Simple_QCNN_MNIST.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 150307,
     "status": "ok",
     "timestamp": 1594969572300,
     "user": {
      "displayName": "오승혁",
      "photoUrl": "",
      "userId": "17576348803238599032"
     },
     "user_tz": -540
    },
    "id": "v443mtS5EPCl"
   },
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "import tensorflow_quantum as tfq\n",
    "\n",
    "from tensorflow.keras import datasets, layers, models\n",
    "\n",
    "import cirq\n",
    "import sympy\n",
    "import numpy as np\n",
    "import collections\n",
    "\n",
    "# visualization tools\n",
    "%matplotlib inline\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt\n",
    "from cirq.contrib.svg import SVGCircuit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 89
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 150776,
     "status": "ok",
     "timestamp": 1594969572784,
     "user": {
      "displayName": "오승혁",
      "photoUrl": "",
      "userId": "17576348803238599032"
     },
     "user_tz": -540
    },
    "id": "73KuuBfoEPCp",
    "outputId": "6bd9a172-b35d-4489-ebec-ae6725324686"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of original training examples: 60000\n",
      "Number of original test examples: 10000\n"
     ]
    }
   ],
   "source": [
    "(x_train, y_train), (x_test, y_test) = tf.keras.datasets.mnist.load_data()\n",
    "\n",
    "# Rescale the images from [0,255] to the [0.0,1.0] range.\n",
    "x_train, x_test = x_train[..., np.newaxis]/255.0, x_test[..., np.newaxis]/255.0\n",
    "\n",
    "print(\"Number of original training examples:\", len(x_train))\n",
    "print(\"Number of original test examples:\", len(x_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 3294,
     "status": "ok",
     "timestamp": 1594969576122,
     "user": {
      "displayName": "오승혁",
      "photoUrl": "",
      "userId": "17576348803238599032"
     },
     "user_tz": -540
    },
    "id": "Hupa2OMlEPCs"
   },
   "outputs": [],
   "source": [
    "x_train= tf.cast(x_train, tf.float32)\n",
    "x_test=tf.cast(x_test, tf.float32)\n",
    "\n",
    "x_train = tf.image.resize(x_train[:], (10,10)).numpy()\n",
    "x_test = tf.image.resize(x_test[:], (10,10)).numpy()\n",
    "\n",
    "y_train = y_train[:]\n",
    "y_test = y_test[:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 301
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 3232,
     "status": "ok",
     "timestamp": 1594969576129,
     "user": {
      "displayName": "오승혁",
      "photoUrl": "",
      "userId": "17576348803238599032"
     },
     "user_tz": -540
    },
    "id": "jE3niAxiEPCu",
    "outputId": "e310e5f3-74db-4080-fd3a-7ed42e619197"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "print(y_train[0])\n",
    "fig = plt.figure()\n",
    "plt.imshow(x_train[0, :, :, 0], cmap='gray_r')\n",
    "plt.colorbar()\n",
    "fig.savefig('sample', dpi=300)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 3186,
     "status": "ok",
     "timestamp": 1594969576130,
     "user": {
      "displayName": "오승혁",
      "photoUrl": "",
      "userId": "17576348803238599032"
     },
     "user_tz": -540
    },
    "id": "1TOeQilXEPCw",
    "outputId": "52a80e5b-2942-4b98-ce5f-bfc6a2b7f6dc"
   },
   "source": [
    "# Fully Connected Layer Only"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "width = np.shape(x_train)[1]\n",
    "height = np.shape(x_train)[2]\n",
    "\n",
    "fc_model = models.Sequential()\n",
    "\n",
    "fc_model.add(layers.Flatten(input_shape=(width,height,1)))\n",
    "fc_model.add(layers.Dense(32, activation='relu'))\n",
    "fc_model.add(layers.Dense(10, activation='softmax'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 1.6229 - accuracy: 0.5556 - val_loss: 1.0329 - val_accuracy: 0.7449\n",
      "Epoch 2/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.8126 - accuracy: 0.7936 - val_loss: 0.6540 - val_accuracy: 0.8314\n",
      "Epoch 3/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.6049 - accuracy: 0.8332 - val_loss: 0.5307 - val_accuracy: 0.8525\n",
      "Epoch 4/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.5417 - accuracy: 0.8492 - val_loss: 0.4667 - val_accuracy: 0.8706\n",
      "Epoch 5/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.4684 - accuracy: 0.8680 - val_loss: 0.4353 - val_accuracy: 0.8770\n",
      "Epoch 6/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.4218 - accuracy: 0.8824 - val_loss: 0.4007 - val_accuracy: 0.8856\n",
      "Epoch 7/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.4353 - accuracy: 0.8728 - val_loss: 0.3942 - val_accuracy: 0.8834\n",
      "Epoch 8/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.3901 - accuracy: 0.8848 - val_loss: 0.3647 - val_accuracy: 0.8976\n",
      "Epoch 9/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.3533 - accuracy: 0.9000 - val_loss: 0.3552 - val_accuracy: 0.8963\n",
      "Epoch 10/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.3813 - accuracy: 0.8848 - val_loss: 0.3416 - val_accuracy: 0.9009\n",
      "Epoch 11/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.3392 - accuracy: 0.8936 - val_loss: 0.3345 - val_accuracy: 0.9021\n",
      "Epoch 12/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.3724 - accuracy: 0.8892 - val_loss: 0.3293 - val_accuracy: 0.9027\n",
      "Epoch 13/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.3612 - accuracy: 0.8936 - val_loss: 0.3235 - val_accuracy: 0.9055\n",
      "Epoch 14/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.3608 - accuracy: 0.8904 - val_loss: 0.3117 - val_accuracy: 0.9120\n",
      "Epoch 15/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.3532 - accuracy: 0.8900 - val_loss: 0.3031 - val_accuracy: 0.9139\n",
      "Epoch 16/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.3719 - accuracy: 0.8904 - val_loss: 0.3067 - val_accuracy: 0.9107\n",
      "Epoch 17/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.3113 - accuracy: 0.9036 - val_loss: 0.2952 - val_accuracy: 0.9163\n",
      "Epoch 18/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.3093 - accuracy: 0.9048 - val_loss: 0.2912 - val_accuracy: 0.9152\n",
      "Epoch 19/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.3047 - accuracy: 0.9156 - val_loss: 0.2879 - val_accuracy: 0.9166\n",
      "Epoch 20/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.3081 - accuracy: 0.9136 - val_loss: 0.2876 - val_accuracy: 0.9175\n",
