flower-classification-cnn.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np \n",
    "import itertools\n",
    "import keras\n",
    "from sklearn import metrics\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from keras.preprocessing.image import ImageDataGenerator, img_to_array, load_img \n",
    "from keras.models import Sequential \n",
    "from keras import optimizers\n",
    "from keras.preprocessing import image\n",
    "from keras.layers import Dropout, Flatten, Dense \n",
    "from keras import applications \n",
    "from keras.utils.np_utils import to_categorical \n",
    "import matplotlib.pyplot as plt \n",
    "import matplotlib.image as mpimg\n",
    "%matplotlib inline\n",
    "import math \n",
    "import datetime\n",
    "import time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# set default dimentions for the images\n",
    "img_width, img_height = 224, 224\n",
    "\n",
    "#create bottleneck file is conver all the img pixels into numpy array\n",
    "top_model_weights_path = 'bottleneck_fc_model.h5'\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "#loading datasets\n",
    "train_data_dir = 'data/train'\n",
    "validation_data_dir = 'data/validation'\n",
    "test_data_dir = 'data/test'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# number of epoch to train top model\n",
    "epoch = 7 #this has been changed after multiple model run\n",
    "\n",
    "\n",
    "#batch size used by flow_from_directory and predict_generator\n",
    "batch_size = 5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From C:\\Users\\LENOVO\\Anaconda3\\lib\\site-packages\\keras\\backend\\tensorflow_backend.py:74: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
      "\n",
      "WARNING:tensorflow:From C:\\Users\\LENOVO\\Anaconda3\\lib\\site-packages\\keras\\backend\\tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
      "\n",
      "WARNING:tensorflow:From C:\\Users\\LENOVO\\Anaconda3\\lib\\site-packages\\keras\\backend\\tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n",
      "\n",
      "WARNING:tensorflow:From C:\\Users\\LENOVO\\Anaconda3\\lib\\site-packages\\keras\\backend\\tensorflow_backend.py:3976: The name tf.nn.max_pool is deprecated. Please use tf.nn.max_pool2d instead.\n",
      "\n",
      "WARNING:tensorflow:From C:\\Users\\LENOVO\\Anaconda3\\lib\\site-packages\\keras\\backend\\tensorflow_backend.py:174: The name tf.get_default_session is deprecated. Please use tf.compat.v1.get_default_session instead.\n",
      "\n",
      "WARNING:tensorflow:From C:\\Users\\LENOVO\\Anaconda3\\lib\\site-packages\\keras\\backend\\tensorflow_backend.py:181: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "\n",
    "#Loading vgc16 model\n",
    "vgg16 = applications.VGG16(include_top=False, weights='imagenet')\n",
    "\n",
    "datagen = ImageDataGenerator(rescale=1. / 255) \n",
    "#needed to create the bottleneck .npy files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 4027 images belonging to 5 classes.\n",
      "Time:  0:30:12.160282\n"
     ]
    }
   ],
   "source": [
    "\n",
    "#__this can take an hour and half to run so only run it once. \n",
    "#once the npy files have been created, no need to run again. Convert this cell to a code cell to run.__\n",
    "\n",
    "start = datetime.datetime.now()\n",
    "   \n",
    "generator = datagen.flow_from_directory(  \n",
    "     train_data_dir,  \n",
    "     target_size=(img_width, img_height),  \n",
    "     batch_size=batch_size,  \n",
    "     class_mode=None,  \n",
    "     shuffle=False)  \n",
    "   \n",
    "nb_train_samples = len(generator.filenames)  \n",
    "num_classes = len(generator.class_indices)  \n",
    "   \n",
    "predict_size_train = int(math.ceil(nb_train_samples / batch_size))  \n",
    "   \n",
    "bottleneck_features_train = vgg16.predict_generator(generator, predict_size_train)  \n",
    "   \n",
    "np.save('bottleneck_features_train.npy', bottleneck_features_train)\n",
    "end= datetime.datetime.now()\n",
    "elapsed= end-start\n",
    "print ('Time: ', elapsed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "---------------------------------------------------------------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "print('-'*117)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 866 images belonging to 5 classes.\n",
      "Time:  0:06:29.752743\n"
     ]
    }
   ],
   "source": [
    "#__this can take half an hour to run so only run it once. once the npy files have been created, no need to run again. Convert this cell to a code cell to run.__\n",
    "\n",
    "start = datetime.datetime.now()\n",
    "generator = datagen.flow_from_directory(  \n",
    "     validation_data_dir,  \n",
    "     target_size=(img_width, img_height),  \n",
    "     batch_size=batch_size,  \n",
    "     class_mode=None,  \n",
    "     shuffle=False)  \n",
    "   \n",
    "nb_validation_samples = len(generator.filenames)  \n",
    "   \n",
    "predict_size_validation = int(math.ceil(nb_validation_samples / batch_size))  \n",
    "   \n",
    "bottleneck_features_validation = vgg16.predict_generator(  \n",
    "     generator, predict_size_validation)  \n",
    "   \n",
    "np.save('bottleneck_features_validation.npy', bottleneck_features_validation) \n",
    "end= datetime.datetime.now()\n",
    "elapsed= end-start\n",
    "print ('Time: ', elapsed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "---------------------------------------------------------------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "print('-'*117)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 1464 images belonging to 5 classes.\n",
      "Time:  0:11:02.561833\n"
     ]
    }
   ],
   "source": [
    "#__this can take half an hour to run so only run it once. once the npy files have been created, no need to run again. Convert this cell to a code cell to run.__\n",
