IndusNet_7 classes_100_100_1st Experiment.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "resized_images_100_100_train_dir='D:\\\\TensorFlow Programs\\\\fruit_image_classification_Indus Fruit Dataset\\\\Indus Fruit Dataset_Resized Images_100_100\\\\Train'\n",
    "resized_images_100_100_test_dir='D:\\\\TensorFlow Programs\\\\fruit_image_classification_Indus Fruit Dataset\\\\Indus Fruit Dataset_Resized Images_100_100\\\\Test'\n",
    "resized_images_100_100_validation_dir='D:\\\\TensorFlow Programs\\\\fruit_image_classification_Indus Fruit Dataset\\\\Indus Fruit Dataset_Resized Images_100_100\\\\Validation'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From C:\\Users\\Raheel\\Anaconda3\\envs\\EnvName\\lib\\site-packages\\tensorflow\\python\\ops\\resource_variable_ops.py:435: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Colocations handled automatically by placer.\n",
      "WARNING:tensorflow:From C:\\Users\\Raheel\\Anaconda3\\envs\\EnvName\\lib\\site-packages\\tensorflow\\python\\keras\\layers\\core.py:143: 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",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d (Conv2D)              (None, 98, 98, 128)       3584      \n",
      "_________________________________________________________________\n",
      "conv2d_1 (Conv2D)            (None, 96, 96, 128)       147584    \n",
      "_________________________________________________________________\n",
      "conv2d_2 (Conv2D)            (None, 94, 94, 256)       295168    \n",
      "_________________________________________________________________\n",
      "conv2d_3 (Conv2D)            (None, 92, 92, 256)       590080    \n",
      "_________________________________________________________________\n",
      "max_pooling2d (MaxPooling2D) (None, 46, 46, 256)       0         \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 46, 46, 256)       0         \n",
      "_________________________________________________________________\n",
      "conv2d_4 (Conv2D)            (None, 44, 44, 256)       590080    \n",
      "_________________________________________________________________\n",
      "conv2d_5 (Conv2D)            (None, 42, 42, 256)       590080    \n",
      "_________________________________________________________________\n",
      "conv2d_6 (Conv2D)            (None, 40, 40, 512)       1180160   \n",
      "_________________________________________________________________\n",
      "conv2d_7 (Conv2D)            (None, 38, 38, 512)       2359808   \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 19, 19, 512)       0         \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 19, 19, 512)       0         \n",
      "_________________________________________________________________\n",
      "conv2d_8 (Conv2D)            (None, 17, 17, 512)       2359808   \n",
      "_________________________________________________________________\n",
      "conv2d_9 (Conv2D)            (None, 15, 15, 1024)      4719616   \n",
      "_________________________________________________________________\n",
      "max_pooling2d_2 (MaxPooling2 (None, 7, 7, 1024)        0         \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 7, 7, 1024)        0         \n",
      "_________________________________________________________________\n",
      "flatten (Flatten)            (None, 50176)             0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 32)                1605664   \n",
      "_________________________________________________________________\n",
      "dropout_3 (Dropout)          (None, 32)                0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 7)                 231       \n",
      "=================================================================\n",
      "Total params: 14,441,863\n",
      "Trainable params: 14,441,863\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import keras\n",
    "from tensorflow.python.keras import models\n",
    "from tensorflow.python.keras import layers\n",
    "from tensorflow.python.keras.preprocessing.image import ImageDataGenerator\n",
    "from tensorflow.python.keras import regularizers\n",
    "from tensorflow.python.keras import optimizers\n",
    "from tensorflow.python.keras.callbacks import ModelCheckpoint\n",
    "\n",
    "indusNetModel=models.Sequential()\n",