      "Epoch 21/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2895 - accuracy: 0.9112 - val_loss: 0.2817 - val_accuracy: 0.9189\n",
      "Epoch 22/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2853 - accuracy: 0.9068 - val_loss: 0.2776 - val_accuracy: 0.9202\n",
      "Epoch 23/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2875 - accuracy: 0.9100 - val_loss: 0.2859 - val_accuracy: 0.9146\n",
      "Epoch 24/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.3073 - accuracy: 0.9064 - val_loss: 0.2733 - val_accuracy: 0.9194\n",
      "Epoch 25/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2746 - accuracy: 0.9180 - val_loss: 0.2710 - val_accuracy: 0.9195\n",
      "Epoch 26/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.3064 - accuracy: 0.9012 - val_loss: 0.2763 - val_accuracy: 0.9170\n",
      "Epoch 27/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2763 - accuracy: 0.9180 - val_loss: 0.2538 - val_accuracy: 0.9247\n",
      "Epoch 28/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2487 - accuracy: 0.9280 - val_loss: 0.2563 - val_accuracy: 0.9248\n",
      "Epoch 29/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.3002 - accuracy: 0.9088 - val_loss: 0.2509 - val_accuracy: 0.9274\n",
      "Epoch 30/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2676 - accuracy: 0.9232 - val_loss: 0.2468 - val_accuracy: 0.9296\n",
      "Epoch 31/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.2927 - accuracy: 0.9124 - val_loss: 0.2462 - val_accuracy: 0.9268\n",
      "Epoch 32/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2312 - accuracy: 0.9316 - val_loss: 0.2439 - val_accuracy: 0.9275\n",
      "Epoch 33/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2593 - accuracy: 0.9192 - val_loss: 0.2391 - val_accuracy: 0.9306\n",
      "Epoch 34/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2526 - accuracy: 0.9236 - val_loss: 0.2391 - val_accuracy: 0.9303\n",
      "Epoch 35/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2751 - accuracy: 0.9184 - val_loss: 0.2414 - val_accuracy: 0.9290\n",
      "Epoch 36/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2389 - accuracy: 0.9280 - val_loss: 0.2393 - val_accuracy: 0.9301\n",
      "Epoch 37/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2469 - accuracy: 0.9260 - val_loss: 0.2338 - val_accuracy: 0.9321\n",
      "Epoch 38/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2522 - accuracy: 0.9264 - val_loss: 0.2360 - val_accuracy: 0.9321\n",
      "Epoch 39/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2489 - accuracy: 0.9320 - val_loss: 0.2270 - val_accuracy: 0.9327\n",
      "Epoch 40/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2571 - accuracy: 0.9240 - val_loss: 0.2324 - val_accuracy: 0.9316\n",
      "Epoch 41/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2606 - accuracy: 0.9204 - val_loss: 0.2310 - val_accuracy: 0.9334\n",
      "Epoch 42/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2398 - accuracy: 0.9268 - val_loss: 0.2210 - val_accuracy: 0.9377\n",
      "Epoch 43/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2410 - accuracy: 0.9248 - val_loss: 0.2209 - val_accuracy: 0.9364\n",
      "Epoch 44/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2467 - accuracy: 0.9268 - val_loss: 0.2214 - val_accuracy: 0.9368\n",
      "Epoch 45/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2273 - accuracy: 0.9304 - val_loss: 0.2258 - val_accuracy: 0.9348\n",
      "Epoch 46/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2157 - accuracy: 0.9340 - val_loss: 0.2171 - val_accuracy: 0.9383\n",
      "Epoch 47/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2384 - accuracy: 0.9248 - val_loss: 0.2208 - val_accuracy: 0.9382\n",
      "Epoch 48/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2518 - accuracy: 0.9212 - val_loss: 0.2162 - val_accuracy: 0.9381\n",
      "Epoch 49/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.2497 - accuracy: 0.9292 - val_loss: 0.2101 - val_accuracy: 0.9408\n",
      "Epoch 50/50\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2151 - accuracy: 0.9352 - val_loss: 0.2162 - val_accuracy: 0.9391\n"
     ]
    }
   ],
   "source": [
    "fc_model.compile(optimizer='adam',\n",
    "              loss='sparse_categorical_crossentropy',\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "fc_history = fc_model.fit(x_train, y_train, steps_per_epoch=500,\n",
    "                        validation_data=(x_test, y_test), \n",
    "                        epochs=50, batch_size=5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Convolutional Layer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1200,
     "status": "ok",
     "timestamp": 1594971558074,
     "user": {
      "displayName": "오승혁",
      "photoUrl": "",
      "userId": "17576348803238599032"
     },
     "user_tz": -540
    },
    "id": "eMVX5OOQEPCy"
   },
   "outputs": [],
   "source": [
    "width = np.shape(x_train)[1]\n",
    "height = np.shape(x_train)[2]\n",
    "\n",
    "cnn_model = models.Sequential()\n",
    "\n",
    "cnn_model.add(layers.Conv2D(8, (2, 2), activation='relu', input_shape=(width, height, 1)))\n",
    "#model.add(layers.MaxPooling2D((2, 2)))\n",
    "#model.add(layers.Conv2D(16, (2, 2), activation='relu'))\n",
    "\n",
    "cnn_model.add(layers.Flatten())\n",
    "cnn_model.add(layers.Dense(32, activation='relu'))\n",
    "cnn_model.add(layers.Dense(10, activation='softmax'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 305
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 917,
     "status": "ok",
     "timestamp": 1594971558076,
     "user": {
      "displayName": "오승혁",
      "photoUrl": "",
      "userId": "17576348803238599032"
     },
     "user_tz": -540
    },
    "id": "4YieE0zaEPC0",
    "outputId": "8ac1e269-2e43-458b-ccd8-63f0c632f66d"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d (Conv2D)              (None, 9, 9, 8)           40        \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 648)               0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 32)                20768     \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 10)                330       \n",
      "=================================================================\n",
      "Total params: 21,138\n",
      "Trainable params: 21,138\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "cnn_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 61419,
     "status": "ok",
     "timestamp": 1594971626644,
     "user": {
      "displayName": "오승혁",
      "photoUrl": "",