    "\n",
    "start = datetime.datetime.now()\n",
    "generator = datagen.flow_from_directory(  \n",
    "     test_data_dir,  \n",
    "     target_size=(img_width, img_height),  \n",
    "     batch_size=batch_size,  \n",
    "     class_mode=None,  \n",
    "     shuffle=False)  \n",
    "   \n",
    "nb_test_samples = len(generator.filenames)  \n",
    "   \n",
    "predict_size_test = int(math.ceil(nb_test_samples / batch_size))  \n",
    "   \n",
    "bottleneck_features_test = vgg16.predict_generator(  \n",
    "     generator, predict_size_test)  \n",
    "   \n",
    "np.save('bottleneck_features_test.npy', bottleneck_features_test) \n",
    "end= datetime.datetime.now()\n",
    "elapsed= end-start\n",
    "print ('Time: ', elapsed)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# LOAD TRAINING, VALIDATION and TESTING data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 4027 images belonging to 5 classes.\n"
     ]
    }
   ],
   "source": [
    "#training data\n",
    "generator_top = datagen.flow_from_directory(  \n",
    "         train_data_dir,  \n",
    "         target_size=(img_width, img_height),  \n",
    "         batch_size=batch_size,  \n",
    "         class_mode='categorical',  \n",
    "         shuffle=False)  \n",
    "   \n",
    "nb_train_samples = len(generator_top.filenames)  \n",
    "num_classes = len(generator_top.class_indices)  \n",
    "   \n",
    "# load the bottleneck features saved earlier  \n",
    "train_data = np.load('bottleneck_features_train.npy')  \n",
    "   \n",
    "# get the class lebels for the training data, in the original order  \n",
    "train_labels = generator_top.classes  \n",
    "   \n",
    "# convert the training labels to categorical vectors  \n",
    "train_labels = to_categorical(train_labels, num_classes=num_classes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 866 images belonging to 5 classes.\n"
     ]
    }
   ],
   "source": [
    "#validation data\n",
    "generator_top = datagen.flow_from_directory(  \n",
    "         validation_data_dir,  \n",
    "         target_size=(img_width, img_height),  \n",
    "         batch_size=batch_size,  \n",
    "         class_mode=None,  \n",
    "         shuffle=False)  \n",
    "   \n",
    "nb_validation_samples = len(generator_top.filenames)  \n",
    "   \n",
    "validation_data = np.load('bottleneck_features_validation.npy')  \n",
    "   \n",
    "\n",
    "validation_labels = generator_top.classes  \n",
    "validation_labels = to_categorical(validation_labels, num_classes=num_classes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 1464 images belonging to 5 classes.\n"
     ]
    }
   ],
   "source": [
    "#testing data\n",
    "generator_top = datagen.flow_from_directory(  \n",
    "         test_data_dir,  \n",
    "         target_size=(img_width, img_height),  \n",
    "         batch_size=batch_size,  \n",
    "         class_mode=None,  \n",
    "         shuffle=False)  \n",
    "   \n",
    "nb_test_samples = len(generator_top.filenames)  \n",
    "   \n",
    "test_data = np.load('bottleneck_features_test.npy')  \n",
    "   \n",
    "\n",
    "test_labels = generator_top.classes  \n",
    "test_labels = to_categorical(test_labels, num_classes=num_classes)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Training of model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From C:\\Users\\LENOVO\\Anaconda3\\lib\\site-packages\\keras\\backend\\tensorflow_backend.py:3445: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n",
      "WARNING:tensorflow:From C:\\Users\\LENOVO\\Anaconda3\\lib\\site-packages\\keras\\optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\LENOVO\\Anaconda3\\lib\\site-packages\\keras\\activations.py:211: UserWarning: Do not pass a layer instance (such as LeakyReLU) as the activation argument of another layer. Instead, advanced activation layers should be used just like any other layer in a model.\n",
      "  identifier=identifier.__class__.__name__))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From C:\\Users\\LENOVO\\Anaconda3\\lib\\site-packages\\tensorflow\\python\\ops\\math_grad.py:1250: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.where in 2.0, which has the same broadcast rule as np.where\n",
      "Train on 4027 samples, validate on 866 samples\n",
      "Epoch 1/7\n",
      "4027/4027 [==============================] - 26s 7ms/step - loss: 1.1281 - acc: 0.5756 - val_loss: 0.5994 - val_acc: 0.7910\n",
      "Epoch 2/7\n",
      "4027/4027 [==============================] - 18s 5ms/step - loss: 0.7727 - acc: 0.7388 - val_loss: 0.4908 - val_acc: 0.8233\n",
      "Epoch 3/7\n",
      "4027/4027 [==============================] - 19s 5ms/step - loss: 0.6360 - acc: 0.7810 - val_loss: 0.4350 - val_acc: 0.8522\n",
      "Epoch 4/7\n",
      "4027/4027 [==============================] - 23s 6ms/step - loss: 0.5480 - acc: 0.8120 - val_loss: 0.3382 - val_acc: 0.8903\n",
      "Epoch 5/7\n",
      "4027/4027 [==============================] - 19s 5ms/step - loss: 0.4645 - acc: 0.8401 - val_loss: 0.2911 - val_acc: 0.9018\n",
      "Epoch 6/7\n",
      "4027/4027 [==============================] - 18s 5ms/step - loss: 0.4214 - acc: 0.8532 - val_loss: 0.3079 - val_acc: 0.8891\n",
      "Epoch 7/7\n",
      "4027/4027 [==============================] - 22s 5ms/step - loss: 0.3807 - acc: 0.8721 - val_loss: 0.2367 - val_acc: 0.9215\n",
      "866/866 [==============================] - 0s 391us/step\n",
      "[INFO] accuracy: 92.15%\n",
      "[INFO] Loss: 0.23665985567771813\n",
      "Time:  0:02:26.829966\n"
     ]
    }
   ],
   "source": [
    "#This is the best model we found. For additional models, check out I_notebook.ipynb\n",
    "start = datetime.datetime.now()\n",
