Tensorflow/FunctionalAPI-2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "view-in-github"
   },
   "source": [
    "<a href=\"https://colab.research.google.com/github/aim97/Neural-networks/blob/master/Tensorflow/FunctionalAPI-enenrgy%2Befficiency.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "7d0eUfvZ2AbS"
   },
   "source": [
    "# Tensorflow Functional API\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "id": "YKNALtjTIkua"
   },
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow.keras.layers import Dense, Input, Flatten, Dropout, Lambda\n",
    "import tensorflow.keras.backend as K\n",
    "from tensorflow.python.keras.utils.vis_utils import plot_model\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Siamese network\n",
    "\n",
    "It's a sub model that recieves multiple inputs, and uses the same weights for all of them, in the image below, you can see two identical parts of the model, marked as having shared weights, in code we use one network and simply pass the two inputs through it.  \n",
    "\n",
    "\n",
    "![Siamese network example](../images/Siamese-network-example.png)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# construct a Siamese network with shared weights\n",
    "# The input is 28x28 images.\n",
    "# The output is a vector of size 128.\n",
    "def siamese_network(prefex=''):\n",
    "  xs = Input((28, 28), name=prefex+'input')\n",
    "  x = Flatten(name=prefex+'_Flatten')(xs)\n",
    "  x = Dense(128, activation='relu', name=prefex+'_D1')(x)\n",
    "  x = Dropout(0.1, name=prefex+'_D1_dropout')(x)\n",
    "  x = Dense(128, activation='relu', name=prefex+'_D2')(x)\n",
    "  x = Dropout(0.1, name=prefex+'_D2_dropout')(x)\n",
    "  x = Dense(128, activation='relu', name=prefex+'_D3')(x)\n",
    "  return tf.keras.Model(inputs=[xs], outputs=[x])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "# Siamese network in Tensorflow\n",
    "\n",
    "In the code above we have a function that *returns* a Siamese network model, each time this function is called a new model is created.  \n",
    "The Siamese network should be shared between two (or more) pathes in the same model, that's why the function must return a model (or more accurately a sub-model) and not just some output tensor.  \n",
    "\n",
    "That model will be passed the two inputs, so it's shared by both pathes.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Construct the model that uses the Siamese network\n",
    "# create two inputs for the siamese network\n",
    "left_input = Input((28, 28), name='left_input')\n",
    "right_input = Input((28, 28), name='right_input')\n",
    "\n",
    "# create the siamese network\n",
    "siamese_net = siamese_network('siamese_')\n",
    "\n",
    "# create the outputs of the siamese network\n",
    "left_output = siamese_net(left_input)\n",
    "right_output = siamese_net(right_input)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Distance layer\n",
    "\n",
    "Our model should compute how *close* 2 input images are, so we need a distance layer. the job of the Siamese network is to generate something like a sematic vector represeting the image, then the distance layer will compare the two vectors and return the distance between them.\n",
    "\n",
    "For creating such a layer we use `Lambda` which allows the execution of an arbitrary simple operation as a layer within the model.  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# define euclidean distance function\n",
    "def euclidean_distance(vects):\n",
    "  x, y = vects\n",
    "  return K.sqrt(K.maximum(K.sum(K.square(x - y), axis=1, keepdims=True), K.epsilon()))\n",
    "\n",
    "def eucl_dist_output_shape(shapes):\n",
    "  shape1, shape2 = shapes\n",
    "  return (shape1[0], 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "# merge the outputs of the siamese network with euclidean distance\n",
    "euclidean_distance = Lambda(euclidean_distance, output_shape=eucl_dist_output_shape)([left_output, right_output])\n",
    "\n",