      "userId": "17576348803238599032"
     },
     "user_tz": -540
    },
    "id": "cnGpap70EPC2",
    "outputId": "4f21ca01-5430-4504-9d4c-fb24df3c6e99",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "500/500 [==============================] - 3s 5ms/step - loss: 1.0500 - accuracy: 0.7032 - val_loss: 0.5254 - val_accuracy: 0.8512\n",
      "Epoch 2/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.5107 - accuracy: 0.8492 - val_loss: 0.4711 - val_accuracy: 0.8539\n",
      "Epoch 3/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.4478 - accuracy: 0.8644 - val_loss: 0.3784 - val_accuracy: 0.8831\n",
      "Epoch 4/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.3858 - accuracy: 0.8872 - val_loss: 0.3595 - val_accuracy: 0.8920\n",
      "Epoch 5/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.3865 - accuracy: 0.8852 - val_loss: 0.3172 - val_accuracy: 0.9053\n",
      "Epoch 6/50\n",
      "500/500 [==============================] - 3s 5ms/step - loss: 0.3391 - accuracy: 0.8960 - val_loss: 0.3271 - val_accuracy: 0.8981\n",
      "Epoch 7/50\n",
      "500/500 [==============================] - 3s 5ms/step - loss: 0.3564 - accuracy: 0.8816 - val_loss: 0.2730 - val_accuracy: 0.9208\n",
      "Epoch 8/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.2985 - accuracy: 0.9128 - val_loss: 0.2728 - val_accuracy: 0.9199\n",
      "Epoch 9/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.2964 - accuracy: 0.9080 - val_loss: 0.2686 - val_accuracy: 0.9198\n",
      "Epoch 10/50\n",
      "500/500 [==============================] - 3s 5ms/step - loss: 0.2950 - accuracy: 0.9156 - val_loss: 0.2483 - val_accuracy: 0.9270\n",
      "Epoch 11/50\n",
      "500/500 [==============================] - 3s 5ms/step - loss: 0.2869 - accuracy: 0.9100 - val_loss: 0.2217 - val_accuracy: 0.9377\n",
      "Epoch 12/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.2376 - accuracy: 0.9264 - val_loss: 0.2334 - val_accuracy: 0.9299\n",
      "Epoch 13/50\n",
      "500/500 [==============================] - 3s 5ms/step - loss: 0.2354 - accuracy: 0.9276 - val_loss: 0.2043 - val_accuracy: 0.9415\n",
      "Epoch 14/50\n",
      "500/500 [==============================] - 3s 5ms/step - loss: 0.2458 - accuracy: 0.9280 - val_loss: 0.2169 - val_accuracy: 0.9340\n",
      "Epoch 15/50\n",
      "500/500 [==============================] - 3s 5ms/step - loss: 0.2254 - accuracy: 0.9328 - val_loss: 0.2013 - val_accuracy: 0.9402\n",
      "Epoch 16/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.2202 - accuracy: 0.9312 - val_loss: 0.2123 - val_accuracy: 0.9362\n",
      "Epoch 17/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.1823 - accuracy: 0.9472 - val_loss: 0.2137 - val_accuracy: 0.9320\n",
      "Epoch 18/50\n",
      "500/500 [==============================] - 3s 5ms/step - loss: 0.2153 - accuracy: 0.9308 - val_loss: 0.1759 - val_accuracy: 0.9487\n",
      "Epoch 19/50\n",
      "500/500 [==============================] - 3s 5ms/step - loss: 0.2029 - accuracy: 0.9424 - val_loss: 0.1914 - val_accuracy: 0.9404\n",
      "Epoch 20/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.1820 - accuracy: 0.9488 - val_loss: 0.1638 - val_accuracy: 0.9503\n",
      "Epoch 21/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.2013 - accuracy: 0.9348 - val_loss: 0.1738 - val_accuracy: 0.9457\n",
      "Epoch 22/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.1825 - accuracy: 0.9408 - val_loss: 0.1699 - val_accuracy: 0.9487\n",
      "Epoch 23/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.1978 - accuracy: 0.9420 - val_loss: 0.1646 - val_accuracy: 0.9490\n",
      "Epoch 24/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.1902 - accuracy: 0.9344 - val_loss: 0.1611 - val_accuracy: 0.9506\n",
      "Epoch 25/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.1806 - accuracy: 0.9428 - val_loss: 0.1607 - val_accuracy: 0.9507\n",
      "Epoch 26/50\n",
      "500/500 [==============================] - 3s 5ms/step - loss: 0.1662 - accuracy: 0.9464 - val_loss: 0.1595 - val_accuracy: 0.9527\n",
      "Epoch 27/50\n",
      "500/500 [==============================] - 3s 5ms/step - loss: 0.1585 - accuracy: 0.9512 - val_loss: 0.1684 - val_accuracy: 0.9480\n",
      "Epoch 28/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.1751 - accuracy: 0.9512 - val_loss: 0.1611 - val_accuracy: 0.9490\n",
      "Epoch 29/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.1877 - accuracy: 0.9416 - val_loss: 0.1634 - val_accuracy: 0.9485\n",
      "Epoch 30/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.1519 - accuracy: 0.9572 - val_loss: 0.1438 - val_accuracy: 0.9553\n",
      "Epoch 31/50\n",
      "500/500 [==============================] - 3s 5ms/step - loss: 0.1673 - accuracy: 0.9528 - val_loss: 0.1471 - val_accuracy: 0.9570\n",
      "Epoch 32/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.1587 - accuracy: 0.9504 - val_loss: 0.1441 - val_accuracy: 0.9562\n",
      "Epoch 33/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.1378 - accuracy: 0.9572 - val_loss: 0.1508 - val_accuracy: 0.9527\n",
      "Epoch 34/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.1456 - accuracy: 0.9560 - val_loss: 0.1566 - val_accuracy: 0.9499\n",
      "Epoch 35/50\n",
      "500/500 [==============================] - 3s 5ms/step - loss: 0.1593 - accuracy: 0.9528 - val_loss: 0.1340 - val_accuracy: 0.9594\n",
      "Epoch 36/50\n",
      "500/500 [==============================] - 3s 5ms/step - loss: 0.1635 - accuracy: 0.9480 - val_loss: 0.1364 - val_accuracy: 0.9582\n",
      "Epoch 37/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.1561 - accuracy: 0.9456 - val_loss: 0.1430 - val_accuracy: 0.9577\n",
      "Epoch 38/50\n",
      "500/500 [==============================] - 3s 5ms/step - loss: 0.1722 - accuracy: 0.9480 - val_loss: 0.1388 - val_accuracy: 0.9587\n",
      "Epoch 39/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.1513 - accuracy: 0.9472 - val_loss: 0.1579 - val_accuracy: 0.9507\n",
      "Epoch 40/50\n",
      "500/500 [==============================] - 3s 5ms/step - loss: 0.1379 - accuracy: 0.9536 - val_loss: 0.1254 - val_accuracy: 0.9606\n",
      "Epoch 41/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.1408 - accuracy: 0.9580 - val_loss: 0.1314 - val_accuracy: 0.9604\n",
      "Epoch 42/50\n",
      "500/500 [==============================] - 3s 5ms/step - loss: 0.1383 - accuracy: 0.9612 - val_loss: 0.1378 - val_accuracy: 0.9581\n",