    "model = Sequential()  \n",
    "model.add(Flatten(input_shape=train_data.shape[1:]))  \n",
    "model.add(Dense(100, activation=keras.layers.LeakyReLU(alpha=0.3)))  \n",
    "model.add(Dropout(0.5))  \n",
    "model.add(Dense(50, activation=keras.layers.LeakyReLU(alpha=0.3)))  \n",
    "model.add(Dropout(0.3)) \n",
    "model.add(Dense(num_classes, activation='softmax'))  \n",
    "\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer=optimizers.RMSprop(lr=1e-4),\n",
    "              metrics=['acc'])  \n",
    "\n",
    "history = model.fit(train_data, train_labels,  \n",
    "      epochs=7,\n",
    "      batch_size=batch_size,  \n",
    "      validation_data=(validation_data, validation_labels))  \n",
    "\n",
    "model.save_weights(top_model_weights_path)  \n",
    "\n",
    "(eval_loss, eval_accuracy) = model.evaluate(  \n",
    " validation_data, validation_labels, batch_size=batch_size, verbose=1)\n",
    "\n",
    "print(\"[INFO] accuracy: {:.2f}%\".format(eval_accuracy * 100))  \n",
    "print(\"[INFO] Loss: {}\".format(eval_loss))  \n",
    "end= datetime.datetime.now()\n",
    "elapsed= end-start\n",
    "print ('Time: ', elapsed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "flatten_1 (Flatten)          (None, 25088)             0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 100)               2508900   \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 100)               0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 50)                5050      \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 50)                0         \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 5)                 255       \n",
      "=================================================================\n",
      "Total params: 2,514,205\n",
      "Trainable params: 2,514,205\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#Model summary\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "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"
    },
    {
     "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": [
    "#Graphing our training and validation\n",
    "acc = history.history['acc']\n",
    "val_acc = history.history['val_acc']\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "epochs = range(len(acc))\n",
    "plt.plot(epochs, acc, 'r', label='Training acc')\n",
    "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.ylabel('accuracy')  \n",
    "plt.xlabel('epoch')\n",
    "plt.legend()\n",
    "plt.figure()\n",
    "plt.plot(epochs, loss, 'r', label='Training loss')\n",
    "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.ylabel('loss')  \n",
    "plt.xlabel('epoch')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1464/1464 [==============================] - 1s 374us/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.2890122901709353, 0.9016393442622951]"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.evaluate(test_data, test_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "test data [[[[0.00000000e+00 0.00000000e+00 6.02312207e-01 ... 1.96461514e-01\n",
      "    9.94315326e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 7.62413070e-02 ... 0.00000000e+00\n",
      "    1.12581837e+00 0.00000000e+00]\n",
      "   [1.58983786e-02 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    1.10688949e+00 0.00000000e+00]\n",
      "   ...\n",
      "   [0.00000000e+00 0.00000000e+00 5.35643585e-02 ... 0.00000000e+00\n",
      "    9.82706130e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 2.95575351e-01 ... 0.00000000e+00\n",
      "    1.45355356e+00 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 5.59204340e-01 ... 0.00000000e+00\n",
      "    1.36669838e+00 0.00000000e+00]]\n",
      "\n",
      "  [[2.92393982e-01 0.00000000e+00 3.09342712e-01 ... 4.28359032e-01\n",
      "    7.17160523e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 6.84435293e-02 ... 1.14029959e-01\n",
      "    1.09386301e+00 0.00000000e+00]\n",
      "   [7.66547740e-01 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    9.84859943e-01 0.00000000e+00]\n",
      "   ...\n",
      "   [9.29225460e-02 0.00000000e+00 3.15879732e-01 ... 0.00000000e+00\n",
      "    9.12516356e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    1.35863352e+00 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 2.34041959e-01 ... 0.00000000e+00\n",
      "    1.44619536e+00 0.00000000e+00]]\n",
      "\n",
      "  [[7.36100316e-01 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    6.20411634e-01 0.00000000e+00]\n",
      "   [3.73666465e-01 0.00000000e+00 5.64455867e-01 ... 0.00000000e+00\n",
      "    9.73024487e-01 0.00000000e+00]\n",
      "   [3.59219253e-01 0.00000000e+00 4.24971104e-01 ... 0.00000000e+00\n",
      "    7.89776087e-01 0.00000000e+00]\n",
      "   ...\n",
      "   [0.00000000e+00 0.00000000e+00 6.40887022e-01 ... 0.00000000e+00\n",
      "    9.27794993e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 4.63500917e-01 ... 0.00000000e+00\n",
      "    1.34525073e+00 0.00000000e+00]\n",
      "   [3.11801787e-02 0.00000000e+00 5.80436885e-01 ... 0.00000000e+00\n",
      "    1.60450864e+00 0.00000000e+00]]\n",
      "\n",
      "  ...\n",
      "\n",
      "  [[0.00000000e+00 0.00000000e+00 1.39810658e+00 ... 0.00000000e+00\n",
      "    1.87364817e-01 0.00000000e+00]\n",
      "   [9.14718747e-01 0.00000000e+00 9.70833123e-01 ... 0.00000000e+00\n",
      "    6.94830298e-01 0.00000000e+00]\n",
      "   [1.30648398e+00 0.00000000e+00 2.58002579e-01 ... 0.00000000e+00\n",
      "    5.58874249e-01 0.00000000e+00]\n",
      "   ...\n",
      "   [1.01620615e+00 0.00000000e+00 2.72896886e-01 ... 0.00000000e+00\n",
      "    5.91135502e-01 0.00000000e+00]\n",
      "   [8.96849394e-01 0.00000000e+00 1.21209204e+00 ... 0.00000000e+00\n",