    "\n",
    "indusNetModel.add(layers.Conv2D(128, (3, 3), activation='relu', input_shape=(100, 100, 3)))\n",
    "indusNetModel.add(layers.Conv2D(128, (3, 3), activation='relu'))\n",
    "indusNetModel.add(layers.Conv2D(256, (3, 3), activation='relu'))\n",
    "indusNetModel.add(layers.Conv2D(256, (3, 3), activation='relu'))\n",
    "indusNetModel.add(layers.MaxPooling2D(pool_size=(2, 2)))\n",
    "indusNetModel.add(layers.Dropout(0.2))\n",
    "\n",
    "indusNetModel.add(layers.Conv2D(256, (3, 3), activation='relu'))\n",
    "indusNetModel.add(layers.Conv2D(256, (3, 3), activation='relu'))\n",
    "indusNetModel.add(layers.Conv2D(512, (3, 3), activation='relu'))\n",
    "indusNetModel.add(layers.Conv2D(512, (3, 3), activation='relu'))\n",
    "indusNetModel.add(layers.MaxPooling2D(pool_size=(2, 2)))\n",
    "indusNetModel.add(layers.Dropout(0.2))\n",
    "\n",
    "indusNetModel.add(layers.Conv2D(512, (3, 3), activation='relu'))\n",
    "indusNetModel.add(layers.Conv2D(1024, (3, 3), activation='relu'))\n",
    "indusNetModel.add(layers.MaxPooling2D(pool_size=(2, 2)))\n",
    "indusNetModel.add(layers.Dropout(0.2))\n",
    "\n",
    "indusNetModel.add(layers.Flatten())\n",
    "indusNetModel.add(layers.Dense(32, kernel_regularizer=regularizers.l2(0.001), activation='relu'))\n",
    "indusNetModel.add(layers.Dropout(0.2))\n",
    "indusNetModel.add(layers.Dense(7, activation='softmax'))\n",
    "\n",
    "indusNetModel.compile(loss='categorical_crossentropy',optimizer=optimizers.Adam(lr=1e-4),metrics=['categorical_accuracy'])\n",
    "indusNetModel.summary()\n",
    "\n",
    "filepath=\"IndusNet_7_classes_100_100_1st Experiment.h5\"\n",
    "checkpoint = ModelCheckpoint(filepath, save_best_only=True)\n",
    "callbacks_list = [checkpoint]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 2800 images belonging to 7 classes.\n",
      "Found 280 images belonging to 7 classes.\n",
      "Found 560 images belonging to 7 classes.\n"
     ]
    }
   ],
   "source": [
    "batch_size=16\n",
    "\n",
    "train_datagen = ImageDataGenerator(\n",
    "    rescale=1./255,\n",
    "    rotation_range=40,\n",
    "    width_shift_range=0.2,\n",
    "    height_shift_range=0.2,\n",
    "    shear_range=0.2,\n",
    "    zoom_range=0.2,\n",
    "    horizontal_flip=True,)\n",
    "\n",
    "# Note that the validation  and test data should not be augmented!\n",
    "validation_datagen=ImageDataGenerator(rescale=1./255)\n",
    "\n",
    "test_datagen = ImageDataGenerator(rescale=1./255)\n",
    "\n",
    "train_generator = train_datagen.flow_from_directory(\n",
    "        # This is the target directory\n",
    "        resized_images_100_100_train_dir,\n",
    "        target_size=(100, 100),\n",
    "        batch_size=batch_size)\n",
    "\n",
    "validation_generator = validation_datagen.flow_from_directory(\n",
    "        resized_images_100_100_validation_dir,\n",
    "        target_size=(100, 100),\n",
    "        batch_size=batch_size, shuffle=False)\n",
    "\n",
    "test_generator = test_datagen.flow_from_directory(\n",
    "        resized_images_100_100_test_dir,\n",
    "        target_size=(100, 100),\n",
    "        batch_size=batch_size, shuffle=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "number_of_training_samples=2800\n",
    "number_of_validation_samples=280\n",
    "number_of_test_samples=560"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From C:\\Users\\Raheel\\Anaconda3\\envs\\EnvName\\lib\\site-packages\\tensorflow\\python\\ops\\math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.cast instead.\n",
      "Epoch 1/400\n",
      "18/18 [==============================] - 5s 255ms/step - loss: 1.7376 - categorical_accuracy: 0.2714\n",
      "175/175 [==============================] - 100s 570ms/step - loss: 1.9152 - categorical_accuracy: 0.1946 - val_loss: 1.7376 - val_categorical_accuracy: 0.2714\n",
      "Epoch 2/400\n",
      "18/18 [==============================] - 3s 170ms/step - loss: 1.6725 - categorical_accuracy: 0.3071\n",
      "175/175 [==============================] - 90s 517ms/step - loss: 1.7647 - categorical_accuracy: 0.2604 - val_loss: 1.6725 - val_categorical_accuracy: 0.3071\n",
      "Epoch 3/400\n",
      "18/18 [==============================] - 3s 171ms/step - loss: 1.6420 - categorical_accuracy: 0.3000\n",