      "Epoch 43/50\n",
      "500/500 [==============================] - 3s 5ms/step - loss: 0.1549 - accuracy: 0.9516 - val_loss: 0.1407 - val_accuracy: 0.9572\n",
      "Epoch 44/50\n",
      "500/500 [==============================] - 3s 5ms/step - loss: 0.1353 - accuracy: 0.9580 - val_loss: 0.1253 - val_accuracy: 0.9641\n",
      "Epoch 45/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.1640 - accuracy: 0.9552 - val_loss: 0.1297 - val_accuracy: 0.9594\n",
      "Epoch 46/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.1419 - accuracy: 0.9516 - val_loss: 0.1326 - val_accuracy: 0.9600\n",
      "Epoch 47/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.1316 - accuracy: 0.9604 - val_loss: 0.1185 - val_accuracy: 0.9645\n",
      "Epoch 48/50\n",
      "500/500 [==============================] - 3s 5ms/step - loss: 0.1612 - accuracy: 0.9516 - val_loss: 0.1302 - val_accuracy: 0.9609\n",
      "Epoch 49/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.1254 - accuracy: 0.9572 - val_loss: 0.1284 - val_accuracy: 0.9613\n",
      "Epoch 50/50\n",
      "500/500 [==============================] - 2s 5ms/step - loss: 0.1228 - accuracy: 0.9636 - val_loss: 0.1393 - val_accuracy: 0.9570\n"
     ]
    }
   ],
   "source": [
    "cnn_model.compile(optimizer=tf.optimizers.Adam(),\n",
    "              loss='sparse_categorical_crossentropy',\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "cnn_history = cnn_model.fit(x_train, y_train, steps_per_epoch=500,\n",
    "                        validation_data=(x_test, y_test), \n",
    "                        epochs=50, batch_size=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1462,
     "status": "ok",
     "timestamp": 1594971634260,
     "user": {
      "displayName": "오승혁",
      "photoUrl": "",
      "userId": "17576348803238599032"
     },
     "user_tz": -540
    },
    "id": "8eKXWRjOEPC4",
    "outputId": "9be50887-154e-4129-9952-0922ba5f42a7"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 - 1s - loss: 0.1393 - accuracy: 0.9570\n"
     ]
    }
   ],
   "source": [
    "test_loss, test_acc = cnn_model.evaluate(x_test, y_test, verbose=2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "u4oNotFgEPC5"
   },
   "source": [
    "# Make Quantum Convolution Layer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1212,
     "status": "ok",
     "timestamp": 1594977131092,
     "user": {
      "displayName": "오승혁",
      "photoUrl": "",
      "userId": "17576348803238599032"
     },
     "user_tz": -540
    },
    "id": "dmnXSyM5EPC6"
   },
   "outputs": [],
   "source": [
    "class QConv(tf.keras.layers.Layer):\n",
    "    def __init__(self, filter_size, depth, activation=None, name=None, kernel_regularizer=None, **kwangs):\n",
    "        super(QConv, self).__init__(name=name, **kwangs)\n",
    "        self.filter_size = filter_size\n",
    "        self.depth = depth\n",
    "        self.learning_params = []\n",
    "        self.QCNN_layer_gen()\n",
    "        # self.circuit_tensor = tfq.convert_to_tensor([self.circuit])\n",
    "        self.activation = tf.keras.layers.Activation(activation)\n",
    "        self.kernel_regularizer = kernel_regularizer\n",
    "\n",
    "    def _next_qubit_set(self, original_size, next_size, qubits):\n",
    "        step = original_size // next_size\n",
    "        qubit_list = []\n",
    "        for i in range(0, original_size, step):\n",
    "            for j in range(0, original_size, step):\n",
    "                qubit_list.append(qubits[original_size*i + j])\n",
    "        return qubit_list\n",
    "\n",
    "    def _get_new_param(self):\n",
    "        \"\"\"\n",
    "        return new learnable parameter\n",
    "        all returned parameter saved in self.learning_params\n",
    "        \"\"\"\n",
    "        new_param = sympy.symbols(\"p\"+str(len(self.learning_params)))\n",
    "        self.learning_params.append(new_param)\n",
    "        return new_param\n",
    "    \n",
    "    def _QConv(self, step, target, qubits):\n",
    "        \"\"\"\n",
    "        apply learnable gates each quantum convolutional layer level\n",
    "        \"\"\"\n",
    "        yield cirq.CZPowGate(exponent=self._get_new_param())(qubits[target], qubits[target+step])\n",
    "        yield cirq.CXPowGate(exponent=self._get_new_param())(qubits[target], qubits[target+step])\n",
    "        \n",
    "    def QCNN_layer_gen(self):\n",
    "        \"\"\"\n",
    "        make quantum convolutional layer in QConv layer\n",
    "        \"\"\"\n",
    "        pixels = self.filter_size**2\n",
    "        # filter size: 2^n only for this version!\n",
    "        if np.log2(pixels) % 1 != 0:\n",
    "            raise NotImplementedError(\"filter size: 2^n only available\")\n",
    "        cirq_qubits = cirq.GridQubit.rect(self.filter_size, self.filter_size)\n",
    "        # mapping input data to circuit\n",
    "        input_circuit = cirq.Circuit()\n",
    "        input_params = [sympy.symbols('a%d' %i) for i in range(pixels)]\n",
    "        for i, qubit in enumerate(cirq_qubits):\n",
    "            input_circuit.append(cirq.rx(np.pi*input_params[i])(qubit))\n",
    "        # apply learnable gate set to QCNN circuit\n",
    "        QCNN_circuit = cirq.Circuit()\n",
    "        step_size = [2**i for i in range(np.log2(pixels).astype(np.int32))]\n",
    "        for step in step_size:\n",
    "            for target in range(0, pixels, 2*step):\n",
    "                QCNN_circuit.append(self._QConv(step, target, cirq_qubits))\n",
    "        # merge the circuits\n",
    "        full_circuit = cirq.Circuit()\n",
    "        full_circuit.append(input_circuit)\n",
    "        full_circuit.append(QCNN_circuit)\n",
    "        self.circuit = full_circuit # save circuit to the QCNN layer obj.\n",
    "        self.params = input_params + self.learning_params\n",
    "        self.op = cirq.Z(cirq_qubits[0])\n",
    "        \n",
    "    def build(self, input_shape):\n",
    "        self.width = input_shape[1]\n",
    "        self.height = input_shape[2]\n",
    "        self.channel = input_shape[3]\n",
    "        self.num_x = self.width - self.filter_size + 1\n",
    "        self.num_y = self.height - self.filter_size + 1\n",
    "        \n",
    "        self.kernel = self.add_weight(name=\"kenel\", \n",
    "                                      shape=[self.depth, \n",
    "                                             self.channel, \n",