      "    7.77776182e-01 0.00000000e+00]\n",
      "   [5.61451972e-01 0.00000000e+00 1.12301946e+00 ... 0.00000000e+00\n",
      "    1.02534091e+00 0.00000000e+00]]\n",
      "\n",
      "  [[0.00000000e+00 0.00000000e+00 5.75070381e-01 ... 0.00000000e+00\n",
      "    3.39955837e-01 0.00000000e+00]\n",
      "   [1.27157152e-01 0.00000000e+00 5.05763650e-01 ... 0.00000000e+00\n",
      "    6.99965119e-01 0.00000000e+00]\n",
      "   [4.16388899e-01 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    9.38457072e-01 0.00000000e+00]\n",
      "   ...\n",
      "   [2.21671425e-02 0.00000000e+00 4.17030334e-01 ... 0.00000000e+00\n",
      "    3.76008689e-01 0.00000000e+00]\n",
      "   [1.40519276e-01 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    2.27336660e-01 0.00000000e+00]\n",
      "   [1.20339975e-01 0.00000000e+00 5.26353791e-02 ... 0.00000000e+00\n",
      "    8.86999190e-01 0.00000000e+00]]\n",
      "\n",
      "  [[0.00000000e+00 0.00000000e+00 8.91470164e-02 ... 0.00000000e+00\n",
      "    9.22302008e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.37207896e-01 ... 0.00000000e+00\n",
      "    5.07726967e-01 0.00000000e+00]\n",
      "   [1.07378013e-01 0.00000000e+00 3.01991075e-01 ... 0.00000000e+00\n",
      "    1.26713336e+00 0.00000000e+00]\n",
      "   ...\n",
      "   [0.00000000e+00 0.00000000e+00 3.87214929e-01 ... 0.00000000e+00\n",
      "    8.16506803e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    6.51575685e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.31042287e-01 ... 0.00000000e+00\n",
      "    9.56049144e-01 0.00000000e+00]]]\n",
      "\n",
      "\n",
      " [[[0.00000000e+00 0.00000000e+00 1.63744643e-01 ... 0.00000000e+00\n",
      "    8.24313700e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.05089629e+00 ... 0.00000000e+00\n",
      "    7.46338129e-01 0.00000000e+00]\n",
      "   [1.92977805e-02 0.00000000e+00 9.22628045e-01 ... 0.00000000e+00\n",
      "    6.12195313e-01 0.00000000e+00]\n",
      "   ...\n",
      "   [0.00000000e+00 0.00000000e+00 3.24079901e-01 ... 0.00000000e+00\n",
      "    7.05191195e-01 0.00000000e+00]\n",
      "   [4.66569215e-02 0.00000000e+00 6.98145390e-01 ... 0.00000000e+00\n",
      "    5.13998508e-01 0.00000000e+00]\n",
      "   [5.17619513e-02 0.00000000e+00 2.32970744e-01 ... 0.00000000e+00\n",
      "    7.00614929e-01 0.00000000e+00]]\n",
      "\n",
      "  [[0.00000000e+00 0.00000000e+00 7.27600157e-01 ... 0.00000000e+00\n",
      "    6.47647619e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.28688979e+00 ... 0.00000000e+00\n",
      "    6.14665091e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.20061994e+00 ... 0.00000000e+00\n",
      "    5.55189312e-01 0.00000000e+00]\n",
      "   ...\n",
      "   [0.00000000e+00 0.00000000e+00 1.07678413e+00 ... 0.00000000e+00\n",
      "    1.74386665e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.27304077e+00 ... 0.00000000e+00\n",
      "    3.41448426e-01 3.51253897e-02]\n",
      "   [0.00000000e+00 0.00000000e+00 5.77638447e-01 ... 0.00000000e+00\n",
      "    6.28854036e-01 0.00000000e+00]]\n",
      "\n",
      "  [[0.00000000e+00 0.00000000e+00 6.90288961e-01 ... 0.00000000e+00\n",
      "    5.14543176e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.19314063e+00 ... 0.00000000e+00\n",
      "    5.29698312e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 7.95895398e-01 ... 0.00000000e+00\n",
      "    5.92150033e-01 0.00000000e+00]\n",
      "   ...\n",
      "   [0.00000000e+00 0.00000000e+00 6.61949575e-01 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 7.41784811e-01 ... 0.00000000e+00\n",
      "    6.39935017e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.09689355e+00 ... 0.00000000e+00\n",
      "    9.84457970e-01 0.00000000e+00]]\n",
      "\n",
      "  ...\n",
      "\n",
      "  [[3.99086595e-01 0.00000000e+00 4.29216981e-01 ... 0.00000000e+00\n",
      "    1.15286231e+00 0.00000000e+00]\n",
      "   [3.91632825e-01 0.00000000e+00 1.37190688e+00 ... 0.00000000e+00\n",
      "    2.70574540e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 8.19551110e-01 ... 0.00000000e+00\n",
      "    4.00886014e-02 0.00000000e+00]\n",
      "   ...\n",
      "   [1.03250265e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   [1.02370775e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   [6.63924694e-01 0.00000000e+00 6.56389236e-01 ... 0.00000000e+00\n",
      "    6.87474787e-01 0.00000000e+00]]\n",
      "\n",
      "  [[4.80567902e-01 0.00000000e+00 7.79882550e-01 ... 0.00000000e+00\n",
      "    1.11835825e+00 0.00000000e+00]\n",
      "   [5.68710566e-01 0.00000000e+00 1.82448113e+00 ... 0.00000000e+00\n",
      "    6.40903115e-01 0.00000000e+00]\n",
      "   [1.87225282e-01 0.00000000e+00 1.75477636e+00 ... 0.00000000e+00\n",
      "    4.88265693e-01 0.00000000e+00]\n",
      "   ...\n",
      "   [8.34046245e-01 0.00000000e+00 7.79555023e-01 ... 0.00000000e+00\n",
      "    7.44924545e-01 0.00000000e+00]\n",
      "   [3.72648776e-01 0.00000000e+00 1.86737394e+00 ... 0.00000000e+00\n",
      "    6.56993032e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 2.20197821e+00 ... 0.00000000e+00\n",
      "    6.67209864e-01 0.00000000e+00]]\n",
      "\n",
      "  [[5.63766733e-02 0.00000000e+00 1.57166982e+00 ... 0.00000000e+00\n",
      "    1.06041229e+00 0.00000000e+00]\n",
      "   [1.93001688e-01 0.00000000e+00 1.71140563e+00 ... 0.00000000e+00\n",
      "    8.10392618e-01 0.00000000e+00]\n",
      "   [1.60508648e-01 0.00000000e+00 1.78165925e+00 ... 0.00000000e+00\n",
      "    7.56978333e-01 0.00000000e+00]\n",
      "   ...\n",
      "   [6.89613163e-01 0.00000000e+00 1.26700377e+00 ... 0.00000000e+00\n",
      "    4.60856855e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 2.35172033e+00 ... 0.00000000e+00\n",