      "175/175 [==============================] - 91s 520ms/step - loss: 1.7194 - categorical_accuracy: 0.2804 - val_loss: 1.6420 - val_categorical_accuracy: 0.3000\n",
      "Epoch 4/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 1.6066 - categorical_accuracy: 0.2964\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 1.6742 - categorical_accuracy: 0.2911 - val_loss: 1.6066 - val_categorical_accuracy: 0.2964\n",
      "Epoch 5/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 1.5353 - categorical_accuracy: 0.3393\n",
      "175/175 [==============================] - 91s 517ms/step - loss: 1.6376 - categorical_accuracy: 0.3029 - val_loss: 1.5353 - val_categorical_accuracy: 0.3393\n",
      "Epoch 6/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 1.4574 - categorical_accuracy: 0.3821\n",
      "175/175 [==============================] - 90s 517ms/step - loss: 1.6143 - categorical_accuracy: 0.3261 - val_loss: 1.4574 - val_categorical_accuracy: 0.3821\n",
      "Epoch 7/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 1.3899 - categorical_accuracy: 0.4179\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 1.4815 - categorical_accuracy: 0.3907 - val_loss: 1.3899 - val_categorical_accuracy: 0.4179\n",
      "Epoch 8/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 1.3498 - categorical_accuracy: 0.4393\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 1.4137 - categorical_accuracy: 0.4061 - val_loss: 1.3498 - val_categorical_accuracy: 0.4393\n",
      "Epoch 9/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 1.3540 - categorical_accuracy: 0.4179\n",
      "175/175 [==============================] - 89s 510ms/step - loss: 1.3941 - categorical_accuracy: 0.4189 - val_loss: 1.3540 - val_categorical_accuracy: 0.4179\n",
      "Epoch 10/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 1.2093 - categorical_accuracy: 0.4679\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 1.3576 - categorical_accuracy: 0.4279 - val_loss: 1.2093 - val_categorical_accuracy: 0.4679\n",
      "Epoch 11/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 1.1931 - categorical_accuracy: 0.5000\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 1.3382 - categorical_accuracy: 0.4393 - val_loss: 1.1931 - val_categorical_accuracy: 0.5000\n",
      "Epoch 12/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 1.1802 - categorical_accuracy: 0.5536\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 1.2966 - categorical_accuracy: 0.4693 - val_loss: 1.1802 - val_categorical_accuracy: 0.5536\n",
      "Epoch 13/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 1.0877 - categorical_accuracy: 0.5571\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 1.2829 - categorical_accuracy: 0.4750 - val_loss: 1.0877 - val_categorical_accuracy: 0.5571\n",
      "Epoch 14/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 1.2743 - categorical_accuracy: 0.5643\n",
      "175/175 [==============================] - 89s 510ms/step - loss: 1.2276 - categorical_accuracy: 0.5111 - val_loss: 1.2743 - val_categorical_accuracy: 0.5643\n",
      "Epoch 15/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 1.1385 - categorical_accuracy: 0.5893\n",
      "175/175 [==============================] - 89s 510ms/step - loss: 1.2453 - categorical_accuracy: 0.4986 - val_loss: 1.1385 - val_categorical_accuracy: 0.5893\n",
      "Epoch 16/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 1.1110 - categorical_accuracy: 0.5714\n",
      "175/175 [==============================] - 89s 510ms/step - loss: 1.2010 - categorical_accuracy: 0.5136 - val_loss: 1.1110 - val_categorical_accuracy: 0.5714\n",
      "Epoch 17/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 1.0839 - categorical_accuracy: 0.5857\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 1.1768 - categorical_accuracy: 0.5382 - val_loss: 1.0839 - val_categorical_accuracy: 0.5857\n",
      "Epoch 18/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 1.0393 - categorical_accuracy: 0.5964\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 1.1490 - categorical_accuracy: 0.5368 - val_loss: 1.0393 - val_categorical_accuracy: 0.5964\n",
      "Epoch 19/400\n",
      "18/18 [==============================] - 3s 164ms/step - loss: 1.0288 - categorical_accuracy: 0.5964\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 1.1201 - categorical_accuracy: 0.5379 - val_loss: 1.0288 - val_categorical_accuracy: 0.5964\n",