    "                                             len(self.learning_params)],\n",
    "                                     initializer=tf.keras.initializers.glorot_normal(),\n",
    "                                     regularizer=self.kernel_regularizer)\n",
    "        self.circuit_tensor = tfq.convert_to_tensor([self.circuit] * self.num_x * self.num_y * self.channel)\n",
    "        \n",
    "    def call(self, inputs):\n",
    "        # input shape: [N, width, height, channel]\n",
    "        # slide and collect data\n",
    "        stack_set = None\n",
    "        for i in range(self.num_x):\n",
    "            for j in range(self.num_y):\n",
    "                slice_part = tf.slice(inputs, [0, i, j, 0], [-1, self.filter_size, self.filter_size, -1])\n",
    "                slice_part = tf.reshape(slice_part, shape=[-1, 1, self.filter_size, self.filter_size, self.channel])\n",
    "                if stack_set == None:\n",
    "                    stack_set = slice_part\n",
    "                else:\n",
    "                    stack_set = tf.concat([stack_set, slice_part], 1)  \n",
    "        # -> shape: [N, num_x*num_y, filter_size, filter_size, channel]\n",
    "        stack_set = tf.transpose(stack_set, perm=[0, 1, 4, 2, 3])\n",
    "        # -> shape: [N, num_x*num_y, channel, filter_size, fiter_size]\n",
    "        stack_set = tf.reshape(stack_set, shape=[-1, self.filter_size**2])\n",
    "        # -> shape: [N*num_x*num_y*channel, filter_size^2]\n",
    "        \n",
    "        # total input citcuits: N * num_x * num_y * channel\n",
    "        circuit_inputs = tf.tile([self.circuit_tensor], [tf.shape(inputs)[0], 1])\n",
    "        circuit_inputs = tf.reshape(circuit_inputs, shape=[-1])\n",
    "        tf.fill([tf.shape(inputs)[0]*self.num_x*self.num_y, 1], 1)\n",
    "        outputs = []\n",
    "        for i in range(self.depth):\n",
    "            controller = tf.tile(self.kernel[i], [tf.shape(inputs)[0]*self.num_x*self.num_y, 1])\n",
    "            outputs.append(self.single_depth_QCNN(stack_set, controller, circuit_inputs))\n",
    "            # shape: [N, num_x, num_y] \n",
    "            \n",
    "        output_tensor = tf.stack(outputs, axis=3)\n",
    "        output_tensor = tf.math.acos(tf.clip_by_value(output_tensor, -1+1e-5, 1-1e-5)) / np.pi\n",
    "        # output_tensor = tf.clip_by_value(tf.math.acos(output_tensor)/np.pi, -1, 1)\n",
    "        return self.activation(output_tensor)\n",
    "          \n",
    "    def single_depth_QCNN(self, input_data, controller, circuit_inputs):\n",
    "        \"\"\"\n",
    "        make QCNN for 1 channel only\n",
    "        \"\"\"\n",
    "        # input shape: [N*num_x*num_y*channel, filter_size^2]\n",
    "        # controller shape: [N*num_x*num_y*channel, len(learning_params)]\n",
    "        input_data = tf.concat([input_data, controller], 1)\n",
    "        # input_data shape: [N*num_x*num_y*channel, len(learning_params)]\n",
    "        QCNN_output = tfq.layers.Expectation()(circuit_inputs, \n",
    "                                               symbol_names=self.params,\n",
    "                                               symbol_values=input_data,\n",
    "                                               operators=self.op)\n",
    "        # QCNN_output shape: [N*num_x*num_y*channel]\n",
    "        QCNN_output = tf.reshape(QCNN_output, shape=[-1, self.num_x, self.num_y, self.channel])\n",
    "        return tf.math.reduce_sum(QCNN_output, 3)\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1249,
     "status": "ok",
     "timestamp": 1594977593117,
     "user": {
      "displayName": "오승혁",
      "photoUrl": "",
      "userId": "17576348803238599032"
     },
     "user_tz": -540
    },
    "id": "j4_hQbAXEPC8"
   },
   "outputs": [],
   "source": [
    "width = np.shape(x_train)[1]\n",
    "height = np.shape(x_train)[2]\n",
    "\n",
    "qcnn_model = models.Sequential()\n",
    "\n",
    "\n",
    "qcnn_model.add(QConv(filter_size=2, depth=8, activation='relu', \n",
    "                     name='qconv1', input_shape=(width, height, 1)))\n",
    "#model.add(layers.Conv2D(16, (2, 2), activation='relu'))\n",
    "qcnn_model.add(layers.Flatten())\n",
    "qcnn_model.add(layers.Dense(32, activation='relu'))\n",
    "qcnn_model.add(layers.Dense(10, activation='softmax'))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 341
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 920,
     "status": "ok",
     "timestamp": 1594977593120,
     "user": {
      "displayName": "오승혁",
      "photoUrl": "",
      "userId": "17576348803238599032"
     },
     "user_tz": -540
    },
    "id": "89ttC7n9EPC9",
    "outputId": "84bcfda8-4b27-49a4-8216-6148d8893cf7"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_2\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "qconv1 (QConv)               (None, 9, 9, 8)           48        \n",
      "_________________________________________________________________\n",
      "flatten_2 (Flatten)          (None, 648)               0         \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 32)                20768     \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 10)                330       \n",
      "=================================================================\n",
      "Total params: 21,146\n",
      "Trainable params: 21,146\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "qcnn_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 220
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 987,
     "status": "ok",
     "timestamp": 1594977593562,
     "user": {
      "displayName": "오승혁",
      "photoUrl": "",
      "userId": "17576348803238599032"
     },
     "user_tz": -540
    },
    "id": "IMjGoNGiEPC_",
    "outputId": "f49fc139-aa8b-4dff-fa0c-b94ec9c1537a",
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "findfont: Font family ['Arial'] not found. Falling back to DejaVu Sans.\n"
     ]
    },
    {
     "data": {
      "image/svg+xml": [