      "    6.12778127e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 2.67540073e+00 ... 0.00000000e+00\n",
      "    6.31675780e-01 0.00000000e+00]]]\n",
      "\n",
      "\n",
      " [[[1.85580060e-01 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    8.84982109e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 4.66149181e-01 ... 0.00000000e+00\n",
      "    1.00331914e+00 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 6.28812253e-01 ... 0.00000000e+00\n",
      "    8.51441145e-01 0.00000000e+00]\n",
      "   ...\n",
      "   [0.00000000e+00 0.00000000e+00 4.49060500e-01 ... 0.00000000e+00\n",
      "    6.23165727e-01 0.00000000e+00]\n",
      "   [3.78525108e-02 0.00000000e+00 5.14497399e-01 ... 0.00000000e+00\n",
      "    5.87730944e-01 0.00000000e+00]\n",
      "   [3.46905529e-01 0.00000000e+00 3.80038083e-01 ... 0.00000000e+00\n",
      "    6.06396139e-01 0.00000000e+00]]\n",
      "\n",
      "  [[0.00000000e+00 0.00000000e+00 3.39390375e-02 ... 0.00000000e+00\n",
      "    2.32400641e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 9.87422466e-01 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.11008799e+00 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   ...\n",
      "   [0.00000000e+00 0.00000000e+00 1.00527883e+00 ... 0.00000000e+00\n",
      "    3.50536466e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 8.97705495e-01 ... 0.00000000e+00\n",
      "    4.74255472e-01 0.00000000e+00]\n",
      "   [6.79931402e-01 0.00000000e+00 1.05566490e+00 ... 8.21788087e-02\n",
      "    7.54162192e-01 0.00000000e+00]]\n",
      "\n",
      "  [[0.00000000e+00 0.00000000e+00 1.11671619e-01 ... 0.00000000e+00\n",
      "    2.16856286e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.03737462e+00 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.22653878e+00 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   ...\n",
      "   [0.00000000e+00 0.00000000e+00 1.12237287e+00 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 5.61456263e-01 ... 0.00000000e+00\n",
      "    3.44775200e-01 0.00000000e+00]\n",
      "   [7.87608549e-02 0.00000000e+00 1.01041830e+00 ... 0.00000000e+00\n",
      "    9.38097775e-01 0.00000000e+00]]\n",
      "\n",
      "  ...\n",
      "\n",
      "  [[3.40605319e-01 0.00000000e+00 4.65791970e-02 ... 0.00000000e+00\n",
      "    8.01000237e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 5.05814373e-01 ... 0.00000000e+00\n",
      "    3.83442819e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   ...\n",
      "   [0.00000000e+00 0.00000000e+00 1.50925219e-01 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.56454459e-01 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   [3.28461558e-01 0.00000000e+00 4.85470630e-02 ... 0.00000000e+00\n",
      "    4.00207639e-01 0.00000000e+00]]\n",
      "\n",
      "  [[5.52365363e-01 0.00000000e+00 1.13530792e-01 ... 0.00000000e+00\n",
      "    1.03119707e+00 0.00000000e+00]\n",
      "   [5.14084995e-01 0.00000000e+00 9.39605162e-02 ... 0.00000000e+00\n",
      "    8.45687389e-01 0.00000000e+00]\n",
      "   [1.00732811e-01 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   ...\n",
      "   [4.39362973e-01 0.00000000e+00 3.66480425e-02 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   [1.46727234e-01 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    4.50715661e-01 0.00000000e+00]]\n",
      "\n",
      "  [[5.37155509e-01 0.00000000e+00 2.89796561e-01 ... 0.00000000e+00\n",
      "    9.83752072e-01 0.00000000e+00]\n",
      "   [4.30532724e-01 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    8.36084604e-01 0.00000000e+00]\n",
      "   [1.64133400e-01 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    4.53784794e-01 0.00000000e+00]\n",
      "   ...\n",
      "   [1.03145671e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    6.98687077e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    6.71702564e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    7.27319002e-01 0.00000000e+00]]]\n",
      "\n",
      "\n",
      " ...\n",
      "\n",
      "\n",
      " [[[0.00000000e+00 0.00000000e+00 4.43328917e-01 ... 0.00000000e+00\n",
      "    1.85624242e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.03217542e+00 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   [1.37921065e-01 0.00000000e+00 1.07494390e+00 ... 0.00000000e+00\n",
      "    1.65077224e-02 0.00000000e+00]\n",
      "   ...\n",
      "   [0.00000000e+00 0.00000000e+00 1.25288451e+00 ... 0.00000000e+00\n",
      "    1.43052235e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.36194313e+00 ... 0.00000000e+00\n",
      "    5.10023713e-01 0.00000000e+00]\n",
      "   [1.59684211e-01 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    5.69818497e-01 0.00000000e+00]]\n",
      "\n",
      "  [[9.05335665e-01 0.00000000e+00 8.34742785e-01 ... 1.15836710e-01\n",
      "    1.47857815e-01 0.00000000e+00]\n",
      "   [1.08214378e+00 0.00000000e+00 6.69787824e-01 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   [1.02337348e+00 0.00000000e+00 7.16852069e-01 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   ...\n",
      "   [7.34674931e-02 0.00000000e+00 1.80416489e+00 ... 9.37641636e-02\n",
      "    4.35314253e-02 5.28765738e-01]\n",
      "   [5.74775115e-02 0.00000000e+00 1.80481923e+00 ... 0.00000000e+00\n",
      "    5.57396829e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 3.98351163e-01 ... 0.00000000e+00\n",
      "    2.48000324e-01 0.00000000e+00]]\n",
      "\n",
      "  [[7.47910500e-01 0.00000000e+00 1.26289790e-02 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   [1.05739188e+00 0.00000000e+00 8.03497583e-02 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   [1.08466053e+00 0.00000000e+00 2.87839830e-01 ... 0.00000000e+00\n",