    "# and create the final model\n",
    "model = tf.keras.Model(inputs=[left_input, right_input], outputs=[euclidean_distance])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "siamese_input (InputLayer)   [(None, 28, 28)]          0         \n",
      "_________________________________________________________________\n",
      "siamese__Flatten (Flatten)   (None, 784)               0         \n",
      "_________________________________________________________________\n",
      "siamese__D1 (Dense)          (None, 128)               100480    \n",
      "_________________________________________________________________\n",
      "siamese__D1_dropout (Dropout (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "siamese__D2 (Dense)          (None, 128)               16512     \n",
      "_________________________________________________________________\n",
      "siamese__D2_dropout (Dropout (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "siamese__D3 (Dense)          (None, 128)               16512     \n",
      "=================================================================\n",
      "Total params: 133,504\n",
      "Trainable params: 133,504\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# show Siamese network architecture and summary\n",
    "print(siamese_net.summary())\n",
    "plot_model(siamese_net, to_file='../ignored/functional-api-2-siamesenet.png' , show_shapes=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_1\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "left_input (InputLayer)         [(None, 28, 28)]     0                                            \n",
      "__________________________________________________________________________________________________\n",
      "right_input (InputLayer)        [(None, 28, 28)]     0                                            \n",
      "__________________________________________________________________________________________________\n",
      "model (Functional)              (None, 128)          133504      left_input[0][0]                 \n",
      "                                                                 right_input[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "lambda (Lambda)                 (None, 1)            0           model[0][0]                      \n",
      "                                                                 model[1][0]                      \n",
      "==================================================================================================\n",
      "Total params: 133,504\n",
      "Trainable params: 133,504\n",
      "Non-trainable params: 0\n",
      "__________________________________________________________________________________________________\n",
      "None\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# show the model architecture\n",
    "print(model.summary())\n",
    "\n",
    "plot_model(model, to_file='../ignored/functional-api-2-model.png' , show_shapes=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Specify the optimizer, and compile the model with loss functions for both outputs\n",
    "optimizer = tf.keras.optimizers.RMSprop(learning_rate=0.001)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Contrastive loss function (we will address this in a future post)\n",
    "def contrastive_loss_with_margin(margin):\n",
    "  def contrastive_loss(y_true, y_pred):\n",
    "    '''Contrastive loss from Hadsell-et-al.'06\n",
    "    http://yann.lecun.com/exdb/publis/pdf/hadsell-chopra-lecun-06.pdf\n",
    "    '''\n",
    "    square_pred = K.square(y_pred)\n",
    "    margin_square = K.square(K.maximum(margin - y_pred, 0))\n",
    "    return K.mean(y_true * square_pred + (1 - y_true) * margin_square)\n",
    "  return contrastive_loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Compile the model\n",
    "model.compile(\n",
    "  optimizer=optimizer,\n",
    "  loss=[contrastive_loss_with_margin(margin=1.0)]\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Shall we load the data?\n",
    "\n",