       "<svg xmlns=\"http://www.w3.org/2000/svg\" width=\"614.768828125\" height=\"200.0\"><line x1=\"34.7588671875\" x2=\"584.768828125\" y1=\"25.0\" y2=\"25.0\" stroke=\"#1967d2\" stroke-width=\"1\" /><line x1=\"34.7588671875\" x2=\"584.768828125\" y1=\"75.0\" y2=\"75.0\" stroke=\"#1967d2\" stroke-width=\"1\" /><line x1=\"34.7588671875\" x2=\"584.768828125\" y1=\"125.0\" y2=\"125.0\" stroke=\"#1967d2\" stroke-width=\"1\" /><line x1=\"34.7588671875\" x2=\"584.768828125\" y1=\"175.0\" y2=\"175.0\" stroke=\"#1967d2\" stroke-width=\"1\" /><line x1=\"324.25851562500003\" x2=\"324.25851562500003\" y1=\"25.0\" y2=\"75.0\" stroke=\"black\" stroke-width=\"3\" /><line x1=\"324.25851562500003\" x2=\"324.25851562500003\" y1=\"125.0\" y2=\"175.0\" stroke=\"black\" stroke-width=\"3\" /><line x1=\"395.83414062500003\" x2=\"395.83414062500003\" y1=\"25.0\" y2=\"75.0\" stroke=\"black\" stroke-width=\"3\" /><line x1=\"395.83414062500003\" x2=\"395.83414062500003\" y1=\"125.0\" y2=\"175.0\" stroke=\"black\" stroke-width=\"3\" /><line x1=\"467.481953125\" x2=\"467.481953125\" y1=\"25.0\" y2=\"125.0\" stroke=\"black\" stroke-width=\"3\" /><line x1=\"539.0794531250001\" x2=\"539.0794531250001\" y1=\"25.0\" y2=\"125.0\" stroke=\"black\" stroke-width=\"3\" /><rect x=\"10.0\" y=\"5.0\" width=\"49.517734375\" height=\"40\" stroke=\"black\" fill=\"white\" stroke-width=\"0\" /><text x=\"34.7588671875\" y=\"25.0\" dominant-baseline=\"middle\" text-anchor=\"middle\" font-size=\"14px\">(0, 0): </text><rect x=\"10.0\" y=\"55.0\" width=\"49.517734375\" height=\"40\" stroke=\"black\" fill=\"white\" stroke-width=\"0\" /><text x=\"34.7588671875\" y=\"75.0\" dominant-baseline=\"middle\" text-anchor=\"middle\" font-size=\"14px\">(0, 1): </text><rect x=\"10.0\" y=\"105.0\" width=\"49.517734375\" height=\"40\" stroke=\"black\" fill=\"white\" stroke-width=\"0\" /><text x=\"34.7588671875\" y=\"125.0\" dominant-baseline=\"middle\" text-anchor=\"middle\" font-size=\"14px\">(1, 0): </text><rect x=\"10.0\" y=\"155.0\" width=\"49.517734375\" height=\"40\" stroke=\"black\" fill=\"white\" stroke-width=\"0\" /><text x=\"34.7588671875\" y=\"175.0\" dominant-baseline=\"middle\" text-anchor=\"middle\" font-size=\"14px\">(1, 1): </text><rect x=\"79.51773437499999\" y=\"5.0\" width=\"198.90484375000003\" height=\"40\" stroke=\"black\" fill=\"white\" stroke-width=\"1\" /><text x=\"178.97015625\" y=\"25.0\" dominant-baseline=\"middle\" text-anchor=\"middle\" font-size=\"14px\">Rx(3.14159265358979*a0)</text><rect x=\"79.51773437499999\" y=\"55.0\" width=\"198.90484375000003\" height=\"40\" stroke=\"black\" fill=\"white\" stroke-width=\"1\" /><text x=\"178.97015625\" y=\"75.0\" dominant-baseline=\"middle\" text-anchor=\"middle\" font-size=\"14px\">Rx(3.14159265358979*a1)</text><rect x=\"79.51773437499999\" y=\"105.0\" width=\"198.90484375000003\" height=\"40\" stroke=\"black\" fill=\"white\" stroke-width=\"1\" /><text x=\"178.97015625\" y=\"125.0\" dominant-baseline=\"middle\" text-anchor=\"middle\" font-size=\"14px\">Rx(3.14159265358979*a2)</text><rect x=\"79.51773437499999\" y=\"155.0\" width=\"198.90484375000003\" height=\"40\" stroke=\"black\" fill=\"white\" stroke-width=\"1\" /><text x=\"178.97015625\" y=\"175.0\" dominant-baseline=\"middle\" text-anchor=\"middle\" font-size=\"14px\">Rx(3.14159265358979*a3)</text><rect x=\"298.42257812500003\" y=\"55.0\" width=\"51.671875\" height=\"40\" stroke=\"black\" fill=\"white\" stroke-width=\"1\" /><text x=\"324.25851562500003\" y=\"75.0\" dominant-baseline=\"middle\" text-anchor=\"middle\" font-size=\"14px\">@^p0</text><circle cx=\"324.25851562500003\" cy=\"25.0\" r=\"10.0\" /><rect x=\"298.42257812500003\" y=\"155.0\" width=\"51.671875\" height=\"40\" stroke=\"black\" fill=\"white\" stroke-width=\"1\" /><text x=\"324.25851562500003\" y=\"175.0\" dominant-baseline=\"middle\" text-anchor=\"middle\" font-size=\"14px\">@^p2</text><circle cx=\"324.25851562500003\" cy=\"125.0\" r=\"10.0\" /><rect x=\"370.09445312500003\" y=\"55.0\" width=\"51.479375000000005\" height=\"40\" stroke=\"black\" fill=\"white\" stroke-width=\"1\" /><text x=\"395.83414062500003\" y=\"75.0\" dominant-baseline=\"middle\" text-anchor=\"middle\" font-size=\"14px\">@^p1</text><rect x=\"370.09445312500003\" y=\"5.0\" width=\"51.479375000000005\" height=\"40\" stroke=\"black\" fill=\"white\" stroke-width=\"1\" /><text x=\"395.83414062500003\" y=\"25.0\" dominant-baseline=\"middle\" text-anchor=\"middle\" font-size=\"18px\">X</text><rect x=\"370.09445312500003\" y=\"155.0\" width=\"51.479375000000005\" height=\"40\" stroke=\"black\" fill=\"white\" stroke-width=\"1\" /><text x=\"395.83414062500003\" y=\"175.0\" dominant-baseline=\"middle\" text-anchor=\"middle\" font-size=\"14px\">@^p3</text><rect x=\"370.09445312500003\" y=\"105.0\" width=\"51.479375000000005\" height=\"40\" stroke=\"black\" fill=\"white\" stroke-width=\"1\" /><text x=\"395.83414062500003\" y=\"125.0\" dominant-baseline=\"middle\" text-anchor=\"middle\" font-size=\"18px\">X</text><rect x=\"441.57382812500003\" y=\"105.0\" width=\"51.81625\" height=\"40\" stroke=\"black\" fill=\"white\" stroke-width=\"1\" /><text x=\"467.481953125\" y=\"125.0\" dominant-baseline=\"middle\" text-anchor=\"middle\" font-size=\"14px\">@^p4</text><circle cx=\"467.481953125\" cy=\"25.0\" r=\"10.0\" /><rect x=\"513.3900781250001\" y=\"105.0\" width=\"51.37875\" height=\"40\" stroke=\"black\" fill=\"white\" stroke-width=\"1\" /><text x=\"539.0794531250001\" y=\"125.0\" dominant-baseline=\"middle\" text-anchor=\"middle\" font-size=\"14px\">@^p5</text><rect x=\"513.3900781250001\" y=\"5.0\" width=\"51.37875\" height=\"40\" stroke=\"black\" fill=\"white\" stroke-width=\"1\" /><text x=\"539.0794531250001\" y=\"25.0\" dominant-baseline=\"middle\" text-anchor=\"middle\" font-size=\"18px\">X</text></svg>"
      ],
      "text/plain": [
       "<cirq.contrib.svg.svg.SVGCircuit at 0x7fae3442a7c0>"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "SVGCircuit(QConv(filter_size=2, depth=0, activation='relu').circuit)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 533
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1492,
     "status": "ok",
     "timestamp": 1594977594367,
     "user": {
      "displayName": "오승혁",
      "photoUrl": "",
      "userId": "17576348803238599032"
     },
     "user_tz": -540
    },
    "id": "a_v5xWKfEPDB",
    "outputId": "461769fe-d978-4cc6-a6ed-0c207eca1eae",
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pydot\n",
    "import graphviz\n",
    "from tensorflow.keras.utils import plot_model\n",
    "\n",
    "plot_model(qcnn_model, to_file='model_shapes.png', show_shapes=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 89
    },
    "colab_type": "code",
    "id": "JhRWSq6UEPDC",
    "outputId": "b605295c-c38d-4fb2-b21a-6e83de8cc99c",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "500/500 [==============================] - 323s 645ms/step - loss: 0.9971 - accuracy: 0.7172 - val_loss: 0.5769 - val_accuracy: 0.8336\n",
      "Epoch 2/50\n",
