    0,   22,    1,    4,    9,    0],\n",
       "       [   0,    3,    2,    0,  969,    0,    0,    1,    3,    4],\n",
       "       [   1,    1,    0,    4,    0,  882,    3,    0,    1,    0],\n",
       "       [   2,    2,    1,    0,    3,    2,  948,    0,    0,    0],\n",
       "       [   0,    5,    7,    2,    4,    1,    0,  995,    5,    9],\n",
       "       [   4,    0,    0,    2,    1,    1,    2,    5,  954,    5],\n",
       "       [   0,    5,    1,    4,    7,    6,    0,    4,    5,  977]])"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred = np.argmax(model.predict(X_test_hog), axis=-1)\n",
    "cm = me.confusion_matrix(y_test, pred)\n",
    "\n",
    "plot_confusion_matrix(cm)"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "MLP_from_HOG.ipynb",
   "provenance": []
  },
  "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": 1
}