      "Epoch 20/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.9850 - categorical_accuracy: 0.6179\n",
      "175/175 [==============================] - 91s 519ms/step - loss: 1.1043 - categorical_accuracy: 0.5768 - val_loss: 0.9850 - val_categorical_accuracy: 0.6179\n",
      "Epoch 21/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.9984 - categorical_accuracy: 0.5964\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 1.0863 - categorical_accuracy: 0.5682 - val_loss: 0.9984 - val_categorical_accuracy: 0.5964\n",
      "Epoch 22/400\n",
      "18/18 [==============================] - 3s 172ms/step - loss: 1.0116 - categorical_accuracy: 0.61072s - loss: 1.5851 \n",
      "175/175 [==============================] - 90s 516ms/step - loss: 1.0887 - categorical_accuracy: 0.5654 - val_loss: 1.0116 - val_categorical_accuracy: 0.6107\n",
      "Epoch 23/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.9679 - categorical_accuracy: 0.6286\n",
      "175/175 [==============================] - 91s 521ms/step - loss: 1.0805 - categorical_accuracy: 0.5832 - val_loss: 0.9679 - val_categorical_accuracy: 0.6286\n",
      "Epoch 24/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.9662 - categorical_accuracy: 0.6429\n",
      "175/175 [==============================] - 91s 518ms/step - loss: 1.0239 - categorical_accuracy: 0.5939 - val_loss: 0.9662 - val_categorical_accuracy: 0.6429\n",
      "Epoch 25/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.8936 - categorical_accuracy: 0.6571\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 1.0323 - categorical_accuracy: 0.5932 - val_loss: 0.8936 - val_categorical_accuracy: 0.6571\n",
      "Epoch 26/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.9986 - categorical_accuracy: 0.6143\n",
      "175/175 [==============================] - 89s 510ms/step - loss: 0.9939 - categorical_accuracy: 0.5979 - val_loss: 0.9986 - val_categorical_accuracy: 0.6143\n",
      "Epoch 27/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.8499 - categorical_accuracy: 0.6821\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.9663 - categorical_accuracy: 0.6254 - val_loss: 0.8499 - val_categorical_accuracy: 0.6821\n",
      "Epoch 28/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.8937 - categorical_accuracy: 0.6893\n",
      "175/175 [==============================] - 89s 510ms/step - loss: 0.9406 - categorical_accuracy: 0.6432 - val_loss: 0.8937 - val_categorical_accuracy: 0.6893\n",
      "Epoch 29/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.7837 - categorical_accuracy: 0.6893\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 0.9524 - categorical_accuracy: 0.6243 - val_loss: 0.7837 - val_categorical_accuracy: 0.6893\n",
      "Epoch 30/400\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18/18 [==============================] - 3s 163ms/step - loss: 0.7578 - categorical_accuracy: 0.6964\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.9159 - categorical_accuracy: 0.6496 - val_loss: 0.7578 - val_categorical_accuracy: 0.6964\n",
      "Epoch 31/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.7796 - categorical_accuracy: 0.7107\n",
      "175/175 [==============================] - 89s 511ms/step - loss: 0.9202 - categorical_accuracy: 0.6382 - val_loss: 0.7796 - val_categorical_accuracy: 0.7107\n",
      "Epoch 32/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.7572 - categorical_accuracy: 0.7214\n",
      "175/175 [==============================] - 91s 518ms/step - loss: 0.8998 - categorical_accuracy: 0.6536 - val_loss: 0.7572 - val_categorical_accuracy: 0.7214\n",
      "Epoch 33/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.7414 - categorical_accuracy: 0.6821\n",
      "175/175 [==============================] - 91s 517ms/step - loss: 0.8534 - categorical_accuracy: 0.6668 - val_loss: 0.7414 - val_categorical_accuracy: 0.6821\n",
      "Epoch 34/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.7525 - categorical_accuracy: 0.7286\n",
      "175/175 [==============================] - 89s 511ms/step - loss: 0.8679 - categorical_accuracy: 0.6729 - val_loss: 0.7525 - val_categorical_accuracy: 0.7286\n",
      "Epoch 35/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.8017 - categorical_accuracy: 0.7250\n",
