Model_5_New_Version.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Model 5 New Version.ipynb",
      "version": "0.3.2",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "xOI82soC8PxJ",
        "colab_type": "code",
        "outputId": "6a87761c-54be-40ea-ecf1-e6b003600c0f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 399
        }
      },
      "source": [
        "!pip install tf-nightly-gpu-2.0-preview\n",
        "import tensorflow as tf\n",
        "device_name = tf.test.gpu_device_name()\n",
        "if device_name != '/device:GPU:0':\n",
        "  raise SystemError('GPU device not found')\n",
        "print('Found GPU at: {}'.format(device_name))\n",
        "import os\n",
        "import matplotlib.pyplot as plt\n",
        "import cv2\n",
        "import numpy as np\n",
        "import random\n",
        "import pickle\n",
        "from google.colab.patches import cv2_imshow\n",
        "from keras.preprocessing.image import ImageDataGenerator\n"
      ],
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: tf-nightly-gpu-2.0-preview in /usr/local/lib/python3.6/dist-packages (2.0.0.dev20190605)\n",
            "Requirement already satisfied: tensorflow-estimator-2.0-preview in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview) (1.14.0.dev2019060500)\n",
            "Requirement already satisfied: protobuf>=3.6.1 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview) (3.7.1)\n",
            "Requirement already satisfied: gast>=0.2.0 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview) (0.2.2)\n",
            "Requirement already satisfied: keras-preprocessing>=1.0.5 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview) (1.0.9)\n",
            "Requirement already satisfied: numpy<2.0,>=1.14.5 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview) (1.16.4)\n",
            "Requirement already satisfied: six>=1.10.0 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview) (1.12.0)\n",
            "Requirement already satisfied: absl-py>=0.7.0 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview) (0.7.1)\n",
            "Requirement already satisfied: wrapt>=1.11.1 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview) (1.11.1)\n",
            "Requirement already satisfied: tb-nightly<1.15.0a0,>=1.14.0a0 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview) (1.14.0a20190605)\n",
            "Requirement already satisfied: wheel>=0.26 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview) (0.33.4)\n",
            "Requirement already satisfied: grpcio>=1.8.6 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview) (1.15.0)\n",
            "Requirement already satisfied: astor>=0.6.0 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview) (0.8.0)\n",
            "Requirement already satisfied: termcolor>=1.1.0 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview) (1.1.0)\n",
            "Requirement already satisfied: keras-applications>=1.0.6 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview) (1.0.7)\n",
            "Requirement already satisfied: google-pasta>=0.1.6 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview) (0.1.7)\n",
            "Requirement already satisfied: setuptools in /usr/local/lib/python3.6/dist-packages (from protobuf>=3.6.1->tf-nightly-gpu-2.0-preview) (41.0.1)\n",
            "Requirement already satisfied: werkzeug>=0.11.15 in /usr/local/lib/python3.6/dist-packages (from tb-nightly<1.15.0a0,>=1.14.0a0->tf-nightly-gpu-2.0-preview) (0.15.4)\n",
            "Requirement already satisfied: markdown>=2.6.8 in /usr/local/lib/python3.6/dist-packages (from tb-nightly<1.15.0a0,>=1.14.0a0->tf-nightly-gpu-2.0-preview) (3.1.1)\n",
            "Requirement already satisfied: h5py in /usr/local/lib/python3.6/dist-packages (from keras-applications>=1.0.6->tf-nightly-gpu-2.0-preview) (2.8.0)\n",
            "Found GPU at: /device:GPU:0\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "Using TensorFlow backend.\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cqy1RBAR9VXy",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#the follow are steps to load data from Google Drive\n",
        "from google.colab import drive"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "a7uzoKzp9i86",
        "colab_type": "code",
        "outputId": "09182868-898c-4fbc-d359-43c0dbc91525",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "drive.mount('/content/gdrive')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Drive already mounted at /content/gdrive; to attempt to forcibly remount, call drive.mount(\"/content/gdrive\", force_remount=True).\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hkTZHq_vCN6L",
        "colab_type": "code",
        "outputId": "de37c9b7-e5e2-4227-d86f-251e97202e1c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "!ls './gdrive/My Drive/Career Work/SERIUS/Pill Images V2' #we can use ! to run terminal commands"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "'Test Set'  'Train Set'\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lCmC5brA-hJO",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "DATA_PATH_TRAIN='./gdrive/My Drive/Career Work/SERIUS/Pill Images V2/Train Set'\n",