      "500/500 [==============================] - 328s 657ms/step - loss: 0.5242 - accuracy: 0.8436 - val_loss: 0.4749 - val_accuracy: 0.8618\n",
      "Epoch 3/50\n",
      "500/500 [==============================] - 301s 603ms/step - loss: 0.4328 - accuracy: 0.8720 - val_loss: 0.3819 - val_accuracy: 0.8893\n",
      "Epoch 4/50\n",
      "500/500 [==============================] - 302s 605ms/step - loss: 0.3766 - accuracy: 0.8816 - val_loss: 0.3329 - val_accuracy: 0.9011\n",
      "Epoch 5/50\n",
      "500/500 [==============================] - 301s 603ms/step - loss: 0.3460 - accuracy: 0.8936 - val_loss: 0.3193 - val_accuracy: 0.9050\n",
      "Epoch 6/50\n",
      "500/500 [==============================] - 280s 560ms/step - loss: 0.3203 - accuracy: 0.9020 - val_loss: 0.2845 - val_accuracy: 0.9138\n",
      "Epoch 7/50\n",
      "500/500 [==============================] - 278s 557ms/step - loss: 0.3522 - accuracy: 0.8996 - val_loss: 0.2743 - val_accuracy: 0.9194\n",
      "Epoch 8/50\n",
      "500/500 [==============================] - 277s 554ms/step - loss: 0.3036 - accuracy: 0.9012 - val_loss: 0.2871 - val_accuracy: 0.9128\n",
      "Epoch 9/50\n",
      "500/500 [==============================] - 277s 553ms/step - loss: 0.2954 - accuracy: 0.9092 - val_loss: 0.2533 - val_accuracy: 0.9246\n",
      "Epoch 10/50\n",
      "500/500 [==============================] - 276s 553ms/step - loss: 0.2724 - accuracy: 0.9120 - val_loss: 0.2291 - val_accuracy: 0.9337\n",
      "Epoch 11/50\n",
      "500/500 [==============================] - 276s 551ms/step - loss: 0.2444 - accuracy: 0.9312 - val_loss: 0.2345 - val_accuracy: 0.9318\n",
      "Epoch 12/50\n",
      "500/500 [==============================] - 275s 549ms/step - loss: 0.2735 - accuracy: 0.9204 - val_loss: 0.2280 - val_accuracy: 0.9304\n",
      "Epoch 13/50\n",
      "500/500 [==============================] - 275s 549ms/step - loss: 0.2391 - accuracy: 0.9356 - val_loss: 0.2149 - val_accuracy: 0.9357\n",
      "Epoch 14/50\n",
      "500/500 [==============================] - 275s 549ms/step - loss: 0.2540 - accuracy: 0.9192 - val_loss: 0.2312 - val_accuracy: 0.9289\n",
      "Epoch 15/50\n",
      "500/500 [==============================] - 273s 547ms/step - loss: 0.1823 - accuracy: 0.9464 - val_loss: 0.2197 - val_accuracy: 0.9318\n",
      "Epoch 16/50\n",
      "500/500 [==============================] - 273s 546ms/step - loss: 0.2529 - accuracy: 0.9260 - val_loss: 0.1985 - val_accuracy: 0.9421\n",
      "Epoch 17/50\n",
      "500/500 [==============================] - 274s 549ms/step - loss: 0.2105 - accuracy: 0.9352 - val_loss: 0.1974 - val_accuracy: 0.9407\n",
      "Epoch 18/50\n",
      "500/500 [==============================] - 274s 548ms/step - loss: 0.2299 - accuracy: 0.9264 - val_loss: 0.1952 - val_accuracy: 0.9386\n",
      "Epoch 19/50\n",
      "500/500 [==============================] - 273s 547ms/step - loss: 0.2351 - accuracy: 0.9300 - val_loss: 0.1789 - val_accuracy: 0.9445\n",
      "Epoch 20/50\n",
      "500/500 [==============================] - 273s 546ms/step - loss: 0.2000 - accuracy: 0.9360 - val_loss: 0.2011 - val_accuracy: 0.9369\n",
      "Epoch 21/50\n",
      "500/500 [==============================] - 273s 546ms/step - loss: 0.2283 - accuracy: 0.9268 - val_loss: 0.2080 - val_accuracy: 0.9359\n",
      "Epoch 22/50\n",
      "500/500 [==============================] - 272s 545ms/step - loss: 0.2306 - accuracy: 0.9264 - val_loss: 0.1880 - val_accuracy: 0.9404\n",
      "Epoch 23/50\n",
      "500/500 [==============================] - 273s 545ms/step - loss: 0.2151 - accuracy: 0.9336 - val_loss: 0.1847 - val_accuracy: 0.9447\n",
      "Epoch 24/50\n",
      "500/500 [==============================] - 274s 547ms/step - loss: 0.2159 - accuracy: 0.9364 - val_loss: 0.1915 - val_accuracy: 0.9407\n",
      "Epoch 25/50\n",
      "500/500 [==============================] - 273s 545ms/step - loss: 0.1682 - accuracy: 0.9476 - val_loss: 0.1781 - val_accuracy: 0.9463\n",
      "Epoch 26/50\n",
      "500/500 [==============================] - 272s 544ms/step - loss: 0.1631 - accuracy: 0.9476 - val_loss: 0.1813 - val_accuracy: 0.9446\n",
      "Epoch 27/50\n",
      "500/500 [==============================] - 273s 546ms/step - loss: 0.1910 - accuracy: 0.9384 - val_loss: 0.1786 - val_accuracy: 0.9439\n",
      "Epoch 28/50\n",
      "500/500 [==============================] - 272s 543ms/step - loss: 0.1910 - accuracy: 0.9412 - val_loss: 0.1876 - val_accuracy: 0.9415\n",
      "Epoch 29/50\n",
      "500/500 [==============================] - 273s 546ms/step - loss: 0.1906 - accuracy: 0.9396 - val_loss: 0.1827 - val_accuracy: 0.9453\n",
      "Epoch 30/50\n",
      "500/500 [==============================] - 274s 548ms/step - loss: 0.2123 - accuracy: 0.9392 - val_loss: 0.1702 - val_accuracy: 0.9474\n",
      "Epoch 31/50\n",
      "500/500 [==============================] - 272s 544ms/step - loss: 0.1787 - accuracy: 0.9448 - val_loss: 0.1646 - val_accuracy: 0.9504\n",
      "Epoch 32/50\n",
      "500/500 [==============================] - 272s 545ms/step - loss: 0.1784 - accuracy: 0.9400 - val_loss: 0.1720 - val_accuracy: 0.9459\n",
      "Epoch 33/50\n",
      "500/500 [==============================] - 273s 545ms/step - loss: 0.1858 - accuracy: 0.9416 - val_loss: 0.1625 - val_accuracy: 0.9493\n",
      "Epoch 34/50\n",
      "500/500 [==============================] - 273s 546ms/step - loss: 0.1693 - accuracy: 0.9468 - val_loss: 0.1693 - val_accuracy: 0.9490\n",
      "Epoch 35/50\n",
      "500/500 [==============================] - 273s 545ms/step - loss: 0.1714 - accuracy: 0.9472 - val_loss: 0.1942 - val_accuracy: 0.9395\n",
      "Epoch 36/50\n",
      "500/500 [==============================] - 272s 543ms/step - loss: 0.1797 - accuracy: 0.9452 - val_loss: 0.1678 - val_accuracy: 0.9489\n",
      "Epoch 37/50\n",
      "500/500 [==============================] - 273s 546ms/step - loss: 0.1972 - accuracy: 0.9356 - val_loss: 0.1559 - val_accuracy: 0.9518\n",
      "Epoch 38/50\n",
      "500/500 [==============================] - 273s 545ms/step - loss: 0.1721 - accuracy: 0.9480 - val_loss: 0.1590 - val_accuracy: 0.9511\n",
      "Epoch 39/50\n",
      "500/500 [==============================] - 273s 547ms/step - loss: 0.1688 - accuracy: 0.9496 - val_loss: 0.1832 - val_accuracy: 0.9413\n",
      "Epoch 40/50\n",
      "500/500 [==============================] - 273s 545ms/step - loss: 0.1517 - accuracy: 0.9556 - val_loss: 0.1531 - val_accuracy: 0.9534\n",