      "175/175 [==============================] - 89s 511ms/step - loss: 0.8165 - categorical_accuracy: 0.6907 - val_loss: 0.8017 - val_categorical_accuracy: 0.7250\n",
      "Epoch 36/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.8173 - categorical_accuracy: 0.6679\n",
      "175/175 [==============================] - 89s 511ms/step - loss: 0.8405 - categorical_accuracy: 0.6764 - val_loss: 0.8173 - val_categorical_accuracy: 0.6679\n",
      "Epoch 37/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.9512 - categorical_accuracy: 0.7000\n",
      "175/175 [==============================] - 89s 511ms/step - loss: 0.8272 - categorical_accuracy: 0.6796 - val_loss: 0.9512 - val_categorical_accuracy: 0.7000\n",
      "Epoch 38/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.7367 - categorical_accuracy: 0.7393\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 0.8027 - categorical_accuracy: 0.7050 - val_loss: 0.7367 - val_categorical_accuracy: 0.7393\n",
      "Epoch 39/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.6894 - categorical_accuracy: 0.7393\n",
      "175/175 [==============================] - 90s 517ms/step - loss: 0.7857 - categorical_accuracy: 0.7150 - val_loss: 0.6894 - val_categorical_accuracy: 0.7393\n",
      "Epoch 40/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.7253 - categorical_accuracy: 0.7536\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.8060 - categorical_accuracy: 0.6996 - val_loss: 0.7253 - val_categorical_accuracy: 0.7536\n",
      "Epoch 41/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.7170 - categorical_accuracy: 0.7214\n",
      "175/175 [==============================] - 89s 511ms/step - loss: 0.7888 - categorical_accuracy: 0.7157 - val_loss: 0.7170 - val_categorical_accuracy: 0.7214\n",
      "Epoch 42/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.7009 - categorical_accuracy: 0.7286\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.7551 - categorical_accuracy: 0.7154 - val_loss: 0.7009 - val_categorical_accuracy: 0.7286\n",
      "Epoch 43/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.7854 - categorical_accuracy: 0.7286\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.7464 - categorical_accuracy: 0.7257 - val_loss: 0.7854 - val_categorical_accuracy: 0.7286\n",
      "Epoch 44/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.7134 - categorical_accuracy: 0.7607\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 0.7209 - categorical_accuracy: 0.7336 - val_loss: 0.7134 - val_categorical_accuracy: 0.7607\n",
      "Epoch 45/400\n",
      "18/18 [==============================] - 3s 173ms/step - loss: 0.6830 - categorical_accuracy: 0.7250\n",
      "175/175 [==============================] - 92s 524ms/step - loss: 0.6950 - categorical_accuracy: 0.7486 - val_loss: 0.6830 - val_categorical_accuracy: 0.7250\n",
      "Epoch 46/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.7805 - categorical_accuracy: 0.7250\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.7163 - categorical_accuracy: 0.7361 - val_loss: 0.7805 - val_categorical_accuracy: 0.7250\n",
      "Epoch 47/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.7182 - categorical_accuracy: 0.7250\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.6871 - categorical_accuracy: 0.7546 - val_loss: 0.7182 - val_categorical_accuracy: 0.7250\n",
      "Epoch 48/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.6799 - categorical_accuracy: 0.7607\n",
      "175/175 [==============================] - 90s 517ms/step - loss: 0.6972 - categorical_accuracy: 0.7486 - val_loss: 0.6799 - val_categorical_accuracy: 0.7607\n",
      "Epoch 49/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.6915 - categorical_accuracy: 0.7929\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.6481 - categorical_accuracy: 0.7621 - val_loss: 0.6915 - val_categorical_accuracy: 0.7929\n",
      "Epoch 50/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.6005 - categorical_accuracy: 0.7893\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 0.6630 - categorical_accuracy: 0.7732 - val_loss: 0.6005 - val_categorical_accuracy: 0.7893\n",
      "Epoch 51/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.7103 - categorical_accuracy: 0.7750\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.6587 - categorical_accuracy: 0.7711 - val_loss: 0.7103 - val_categorical_accuracy: 0.7750\n",