        "DATA_PATH_TEST='./gdrive/My Drive/Career Work/SERIUS/Pill Images V2/Test Set'\n",
        "medication_array=['Pill1','Pill2','Pill3','Pill4','Pill5','Pill6']\n",
        "colour_array=['Red','Orange','White','Black','Dark Red','Blue']\n",
        "shape=['Circular','Elliptical']\n",
        "IMG_SIZE=150\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bS9deh33AFrc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def process_image(img_array,IMG_SIZE): \n",
        "  crop_img = img_array[250:750, 700:1200] #first crop the image\n",
        "  \n",
        "  #next find the contours of the image\n",
        "  new_array=cv2.resize(crop_img,(IMG_SIZE,IMG_SIZE)) #resize every image before passing it in\n",
        "  hsv = cv2.cvtColor(new_array, cv2.COLOR_BGR2HSV) #convert image to hsv\n",
        "\n",
        "  ret, thresh = cv2.threshold(hsv, 127, 255, 0) #threshold our hsv\n",
        "\n",
        "  #im2, contours, hierarchy = cv2.findContours(thresh, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)\n",
        "  #cv2.drawContours(img, contours, -1, (0,255,0), 3)\n",
        "  plt.imshow(hsv,cmap='Greys_r')\n",
        "  #plt.show()    \n",
        "  return hsv"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "outputId": "9826a792-b529-4b18-9ecd-ef985067156c",
        "id": "rs79iZeMjYLa",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 473
        }
      },
      "source": [
        "medication_array=['Pill1','Pill2','Pill3','Pill4','Pill5','Pill6']\n",
        "colour_array=['Red','Orange','White','Black','Dark Red','Blue']\n",
        "shape=['Circular','Elliptical']\n",
        "\n",
        "class dataset:\n",
        "  #in python, don't define any variables here, another it doesn't not belong to the object itself and will get mutated everything another class is called\n",
        "  \n",
        "  def __init__(self,path):\n",
        "    self.PATH=path\n",
        "    self.x=[]\n",
        "    self.y=[]\n",
        "    self.shapes=[]\n",
        "    self.colours=[]\n",
        "    self.data=[] #entire dataset\n",
        "    \n",
        "    \n",
        "  def create_data(self):\n",
        "    for pill in medication_array:\n",
        "      #local variables\n",
        "      print(pill)\n",
        "      my_path=os.path.join(self.PATH,pill)\n",
        "      classnum=medication_array.index(pill) #indexed from 0 to 5 (will add one when showing results)\n",
        "      pill_shape=shape[1]\n",
        "      \n",
        "      if classnum==0 or classnum==1 or classnum==3:\n",
        "        pill_shape=shape[0]\n",
        "        \n",
        "      for image in os.listdir(my_path):\n",
        "        try:\n",
        "          img_array=cv2.imread(os.path.join(my_path,image))\n",
        "          new_array=process_image(img_array,IMG_SIZE)\n",
        "          self.data.append([new_array,classnum,pill_shape,colour_array[classnum]])\n",
        "        except Exception as e:\n",
        "          print('f')\n",
        "          pass\n",
        "    random.shuffle(self.data)\n",
        "    \n",
        "  def structure_data(self):\n",
        "    for image, label, pill_shape, colour in self.data:\n",
        "      self.x.append(image)\n",
        "      self.y.append(label)\n",
        "      self.shapes.append(pill_shape)\n",
        "      self.colours.append(colour)\n",
        "    self.x=np.array(self.x).reshape(-1,IMG_SIZE,IMG_SIZE,3)\n",
        "          \n",
        "\n",
        "training=dataset(DATA_PATH_TRAIN)\n",
        "training.create_data()\n",
        "training.structure_data()\n",
        "\n",
        "testing=dataset(DATA_PATH_TEST)\n",
        "testing.create_data()\n",
        "testing.structure_data()\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Pill1\n",
            "Pill2\n",
            "Pill3\n",
            "Pill4\n",
            "Pill5\n",
            "Pill6\n",
            "Pill1\n",
            "Pill2\n",
            "Pill3\n",
            "Pill4\n",
            "Pill5\n",
            "Pill6\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAQUAAAD8CAYAAAB+fLH0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nBmEKQRA0CFMIgqDBD5ap+L+yK4J7AAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RGS72fc5Usea",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "checkpoint_path = \"training_1/cp.ckpt\"\n",
        "checkpoint_dir = os.path.dirname(checkpoint_path)\n",
        "\n",
        "cp_callback = tf.keras.callbacks.ModelCheckpoint(checkpoint_path, \n",
        "                                                 save_weights_only=True,\n",
        "                                                 verbose=1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7qqHhDwgIqgY",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def show_RBG(image_in):\n",