      "    1.98056945e-03 0.00000000e+00]\n",
      "   ...\n",
      "   [8.37054551e-01 0.00000000e+00 1.94169581e+00 ... 4.42416638e-01\n",
      "    0.00000000e+00 3.29632908e-01]\n",
      "   [7.22313523e-01 0.00000000e+00 1.71571040e+00 ... 6.26648590e-02\n",
      "    3.29723001e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 5.56894600e-01 ... 0.00000000e+00\n",
      "    1.03897769e-02 0.00000000e+00]]\n",
      "\n",
      "  ...\n",
      "\n",
      "  [[0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    6.82451427e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    9.55807567e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    6.26205266e-01 0.00000000e+00]\n",
      "   ...\n",
      "   [9.28180754e-01 0.00000000e+00 8.68633211e-01 ... 0.00000000e+00\n",
      "    2.51636952e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 6.21538460e-01 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]]\n",
      "\n",
      "  [[0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    7.13569760e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 7.07962066e-02 ... 0.00000000e+00\n",
      "    8.43208134e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    7.60911763e-01 0.00000000e+00]\n",
      "   ...\n",
      "   [0.00000000e+00 0.00000000e+00 1.04461920e+00 ... 0.00000000e+00\n",
      "    5.07076383e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 6.88630044e-01 ... 0.00000000e+00\n",
      "    2.42919222e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    1.38951913e-01 0.00000000e+00]]\n",
      "\n",
      "  [[0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    9.90817666e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 5.36929145e-02 ... 0.00000000e+00\n",
      "    9.29165661e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    5.91655612e-01 0.00000000e+00]\n",
      "   ...\n",
      "   [0.00000000e+00 0.00000000e+00 1.54999673e-01 ... 0.00000000e+00\n",
      "    1.14548218e+00 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 2.23290935e-01 ... 0.00000000e+00\n",
      "    6.82767391e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    5.90683818e-01 0.00000000e+00]]]\n",
      "\n",
      "\n",
      " [[[1.68633014e-01 0.00000000e+00 8.61998558e-01 ... 0.00000000e+00\n",
      "    9.27316785e-01 0.00000000e+00]\n",
      "   [2.43124396e-01 0.00000000e+00 5.74798405e-01 ... 0.00000000e+00\n",
      "    9.86201286e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.54755580e+00 ... 0.00000000e+00\n",
      "    1.07207942e+00 0.00000000e+00]\n",
      "   ...\n",
      "   [0.00000000e+00 0.00000000e+00 3.72738123e-01 ... 0.00000000e+00\n",
      "    2.42955595e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.99467695e+00 ... 0.00000000e+00\n",
      "    1.40089318e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 2.38902545e+00 ... 2.23500729e-01\n",
      "    3.55495960e-01 1.22124456e-01]]\n",
      "\n",
      "  [[0.00000000e+00 0.00000000e+00 7.90940762e-01 ... 0.00000000e+00\n",
      "    1.06418538e+00 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.79599345e-01 ... 0.00000000e+00\n",
      "    7.55582809e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.93325090e+00 ... 1.58432603e-01\n",
      "    9.80748296e-01 0.00000000e+00]\n",
      "   ...\n",
      "   [9.33542997e-02 0.00000000e+00 2.64923835e+00 ... 0.00000000e+00\n",
      "    2.18846709e-01 2.49037519e-01]\n",
      "   [0.00000000e+00 0.00000000e+00 2.45201635e+00 ... 0.00000000e+00\n",
      "    3.33429198e-03 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 2.45820045e+00 ... 2.01261818e-01\n",
      "    2.83813149e-01 0.00000000e+00]]\n",
      "\n",
      "  [[0.00000000e+00 0.00000000e+00 4.72815633e-01 ... 0.00000000e+00\n",
      "    9.23390985e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.37038603e-01 ... 0.00000000e+00\n",
      "    5.28170168e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.29586864e+00 ... 0.00000000e+00\n",
      "    7.39466131e-01 0.00000000e+00]\n",
      "   ...\n",
      "   [0.00000000e+00 0.00000000e+00 2.14086676e+00 ... 0.00000000e+00\n",
      "    5.57321757e-02 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.97639287e+00 ... 0.00000000e+00\n",
      "    4.39569503e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.63565695e+00 ... 0.00000000e+00\n",
      "    2.97794849e-01 0.00000000e+00]]\n",
      "\n",
      "  ...\n",
      "\n",
      "  [[0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    4.64761347e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 3.74768764e-01 ... 0.00000000e+00\n",
      "    2.74377137e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 2.33427143e+00 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   ...\n",
      "   [2.91980128e-03 0.00000000e+00 1.43679607e+00 ... 0.00000000e+00\n",
      "    3.03911954e-01 0.00000000e+00]\n",
      "   [3.52033347e-01 0.00000000e+00 6.29391253e-01 ... 0.00000000e+00\n",
      "    3.97366256e-01 0.00000000e+00]\n",
      "   [7.12487817e-01 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    6.82628572e-01 0.00000000e+00]]\n",
      "\n",
      "  [[0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    6.51545644e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 4.07497257e-01 ... 0.00000000e+00\n",
      "    4.30874079e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 8.67068827e-01 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   ...\n",
      "   [2.86989659e-01 0.00000000e+00 2.47161460e+00 ... 0.00000000e+00\n",
      "    5.00600398e-01 0.00000000e+00]\n",
      "   [1.92691728e-01 0.00000000e+00 2.20639467e+00 ... 0.00000000e+00\n",
      "    6.82005823e-01 0.00000000e+00]\n",
      "   [2.95645650e-02 0.00000000e+00 6.29915893e-02 ... 0.00000000e+00\n",