    "The data we will work with is the Fashion MNIST dataset, which is a dataset of 60,000 grayscale images of size 28x28 pixels.  \n",
    "\n",
    "The model we are trying to build should output the distance between the images, namely a pair of images from Fashion MNIST dataset; so we need to reformat the images into pairs with a label of 1 if they are from the same class and 0 otherwise.\n",
    "\n",
    "we want the data in the format\n",
    "\n",
    "image_1 | image_2 | label\n",
    "--------|---------|--------\n",
    "0.png   | 1.png   | 1\n",
    "0.png   | 2.png   | 0\n",
    " ....   | ....    | ...."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load the dataset\n",
    "from tensorflow.keras.datasets import fashion_mnist\n",
    "\n",
    "(train_X, train_y), (test_X, test_y) = fashion_mnist.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Some preprocessing\n",
    "def preprocess(x):\n",
    "  x = x.reshape(x.shape[0], 28, 28).astype('float32')\n",
    "  x /= 255\n",
    "  return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_X = preprocess(train_X)\n",
    "test_X = preprocess(test_X)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Making pairs\n",
    "\n",
    "Splitting the dataset into pairs should be an easy two nested loops, or may be you can do it easier, but anyway, this will give us more data than we need, and will be too time consuming, instead we will try to limit the number of pairs we have by some restrictions:\n",
    "\n",
    "1. if N is the size of smallest class, then we can have 2(N - 1) pairs for each class.\n",
    "2. Half of the pairs per class should be of the same class, hence will be labeled as 1 `match`.\n",
    "3. the other half of pairs will be made of one image of current class, and another from another randomly selected class, hence will be labeled as 0 `mismatch`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "# reformat the data into pairs of images\n",
    "def reformat(images, labels):\n",
    "  # get size of smallest class\n",
    "  min_class_size = min(np.bincount(labels)) - 1\n",
    "  print(min_class_size)\n",
    "  pairs = []            # list of image pairs\n",
    "  areMatching = []      # list of match/missMatch labels\n",
    "\n",
    "  class_indices = [np.where(labels == i)[0] for i in range(10)]\n",
    "    \n",
    "  for cur_class in range(10):\n",
    "    for i in range(min_class_size):\n",
    "      # the current image index is the ith image within the current class\n",
    "      cur_image_index = class_indices[cur_class][i]\n",
    "      # and the matching image is the one after it within the same class\n",
    "      matching_image_index = class_indices[cur_class][i + 1]\n",
    "\n",
    "      # add matching pairs\n",
    "      cur_image = images[cur_image_index]\n",
    "      matching_image = images[matching_image_index]\n",
    "\n",
    "      pairs.append([cur_image, matching_image])\n",
    "      areMatching.append(1)\n",
    "\n",
    "      # add non-matching images\n",
    "      random_class = np.random.randint(0, 9)\n",
    "\n",
    "      # random image index is the ith image within the random class, so that the result is deterministic\n",
    "      random_image_index = class_indices[random_class][i]\n",
    "      non_matching_image = images[random_image_index]\n",
    "\n",
    "      pairs.append([cur_image, non_matching_image])\n",
    "      areMatching.append(0)\n",
    "  \n",
    "  return np.array(pairs), np.array(areMatching, dtype=np.float32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5999\n",
      "999\n"
     ]
    }
   ],
   "source": [
    "# reformat the dataset into pairs of images\n",
    "train_pairs, train_distance = reformat(train_X, train_y)\n",
    "test_pairs, test_distance = reformat(test_X, test_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# # Show some examples of the training data\n",
    "# %matplotlib inline\n",
    "# import matplotlib.pyplot as plt\n",
    "# import matplotlib.gridspec as gridspec\n",
    "\n",