        "  bgrimg = cv2.cvtColor(image_in, cv2.COLOR_HSV2BGR)\n",
        "  image = cv2.cvtColor(bgrimg, cv2.COLOR_BGR2RGB)\n",
        "  pixels = np.array(image)\n",
        "  plt.imshow(pixels)\n",
        "  plt.show()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2jObtCQDEjLL",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "pickle_out=open('train_x.pickle','wb')\n",
        "pickle.dump(training.x,pickle_out)\n",
        "pickle_out.close();\n",
        "\n",
        "pickle_out=open('train_y.pickle','wb')\n",
        "pickle.dump(training.y,pickle_out)\n",
        "pickle_out.close();\n",
        "\n",
        "pickle_out=open('test_x.pickle','wb')\n",
        "pickle.dump(testing.x,pickle_out)\n",
        "pickle_out.close();\n",
        "\n",
        "pickle_out=open('test_y.pickle','wb')\n",
        "pickle.dump(testing.y,pickle_out)\n",
        "pickle_out.close();\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Up9ZJBG8M80o",
        "colab_type": "code",
        "outputId": "3f37cc4e-4367-4f7b-a704-774c90179a0c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "source": [
        "test_x=[]\n",
        "train_x=[]\n",
        "test_y=[]\n",
        "train_y=[]\n",
        "\n",
        "pickle_in=open('test_x.pickle','rb')\n",
        "test_x=pickle.load(pickle_in)\n",
        "\n",
        "pickle_in=open('test_y.pickle','rb')\n",
        "test_y=pickle.load(pickle_in)\n",
        "\n",
        "\n",
        "pickle_in=open('train_x.pickle','rb')\n",
        "train_x=pickle.load(pickle_in)\n",
        "\n",
        "pickle_in=open('train_y.pickle','rb')\n",
        "train_y=pickle.load(pickle_in)\n",
        "print(len(test_y))\n",
        "print(test_x.shape)\n",
        "test_x=test_x/255.0\n",
        "train_x=train_x/255.0\n"
      ],
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "242\n",
            "(242, 150, 150, 3)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "y2gSrOOVFN_R",
        "colab_type": "code",
        "outputId": "d1ff666d-2245-4fd5-bf1d-794f8445937e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 469
        }
      },
      "source": [
        "def createModel():\n",
        "    model = tf.keras.models.Sequential([\n",
        "        tf.keras.layers.Conv2D(16,(3,3),activation='relu',input_shape=train_x.shape[1:]),\n",
        "        #input shape is IMG_SIZExIMG_SIZEx3bytes coloured (HSV))\n",
        "        tf.keras.layers.MaxPooling2D(2,2),\n",
        "        tf.keras.layers.Conv2D(32,(3,3),activation='relu'),\n",
        "        tf.keras.layers.MaxPooling2D(2,2),\n",
        "        tf.keras.layers.Conv2D(64,(3,3),activation='relu'),\n",
        "        tf.keras.layers.MaxPooling2D(2,2),\n",
        "        tf.keras.layers.Flatten(), #flattens the dataset for our input layer\n",
        "        tf.keras.layers.Dense(512,activation=tf.nn.relu),\n",
        "        tf.keras.layers.Dense(6,activation='softmax') #sigmoid for final layer\n",
        "    ])\n",
        "    model.compile(optimizer='adam',\n",
        "              loss='sparse_categorical_crossentropy',\n",
        "              metrics=['accuracy'])\n",
        "    return model\n",
        "model=createModel()\n",
        "model.summary()"
      ],
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_10\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d_30 (Conv2D)           (None, 148, 148, 16)      448       \n",
            "_________________________________________________________________\n",
            "max_pooling2d_30 (MaxPooling (None, 74, 74, 16)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_31 (Conv2D)           (None, 72, 72, 32)        4640      \n",
            "_________________________________________________________________\n",
            "max_pooling2d_31 (MaxPooling (None, 36, 36, 32)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_32 (Conv2D)           (None, 34, 34, 64)        18496     \n",
            "_________________________________________________________________\n",
            "max_pooling2d_32 (MaxPooling (None, 17, 17, 64)        0         \n",
            "_________________________________________________________________\n",
            "flatten_10 (Flatten)         (None, 18496)             0         \n",
            "_________________________________________________________________\n",
            "dense_20 (Dense)             (None, 512)               9470464   \n",
            "_________________________________________________________________\n",
            "dense_21 (Dense)             (None, 6)                 3078      \n",
            "=================================================================\n",
            "Total params: 9,497,126\n",
            "Trainable params: 9,497,126\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6ovZrHUoG2ag",
        "colab_type": "text"
      },
      "source": [
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bgSMK52PHn0U",
        "colab_type": "code",