      "Epoch 52/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.6314 - categorical_accuracy: 0.8000\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.6585 - categorical_accuracy: 0.7725 - val_loss: 0.6314 - val_categorical_accuracy: 0.8000\n",
      "Epoch 53/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.7111 - categorical_accuracy: 0.7679\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.6164 - categorical_accuracy: 0.7996 - val_loss: 0.7111 - val_categorical_accuracy: 0.7679\n",
      "Epoch 54/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.7002 - categorical_accuracy: 0.7750\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.6219 - categorical_accuracy: 0.7839 - val_loss: 0.7002 - val_categorical_accuracy: 0.7750\n",
      "Epoch 55/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5649 - categorical_accuracy: 0.8036\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 0.6087 - categorical_accuracy: 0.7943 - val_loss: 0.5649 - val_categorical_accuracy: 0.8036\n",
      "Epoch 56/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.6180 - categorical_accuracy: 0.7857\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.6056 - categorical_accuracy: 0.7896 - val_loss: 0.6180 - val_categorical_accuracy: 0.7857\n",
      "Epoch 57/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.6745 - categorical_accuracy: 0.7964\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.5826 - categorical_accuracy: 0.7954 - val_loss: 0.6745 - val_categorical_accuracy: 0.7964\n",
      "Epoch 58/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5837 - categorical_accuracy: 0.8036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.5860 - categorical_accuracy: 0.7957 - val_loss: 0.5837 - val_categorical_accuracy: 0.8036\n",
      "Epoch 59/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5558 - categorical_accuracy: 0.8107\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 0.5978 - categorical_accuracy: 0.7932 - val_loss: 0.5558 - val_categorical_accuracy: 0.8107\n",
      "Epoch 60/400\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18/18 [==============================] - 3s 163ms/step - loss: 0.6059 - categorical_accuracy: 0.8000\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.5701 - categorical_accuracy: 0.8004 - val_loss: 0.6059 - val_categorical_accuracy: 0.8000\n",
      "Epoch 61/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5943 - categorical_accuracy: 0.7750\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.5594 - categorical_accuracy: 0.8086 - val_loss: 0.5943 - val_categorical_accuracy: 0.7750\n",
      "Epoch 62/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5849 - categorical_accuracy: 0.7964\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.5694 - categorical_accuracy: 0.7989 - val_loss: 0.5849 - val_categorical_accuracy: 0.7964\n",
      "Epoch 63/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.6889 - categorical_accuracy: 0.7429\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.5386 - categorical_accuracy: 0.8204 - val_loss: 0.6889 - val_categorical_accuracy: 0.7429\n",
      "Epoch 64/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.6733 - categorical_accuracy: 0.7929\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.5359 - categorical_accuracy: 0.8179 - val_loss: 0.6733 - val_categorical_accuracy: 0.7929\n",
      "Epoch 65/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5029 - categorical_accuracy: 0.8250\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 0.5147 - categorical_accuracy: 0.8236 - val_loss: 0.5029 - val_categorical_accuracy: 0.8250\n",
      "Epoch 66/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.6851 - categorical_accuracy: 0.7536\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.5156 - categorical_accuracy: 0.8207 - val_loss: 0.6851 - val_categorical_accuracy: 0.7536\n",
      "Epoch 67/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5336 - categorical_accuracy: 0.8321\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.5127 - categorical_accuracy: 0.8293 - val_loss: 0.5336 - val_categorical_accuracy: 0.8321\n",