    "# # pick 4 random pairs from the training set\n",
    "# random_indices = np.random.randint(0, len(train_pairs), size=4)\n",
    "# random_pairs = train_pairs[random_indices]\n",
    "# random_distance = train_distance[random_indices]\n",
    "\n",
    "# fig = plt.figure(figsize=(20, 10))\n",
    "# outer = gridspec.GridSpec(2, 2, wspace=0.2, hspace=0.2)\n",
    "\n",
    "# for i in range(4):\n",
    "#   inner = gridspec.GridSpecFromSubplotSpec(1, 2,\n",
    "#                   subplot_spec=outer[i], wspace=0.1, hspace=0.1)\n",
    "\n",
    "#   for j in range(2):\n",
    "#     ax = plt.Subplot(fig, inner[j])\n",
    "    \n",
    "#     # show the image\n",
    "#     ax.imshow(random_pairs[i][j])\n",
    "\n",
    "#     # show the label\n",
    "#     ax.text(0, 0, '{}'.format(random_distance[i]),\n",
    "#             size=24, ha='center', va='center', color='w')\n",
    "\n",
    "#     ax.set_xticks([])\n",
    "#     ax.set_yticks([])\n",
    "#     fig.add_subplot(ax)\n",
    "\n",
    "\n",
    "\n",
    "# fig.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "<ipython-input-17-5ab368570464>:37: UserWarning: Matplotlib is currently using module://ipykernel.pylab.backend_inline, which is a non-GUI backend, so cannot show the figure.\n",
      "  fig.show()\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1440x720 with 8 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Show some examples of the training data\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.gridspec as gridspec\n",
    "\n",
    "# pick 4 random pairs from the training set\n",
    "random_indices = np.random.randint(0, len(train_pairs), size=4)\n",
    "random_pairs = train_pairs[random_indices]\n",
    "random_distance = train_distance[random_indices]\n",
    "\n",
    "fig = plt.figure(figsize=(20, 10))\n",
    "\n",
    "subFigs = fig.subfigures(2, 2).flatten()\n",
    "\n",
    "for i in range(4):\n",
    "  subFig = subFigs[i]\n",
    "  label = \"Match\" if random_distance[i] else \"miss-Match\"\n",
    "  color = \"green\" if random_distance[i] else \"red\"\n",
    "  subFig.supxlabel(label, fontsize=26, color=color)\n",
    "\n",
    "  axs = subFig.subplots(1, 2)\n",
    "\n",
    "  for j in range(2):\n",
    "    ax = axs[j]\n",
    "    \n",
    "    # show the image\n",
    "    ax.imshow(random_pairs[i][j])\n",
    "\n",
    "    # show the label\n",
    "    # ax.text(0, 0, '{}'.format(random_distance[i]),\n",
    "    #         size=24, ha='center', va='center', color='w')\n",
    "\n",
    "    ax.set_xticks([])\n",
    "    ax.set_yticks([])\n",
    "    subFig.add_subplot(ax)\n",
    "\n",
    "fig.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# How does a siamese network train\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/20\n",
      "938/938 [==============================] - 7s 7ms/step - loss: 0.1256 - val_loss: 0.1295\n",
      "Epoch 2/20\n",
      "938/938 [==============================] - 6s 7ms/step - loss: 0.1161 - val_loss: 0.1219\n",
      "Epoch 3/20\n",
      "938/938 [==============================] - 6s 7ms/step - loss: 0.1107 - val_loss: 0.1249\n",
      "Epoch 4/20\n",
      "938/938 [==============================] - 6s 7ms/step - loss: 0.1074 - val_loss: 0.1160\n",
      "Epoch 5/20\n",
      "938/938 [==============================] - 6s 7ms/step - loss: 0.1049 - val_loss: 0.1152\n",
      "Epoch 6/20\n",
      "938/938 [==============================] - 6s 7ms/step - loss: 0.1035 - val_loss: 0.1182\n",
      "Epoch 7/20\n",
      "938/938 [==============================] - 6s 7ms/step - loss: 0.1020 - val_loss: 0.1158\n",
      "Epoch 8/20\n",
      "938/938 [==============================] - 6s 7ms/step - loss: 0.1011 - val_loss: 0.1152\n",
      "Epoch 9/20\n",
      "938/938 [==============================] - 6s 7ms/step - loss: 0.1000 - val_loss: 0.1116\n",
      "Epoch 10/20\n",
      "938/938 [==============================] - 6s 7ms/step - loss: 0.0992 - val_loss: 0.1114\n",
      "Epoch 11/20\n",