        "outputId": "7c021d94-122e-45e9-ce74-051d300dd5d8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 399
        }
      },
      "source": [
        "\n",
        "model.fit(train_x,train_y, validation_split=0.2, epochs=5, batch_size=30, callbacks = [cp_callback])"
      ],
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 767 samples, validate on 192 samples\n",
            "Epoch 1/5\n",
            "750/767 [============================>.] - ETA: 0s - loss: 1.3453 - accuracy: 0.4973\n",
            "Epoch 00001: saving model to training_1/cp.ckpt\n",
            "767/767 [==============================] - 1s 2ms/sample - loss: 1.3214 - accuracy: 0.5072 - val_loss: 0.3047 - val_accuracy: 0.9427\n",
            "Epoch 2/5\n",
            "720/767 [===========================>..] - ETA: 0s - loss: 0.1957 - accuracy: 0.9556\n",
            "Epoch 00002: saving model to training_1/cp.ckpt\n",
            "767/767 [==============================] - 1s 1ms/sample - loss: 0.1903 - accuracy: 0.9570 - val_loss: 0.0710 - val_accuracy: 0.9844\n",
            "Epoch 3/5\n",
            "750/767 [============================>.] - ETA: 0s - loss: 0.0254 - accuracy: 0.9973\n",
            "Epoch 00003: saving model to training_1/cp.ckpt\n",
            "767/767 [==============================] - 1s 1ms/sample - loss: 0.0253 - accuracy: 0.9974 - val_loss: 0.0355 - val_accuracy: 0.9948\n",
            "Epoch 4/5\n",
            "750/767 [============================>.] - ETA: 0s - loss: 0.0032 - accuracy: 1.0000\n",
            "Epoch 00004: saving model to training_1/cp.ckpt\n",
            "767/767 [==============================] - 1s 1ms/sample - loss: 0.0032 - accuracy: 1.0000 - val_loss: 0.0012 - val_accuracy: 1.0000\n",
            "Epoch 5/5\n",
            "720/767 [===========================>..] - ETA: 0s - loss: 4.7766e-04 - accuracy: 1.0000\n",
            "Epoch 00005: saving model to training_1/cp.ckpt\n",
            "767/767 [==============================] - 1s 2ms/sample - loss: 4.5718e-04 - accuracy: 1.0000 - val_loss: 7.7995e-04 - val_accuracy: 1.0000\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.keras.callbacks.History at 0x7f22b8340748>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 47
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DNruJ4cIA8Qw",
        "colab_type": "code",
        "outputId": "d12abbd6-20f3-428a-9b9c-ccb3393f118c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 69
        }
      },
      "source": [
        "\n",
        "newModel=createModel()\n",
        "newModel.evaluate(test_x,test_y)\n",
        "newModel.load_weights(checkpoint_path)\n",
        "newModel.evaluate(test_x,test_y)"
      ],
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "242/242 [==============================] - 0s 717us/sample - loss: 1.7967 - accuracy: 0.1570\n",
            "242/242 [==============================] - 0s 517us/sample - loss: 0.0017 - accuracy: 1.0000\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.0017480696472728986, 1.0]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 48
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gM4a4DB_QUjX",
        "colab_type": "code",
        "outputId": "d977a800-dd2b-422a-99cd-c71b9b508946",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "ypr = model.predict(test_x)\n",
        "model.evaluate(test_x,test_y)\n",
        "model.save('Pill Predictor V5')"
      ],
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "242/242 [==============================] - 0s 531us/sample - loss: 0.0017 - accuracy: 1.0000\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5Txk0unKQwLb",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def prepare(filepath):\n",
        "    IMG_SIZE=150\n",
        "    img_array=cv2.imread(filepath)\n",
        "    new_array=process_image(img_array)\n",
        "    return new_array.reshape(-1,IMG_SIZE,IMG_SIZE,1) "
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3uzDE2wnS_my",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.metrics import confusion_matrix"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9MQxwlOxTKtU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "yprr = np.argmax(ypr, axis=1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "A7C2UgrqTOS1",
        "colab_type": "code",
        "outputId": "f10d3ce9-19e1-4678-daed-632ff2ac3737",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 121
        }
      },
      "source": [
        "confusion_matrix(test_y, yprr)"
      ],
      "execution_count": 53,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[40,  0,  0,  0,  0,  0],\n",
              "       [ 0, 41,  0,  0,  0,  0],\n",
              "       [ 0,  0, 41,  0,  0,  0],\n",
              "       [ 0,  0,  0, 38,  0,  0],\n",
              "       [ 0,  0,  0,  0, 41,  0],\n",
              "       [ 0,  0,  0,  0,  0, 41]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 53
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fItoJtDWTXaj",
        "colab_type": "text"
      },
      "source": [
        ""
      ]
    }
  ]
}