      "    4.16909486e-01 0.00000000e+00]]\n",
      "\n",
      "  [[0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    9.36613560e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    1.00579894e+00 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 4.95477915e-01 ... 0.00000000e+00\n",
      "    6.59439325e-01 0.00000000e+00]\n",
      "   ...\n",
      "   [4.69772726e-01 0.00000000e+00 1.60395157e+00 ... 0.00000000e+00\n",
      "    9.07361448e-01 0.00000000e+00]\n",
      "   [3.50840271e-01 0.00000000e+00 1.34495544e+00 ... 0.00000000e+00\n",
      "    1.00947201e+00 0.00000000e+00]\n",
      "   [3.18437696e-01 0.00000000e+00 2.64511377e-01 ... 0.00000000e+00\n",
      "    7.20502615e-01 0.00000000e+00]]]\n",
      "\n",
      "\n",
      " [[[4.06998783e-01 0.00000000e+00 2.23223925e+00 ... 0.00000000e+00\n",
      "    1.32926857e+00 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.21636772e+00 ... 0.00000000e+00\n",
      "    7.46140540e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 3.29853445e-01 ... 0.00000000e+00\n",
      "    9.60217297e-01 0.00000000e+00]\n",
      "   ...\n",
      "   [0.00000000e+00 0.00000000e+00 1.07936990e+00 ... 0.00000000e+00\n",
      "    5.33394873e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 5.38675666e-01 ... 0.00000000e+00\n",
      "    9.03544724e-01 0.00000000e+00]\n",
      "   [2.27869913e-01 0.00000000e+00 4.68942858e-02 ... 0.00000000e+00\n",
      "    1.15583038e+00 0.00000000e+00]]\n",
      "\n",
      "  [[1.23030767e-01 0.00000000e+00 1.38803065e+00 ... 0.00000000e+00\n",
      "    1.32390380e+00 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 6.11472964e-01 ... 0.00000000e+00\n",
      "    8.27341795e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 4.51548934e-01 ... 0.00000000e+00\n",
      "    7.42285907e-01 0.00000000e+00]\n",
      "   ...\n",
      "   [0.00000000e+00 0.00000000e+00 3.52875471e-01 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.41688120e+00 ... 0.00000000e+00\n",
      "    7.41718471e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.06698501e+00 ... 1.65655524e-01\n",
      "    1.15901315e+00 0.00000000e+00]]\n",
      "\n",
      "  [[1.71840861e-01 0.00000000e+00 1.87532425e+00 ... 0.00000000e+00\n",
      "    6.00197077e-01 0.00000000e+00]\n",
      "   [2.08049774e-01 0.00000000e+00 9.62077677e-01 ... 0.00000000e+00\n",
      "    6.45319372e-02 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 4.73302007e-01 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   ...\n",
      "   [0.00000000e+00 0.00000000e+00 1.08351290e+00 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   [8.89165774e-02 0.00000000e+00 2.37306809e+00 ... 0.00000000e+00\n",
      "    1.38784796e-01 0.00000000e+00]\n",
      "   [5.76902628e-01 0.00000000e+00 2.43375230e+00 ... 0.00000000e+00\n",
      "    5.92017055e-01 0.00000000e+00]]\n",
      "\n",
      "  ...\n",
      "\n",
      "  [[0.00000000e+00 0.00000000e+00 4.09975111e-01 ... 0.00000000e+00\n",
      "    1.24330878e+00 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.10736184e-01 ... 0.00000000e+00\n",
      "    1.12172103e+00 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 1.96402818e-01 ... 0.00000000e+00\n",
      "    1.43799067e+00 0.00000000e+00]\n",
      "   ...\n",
      "   [4.87743497e-01 0.00000000e+00 7.58727565e-02 ... 0.00000000e+00\n",
      "    0.00000000e+00 0.00000000e+00]\n",
      "   [5.42009532e-01 0.00000000e+00 6.38387620e-01 ... 0.00000000e+00\n",
      "    7.96588659e-01 0.00000000e+00]\n",
      "   [6.57770038e-01 0.00000000e+00 1.23217463e+00 ... 6.74811017e-05\n",
      "    5.62554777e-01 0.00000000e+00]]\n",
      "\n",
      "  [[0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    1.33076048e+00 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    1.08781779e+00 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    1.40518045e+00 0.00000000e+00]\n",
      "   ...\n",
      "   [0.00000000e+00 0.00000000e+00 7.25166976e-01 ... 6.04034811e-02\n",
      "    8.73281658e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 7.00969577e-01 ... 0.00000000e+00\n",
      "    9.17458057e-01 0.00000000e+00]\n",
      "   [1.21101409e-01 0.00000000e+00 1.03402007e+00 ... 0.00000000e+00\n",
      "    7.07213402e-01 0.00000000e+00]]\n",
      "\n",
      "  [[0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    1.17509520e+00 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    8.84068310e-01 0.00000000e+00]\n",
      "   [0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    1.09978878e+00 0.00000000e+00]\n",
      "   ...\n",
      "   [0.00000000e+00 0.00000000e+00 0.00000000e+00 ... 0.00000000e+00\n",
      "    9.92509007e-01 0.00000000e+00]\n",
      "   [6.65033981e-02 0.00000000e+00 6.85854912e-01 ... 0.00000000e+00\n",
      "    7.01246023e-01 0.00000000e+00]\n",
      "   [1.99218392e-01 0.00000000e+00 8.34564686e-01 ... 0.00000000e+00\n",
      "    8.03791344e-01 0.00000000e+00]]]]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rounded test_labels [[1. 0. 0. 0. 0.]\n",
      " [1. 0. 0. 0. 0.]\n",
      " [1. 0. 0. 0. 0.]\n",
      " ...\n",
      " [0. 0. 0. 0. 1.]\n",
      " [0. 0. 0. 0. 1.]\n",
      " [0. 0. 0. 0. 1.]]\n"
     ]
    }
   ],
   "source": [
    "print('test data', test_data)\n",
    "preds = np.round(model.predict(test_data),0) \n",
    "#to fit them into classification metrics and confusion metrics, some additional modificaitions are required\n",
    "print('rounded test_labels', preds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "       daisy       0.97      0.83      0.90       276\n",
      "   dandelion       0.94      0.90      0.92       330\n",
      "        rose       0.89      0.88      0.89       280\n",