      "938/938 [==============================] - 6s 7ms/step - loss: 0.0985 - val_loss: 0.1124\n",
      "Epoch 12/20\n",
      "938/938 [==============================] - 7s 7ms/step - loss: 0.0975 - val_loss: 0.1133\n",
      "Epoch 13/20\n",
      "938/938 [==============================] - 6s 7ms/step - loss: 0.0968 - val_loss: 0.1130\n",
      "Epoch 14/20\n",
      "938/938 [==============================] - 6s 7ms/step - loss: 0.0964 - val_loss: 0.1117\n",
      "Epoch 15/20\n",
      "938/938 [==============================] - 6s 7ms/step - loss: 0.0961 - val_loss: 0.1119\n",
      "Epoch 16/20\n",
      "938/938 [==============================] - 6s 7ms/step - loss: 0.0954 - val_loss: 0.1123\n",
      "Epoch 17/20\n",
      "938/938 [==============================] - 6s 7ms/step - loss: 0.0950 - val_loss: 0.1105\n",
      "Epoch 18/20\n",
      "938/938 [==============================] - 6s 7ms/step - loss: 0.0948 - val_loss: 0.1132\n",
      "Epoch 19/20\n",
      "938/938 [==============================] - 6s 7ms/step - loss: 0.0942 - val_loss: 0.1103\n",
      "Epoch 20/20\n",
      "938/938 [==============================] - 6s 7ms/step - loss: 0.0940 - val_loss: 0.1095\n"
     ]
    }
   ],
   "source": [
    "# Train the model\n",
    "history = model.fit(\n",
    "  x=[train_pairs[:, 0], train_pairs[:, 1]],\n",
    "  y=train_distance,\n",
    "  validation_data=([test_pairs[:, 0], test_pairs[:, 1]], test_distance),\n",
    "  batch_size=128,\n",
    "  epochs=20,\n",
    "  verbose=1\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "625/625 [==============================] - 1s 2ms/step - loss: 0.1095\n",
      "0.10946211963891983\n"
     ]
    }
   ],
   "source": [
    "# evaluate the model, don't forget the history!\n",
    "print(model.evaluate(x=[test_pairs[:, 0], test_pairs[:, 1]], y=test_distance, verbose=1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\as\\miniconda3\\envs\\TF-2.5\\lib\\site-packages\\matplotlib\\text.py:1215: FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will perform elementwise comparison\n",
      "  if s != self._text:\n",
      "<ipython-input-21-858dd6455cc9>:38: UserWarning: Matplotlib is currently using module://ipykernel.pylab.backend_inline, which is a non-GUI backend, so cannot show the figure.\n",
      "  fig.show()\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1440x720 with 20 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# let's do some plots!\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# pick 4 random pairs from the test set\n",
    "random_indices = np.random.randint(0, len(test_pairs), size=10)\n",
    "random_pairs = test_pairs[random_indices]\n",
    "\n",
    "# predict the distance for each pair\n",
    "predictions = model.predict([random_pairs[:, 0], random_pairs[:, 1]])\n",
    "\n",
    "# plot the results\n",
    "fig = plt.figure(figsize=(20, 10))\n",
    "\n",
    "subFigs = fig.subfigures(2, 5).flatten()\n",
    "\n",
    "for i in range(10):\n",
    "  subFig = subFigs[i]\n",
    "  distance = predictions[i]\n",
    "  color = \"green\" if distance < 0.5 else \"red\"\n",
    "  subFig.supxlabel(distance, fontsize=26, color=color)\n",
    "\n",
    "  axs = subFig.subplots(2, 1)\n",
    "\n",
    "  for j in range(2):\n",
    "    ax = axs[j]\n",
    "    \n",
    "    # show the image\n",
    "    ax.imshow(random_pairs[i][j])\n",
    "\n",
    "    # show the label\n",
    "    # ax.text(0, 0, '{}'.format(random_distance[i]),\n",
    "    #         size=24, ha='center', va='center', color='w')\n",
    "\n",
    "    ax.set_xticks([])\n",
    "    ax.set_yticks([])\n",
    "    subFig.add_subplot(ax)\n",
    "\n",
    "fig.show()"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "authorship_tag": "ABX9TyMUpFUBusWHNNPOpzczEvAw",
   "collapsed_sections": [],
   "include_colab_link": true,
   "name": "enenrgy+efficiency.ipynb",
   "provenance": []
  },
  "interpreter": {
   "hash": "70d077a54fc0094e614fa5b0f03887f72b81a6c521e79d0a3e89b1c317604ecf"
  },
  "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}