      "Epoch 41/50\n",
      "500/500 [==============================] - 273s 546ms/step - loss: 0.1719 - accuracy: 0.9420 - val_loss: 0.1753 - val_accuracy: 0.9441\n",
      "Epoch 42/50\n",
      "500/500 [==============================] - 272s 544ms/step - loss: 0.1762 - accuracy: 0.9476 - val_loss: 0.1640 - val_accuracy: 0.9511\n",
      "Epoch 43/50\n",
      "500/500 [==============================] - 273s 546ms/step - loss: 0.1567 - accuracy: 0.9416 - val_loss: 0.1685 - val_accuracy: 0.9454\n",
      "Epoch 44/50\n",
      "500/500 [==============================] - 274s 548ms/step - loss: 0.1484 - accuracy: 0.9504 - val_loss: 0.1572 - val_accuracy: 0.9513\n",
      "Epoch 45/50\n",
      "500/500 [==============================] - 325s 651ms/step - loss: 0.1799 - accuracy: 0.9452 - val_loss: 0.1448 - val_accuracy: 0.9524\n",
      "Epoch 46/50\n",
      "500/500 [==============================] - 325s 650ms/step - loss: 0.1424 - accuracy: 0.9572 - val_loss: 0.1530 - val_accuracy: 0.9547\n",
      "Epoch 47/50\n",
      "500/500 [==============================] - 321s 641ms/step - loss: 0.1511 - accuracy: 0.9512 - val_loss: 0.1526 - val_accuracy: 0.9537\n",
      "Epoch 48/50\n",
      "500/500 [==============================] - 299s 598ms/step - loss: 0.1613 - accuracy: 0.9516 - val_loss: 0.1658 - val_accuracy: 0.9460\n",
      "Epoch 49/50\n",
      "500/500 [==============================] - 300s 599ms/step - loss: 0.1376 - accuracy: 0.9596 - val_loss: 0.1540 - val_accuracy: 0.9540\n",
      "Epoch 50/50\n",
      "500/500 [==============================] - 295s 590ms/step - loss: 0.1595 - accuracy: 0.9496 - val_loss: 0.1522 - val_accuracy: 0.9532\n"
     ]
    }
   ],
   "source": [
    "qcnn_model.compile(optimizer=tf.keras.optimizers.Adam(),\n",
    "              loss='sparse_categorical_crossentropy',\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "qcnn_history = qcnn_model.fit(x_train, y_train, steps_per_epoch=500,\n",
    "                        validation_data=(x_test, y_test), \n",
    "                        epochs=50, batch_size=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "oSjNaowxEPDE",
    "outputId": "fe340c9f-c64a-4590-c569-7499f326b999"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 - 179s - loss: 0.1522 - accuracy: 0.9532\n"
     ]
    }
   ],
   "source": [
    "test_loss, test_acc = qcnn_model.evaluate(x_test, y_test, verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for i in range(8):\n",
    "    fig = plt.figure()\n",
    "    plt.axis('off')\n",
    "    plt.imshow(qcnn_model.get_layer('qconv1')(x_train[0:1,:,:,:])[0,:,:,i], cmap='gray_r')\n",
    "    fig.savefig('qcnn_%d'%i, dpi=300, bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Comparing Result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_loss_curves(fc_loss, cnn_loss, qcnn_loss):\n",
    "    fig = plt.figure()\n",
    "    plt.plot(np.arange(len(fc_loss)) + 1, fc_loss, \"go-\", label=\"fully-connected\")\n",
    "    plt.plot(np.arange(len(cnn_loss)) + 1, cnn_loss, \"rs-\", label=\"CNN\")\n",
    "    plt.plot(np.arange(len(qcnn_loss)) + 1, qcnn_loss, \"b^-\", label=\"QCNN\")\n",
    "    plt.gca().xaxis.set_major_locator(mpl.ticker.MaxNLocator(integer=True))\n",
    "    plt.axis([1, 50, 0, 1])\n",
    "    plt.legend(fontsize=14)\n",
    "    plt.xlabel(\"Epochs\")\n",
    "    plt.ylabel(\"Test set loss\")\n",
    "    plt.grid(True)\n",
    "    fig.savefig('loss.png', dpi=300)\n",
    "    \n",
    "def plot_loss_curves_small(fc_loss, cnn_loss, qcnn_loss):\n",
    "    fig = plt.figure()\n",
    "    plt.plot(np.arange(len(fc_loss))[::2] + 1, fc_loss[::2], \"go-\", label=\"fully-connected\")\n",
    "    plt.plot(np.arange(len(cnn_loss))[::2] + 1, cnn_loss[::2], \"rs-\", label=\"CNN\")\n",
    "    plt.plot(np.arange(len(qcnn_loss))[::2] + 1, qcnn_loss[::2], \"b^-\", label=\"QCNN\")\n",
    "    plt.gca().xaxis.set_major_locator(mpl.ticker.MaxNLocator(integer=True))\n",
    "    plt.axis([1, 50, 0, 1])\n",
    "    plt.legend(fontsize=14)\n",
    "    plt.xlabel(\"Epochs\")\n",
    "    plt.ylabel(\"Test set loss\")\n",
    "    plt.grid(True)\n",
    "    fig.savefig('loss_small.png', dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_loss_curves(fc_history.history['val_loss'], cnn_history.history['val_loss'], qcnn_history.history['val_loss'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_loss_curves_small(fc_history.history['val_loss'], cnn_history.history['val_loss'], qcnn_history.history['val_loss'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_acc_curves(fc_acc, cnn_acc, qcnn_acc):\n",
    "    fig = plt.figure()\n",
    "    plt.plot(np.arange(len(fc_acc)) + 1, fc_acc, \"go-\", label=\"fully-connected\")\n",
    "    plt.plot(np.arange(len(cnn_acc)) + 1, cnn_acc, \"rs-\", label=\"CNN\")\n",
    "    plt.plot(np.arange(len(qcnn_acc)) + 1, qcnn_acc, \"b^-\", label=\"QCNN\")\n",
    "    plt.gca().xaxis.set_major_locator(mpl.ticker.MaxNLocator(integer=True))\n",
    "    plt.grid()\n",
    "    plt.axis([1, 50, 0.8, 1])\n",
    "    plt.legend(fontsize=14)\n",
    "    plt.xlabel(\"Epochs\")\n",
    "    plt.ylabel(\"Test set accuracy\")\n",
    "    fig.savefig('accuracy.png', dpi=300)\n",
    "    \n",
    "def plot_acc_curves_small(fc_acc, cnn_acc, qcnn_acc):\n",
    "    fig = plt.figure()\n",
    "    plt.plot(np.arange(len(fc_acc))[::2] + 1, fc_acc[::2], \"go-\", label=\"fully-connected\")\n",
    "    plt.plot(np.arange(len(cnn_acc))[::2] + 1, cnn_acc[::2], \"rs-\", label=\"CNN\")\n",
    "    plt.plot(np.arange(len(qcnn_acc))[::2] + 1, qcnn_acc[::2], \"b^-\", label=\"QCNN\")\n",
    "    plt.gca().xaxis.set_major_locator(mpl.ticker.MaxNLocator(integer=True))\n",
    "    plt.grid()\n",
    "    plt.axis([1, 50, 0.8, 1])\n",
    "    plt.legend(fontsize=14)\n",
    "    plt.xlabel(\"Epochs\")\n",
    "    plt.ylabel(\"Test set accuracy\")\n",
    "    fig.savefig('accuracy_small.png', dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_acc_curves(fc_history.history['val_accuracy'], cnn_history.history['val_accuracy'], qcnn_history.history['val_accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_acc_curves_small(fc_history.history['val_accuracy'], cnn_history.history['val_accuracy'], qcnn_history.history['val_accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "name": "QCNN_MNIST-1.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.8.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}