      "Epoch 68/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5857 - categorical_accuracy: 0.8214\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.5091 - categorical_accuracy: 0.8307 - val_loss: 0.5857 - val_categorical_accuracy: 0.8214\n",
      "Epoch 69/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5666 - categorical_accuracy: 0.8071\n",
      "175/175 [==============================] - 89s 511ms/step - loss: 0.5455 - categorical_accuracy: 0.8046 - val_loss: 0.5666 - val_categorical_accuracy: 0.8071\n",
      "Epoch 70/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.6108 - categorical_accuracy: 0.8000\n",
      "175/175 [==============================] - 89s 511ms/step - loss: 0.5316 - categorical_accuracy: 0.8189 - val_loss: 0.6108 - val_categorical_accuracy: 0.8000\n",
      "Epoch 71/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5683 - categorical_accuracy: 0.8500\n",
      "175/175 [==============================] - 90s 511ms/step - loss: 0.4812 - categorical_accuracy: 0.8371 - val_loss: 0.5683 - val_categorical_accuracy: 0.8500\n",
      "Epoch 72/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5411 - categorical_accuracy: 0.8179\n",
      "175/175 [==============================] - 90s 511ms/step - loss: 0.4769 - categorical_accuracy: 0.8379 - val_loss: 0.5411 - val_categorical_accuracy: 0.8179\n",
      "Epoch 73/400\n",
      "18/18 [==============================] - 3s 164ms/step - loss: 0.6380 - categorical_accuracy: 0.8250\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.4545 - categorical_accuracy: 0.8454 - val_loss: 0.6380 - val_categorical_accuracy: 0.8250\n",
      "Epoch 74/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.8258 - categorical_accuracy: 0.8000\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.5032 - categorical_accuracy: 0.8389 - val_loss: 0.8258 - val_categorical_accuracy: 0.8000\n",
      "Epoch 75/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.6715 - categorical_accuracy: 0.7857\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.4483 - categorical_accuracy: 0.8507 - val_loss: 0.6715 - val_categorical_accuracy: 0.7857\n",
      "Epoch 76/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4741 - categorical_accuracy: 0.8214\n",
      "175/175 [==============================] - 91s 517ms/step - loss: 0.4635 - categorical_accuracy: 0.8486 - val_loss: 0.4741 - val_categorical_accuracy: 0.8214\n",
      "Epoch 77/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4837 - categorical_accuracy: 0.8500\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.4285 - categorical_accuracy: 0.8561 - val_loss: 0.4837 - val_categorical_accuracy: 0.8500\n",
      "Epoch 78/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.5939 - categorical_accuracy: 0.8107\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.4652 - categorical_accuracy: 0.8507 - val_loss: 0.5939 - val_categorical_accuracy: 0.8107\n",
      "Epoch 79/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.5614 - categorical_accuracy: 0.8500\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.4438 - categorical_accuracy: 0.8561 - val_loss: 0.5614 - val_categorical_accuracy: 0.8500\n",
      "Epoch 80/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.4876 - categorical_accuracy: 0.8393\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.4625 - categorical_accuracy: 0.8496 - val_loss: 0.4876 - val_categorical_accuracy: 0.8393\n",
      "Epoch 81/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.6225 - categorical_accuracy: 0.8214\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.3993 - categorical_accuracy: 0.8714 - val_loss: 0.6225 - val_categorical_accuracy: 0.8214\n",
      "Epoch 82/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.4911 - categorical_accuracy: 0.8321\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.4615 - categorical_accuracy: 0.8496 - val_loss: 0.4911 - val_categorical_accuracy: 0.8321\n",
      "Epoch 83/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.5352 - categorical_accuracy: 0.8286\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.4251 - categorical_accuracy: 0.8607 - val_loss: 0.5352 - val_categorical_accuracy: 0.8286\n",
      "Epoch 84/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4597 - categorical_accuracy: 0.8536\n",
      "175/175 [==============================] - 91s 519ms/step - loss: 0.4029 - categorical_accuracy: 0.8729 - val_loss: 0.4597 - val_categorical_accuracy: 0.8536\n",