      "   sunflower       0.96      0.89      0.92       265\n",
      "       tulip       0.86      0.93      0.89       313\n",
      "\n",
      "   micro avg       0.92      0.89      0.90      1464\n",
      "   macro avg       0.93      0.89      0.90      1464\n",
      "weighted avg       0.92      0.89      0.90      1464\n",
      " samples avg       0.89      0.89      0.89      1464\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\LENOVO\\Anaconda3\\lib\\site-packages\\sklearn\\metrics\\classification.py:1437: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in samples with no predicted labels.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    }
   ],
   "source": [
    "flowers = ['daisy', 'dandelion', 'rose', 'sunflower', 'tulip']\n",
    "classification_metrics = metrics.classification_report(test_labels, preds, target_names=flowers )\n",
    "print(classification_metrics)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# confution matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Since our data is in dummy format we put the numpy array into a dataframe and call idxmax axis=1 to return the column\n",
    "# label of the maximum value thus creating a categorical variable\n",
    "#Basically, flipping a dummy variable back to it's categorical variable\n",
    "categorical_test_labels = pd.DataFrame(test_labels).idxmax(axis=1)\n",
    "categorical_preds = pd.DataFrame(preds).idxmax(axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "confusion_matrix= confusion_matrix(categorical_test_labels, categorical_preds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "#To get better visual of the confusion matrix:\n",
    "def plot_confusion_matrix(cm, classes,\n",
    "             normalize=False,\n",
    "             title='Confusion matrix',\n",
    "             cmap=plt.cm.Blues):\n",
    "    #Add Normalization Option\n",
    "    '''prints pretty confusion metric with normalization option '''\n",
    "    if normalize:\n",
    "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "        print(\"Normalized confusion matrix\")\n",
    "    else:\n",
    "        print('Confusion matrix, without normalization')\n",
    "    \n",
    "#     print(cm)\n",
    "    \n",
    "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
    "    plt.title(title)\n",
    "    plt.colorbar()\n",
    "    tick_marks = np.arange(len(classes))\n",
    "    plt.xticks(tick_marks, classes, rotation=45)\n",
    "    plt.yticks(tick_marks, classes)\n",
    "    \n",
    "    fmt = '.2f' if normalize else 'd'\n",
    "    thresh = cm.max() / 2.\n",
    "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
    "        plt.text(j, i, format(cm[i, j], fmt), horizontalalignment=\"center\", color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "    \n",
    "    plt.tight_layout()\n",
    "    plt.ylabel('True label')\n",
    "    plt.xlabel('Predicted label')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix, without normalization\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": [
    "plot_confusion_matrix(confusion_matrix, ['daisy', 'dandelion', 'rose', 'sunflower', 'tulip'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Normalized confusion matrix\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": [
    "#Those numbers are all over the place. Now turning normalize= True\n",
    "plot_confusion_matrix(confusion_matrix, \n",
    "                      ['daisy', 'dandelion', 'rose', 'sunflower', 'tulip'],\n",
    "                     normalize=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Testing images on model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "def read_image(file_path):\n",
    "    print(\"[INFO] loading and preprocessing image...\")  \n",
    "    image = load_img(file_path, target_size=(224, 224))  \n",
    "    image = img_to_array(image)  \n",
    "    image = np.expand_dims(image, axis=0)\n",
    "    image /= 255.  \n",
    "    return image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "def test_single_image(path):\n",
    "    flowers = ['daisy', 'dandelion', 'rose', 'sunflower', 'tulip']\n",
    "    images = read_image(path)\n",
    "    time.sleep(.5)\n",
    "    bt_prediction = vgg16.predict(images)  \n",
    "    preds = model.predict_proba(bt_prediction)\n",
    "    for idx, flower, x in zip(range(0,6), flowers , preds[0]):\n",
    "        print(\"ID: {}, Label: {} {}%\".format(idx, flower, round(x*100,2) ))\n",
    "    print('Final Decision:')\n",
    "    time.sleep(.5)\n",
    "    for x in range(3):\n",
    "        print('.'*(x+1))\n",
    "        time.sleep(.2)\n",
    "    class_predicted = model.predict_classes(bt_prediction)\n",
    "    class_dictionary = generator_top.class_indices  \n",
    "    inv_map = {v: k for k, v in class_dictionary.items()}  \n",
    "    print(\"ID: {}, Label: {}\".format(class_predicted[0], inv_map[class_predicted[0]]))  \n",
    "    return load_img(path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "path = 'data/test/dog.jpg'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[INFO] loading and preprocessing image...\n",
      "ID: 0, Label: daisy 3.64%\n",
      "ID: 1, Label: dandelion 3.14%\n",
      "ID: 2, Label: rose 89.91%\n",
      "ID: 3, Label: sunflower 1.52%\n",
      "ID: 4, Label: tulip 1.79%\n",
      "Final Decision:\n",
      ".\n",
      "..\n",
      "...\n",
      "ID: 2, Label: rose\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<PIL.JpegImagePlugin.JpegImageFile image mode=RGB size=275x183 at 0x1A9F52BC048>"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_single_image(path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n"
   ]
  }
 ],
 "metadata": {
  "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": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}