      "Epoch 85/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4524 - categorical_accuracy: 0.8643\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 0.4264 - categorical_accuracy: 0.8643 - val_loss: 0.4524 - val_categorical_accuracy: 0.8643\n",
      "Epoch 86/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5555 - categorical_accuracy: 0.8464\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.4081 - categorical_accuracy: 0.8754 - val_loss: 0.5555 - val_categorical_accuracy: 0.8464\n",
      "Epoch 87/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.4820 - categorical_accuracy: 0.8464\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.4066 - categorical_accuracy: 0.8682 - val_loss: 0.4820 - val_categorical_accuracy: 0.8464\n",
      "Epoch 88/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5384 - categorical_accuracy: 0.8500\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.4038 - categorical_accuracy: 0.8718 - val_loss: 0.5384 - val_categorical_accuracy: 0.8500\n",
      "Epoch 89/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.5171 - categorical_accuracy: 0.8393\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3929 - categorical_accuracy: 0.8754 - val_loss: 0.5171 - val_categorical_accuracy: 0.8393\n",
      "Epoch 90/400\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4345 - categorical_accuracy: 0.8714\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 0.3973 - categorical_accuracy: 0.8696 - val_loss: 0.4345 - val_categorical_accuracy: 0.8714\n",
      "Epoch 91/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4556 - categorical_accuracy: 0.8571\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.3670 - categorical_accuracy: 0.8861 - val_loss: 0.4556 - val_categorical_accuracy: 0.8571\n",
      "Epoch 92/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.6158 - categorical_accuracy: 0.8214\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3747 - categorical_accuracy: 0.8854 - val_loss: 0.6158 - val_categorical_accuracy: 0.8214\n",
      "Epoch 93/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.5445 - categorical_accuracy: 0.8500\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.4249 - categorical_accuracy: 0.8789 - val_loss: 0.5445 - val_categorical_accuracy: 0.8500\n",
      "Epoch 94/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.5048 - categorical_accuracy: 0.8536\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3832 - categorical_accuracy: 0.8743 - val_loss: 0.5048 - val_categorical_accuracy: 0.8536\n",
      "Epoch 95/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.5337 - categorical_accuracy: 0.8571\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3789 - categorical_accuracy: 0.8836 - val_loss: 0.5337 - val_categorical_accuracy: 0.8571\n",
      "Epoch 96/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5315 - categorical_accuracy: 0.8393\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3810 - categorical_accuracy: 0.8846 - val_loss: 0.5315 - val_categorical_accuracy: 0.8393\n",
      "Epoch 97/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4922 - categorical_accuracy: 0.8500\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3508 - categorical_accuracy: 0.8896 - val_loss: 0.4922 - val_categorical_accuracy: 0.8500\n",
      "Epoch 98/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5847 - categorical_accuracy: 0.8536\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3498 - categorical_accuracy: 0.8800 - val_loss: 0.5847 - val_categorical_accuracy: 0.8536\n",
      "Epoch 99/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4477 - categorical_accuracy: 0.8679\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3565 - categorical_accuracy: 0.8879 - val_loss: 0.4477 - val_categorical_accuracy: 0.8679\n",
      "Epoch 100/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4000 - categorical_accuracy: 0.8893\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 0.3686 - categorical_accuracy: 0.8814 - val_loss: 0.4000 - val_categorical_accuracy: 0.8893\n",
      "Epoch 101/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4520 - categorical_accuracy: 0.9000\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3579 - categorical_accuracy: 0.8861 - val_loss: 0.4520 - val_categorical_accuracy: 0.9000\n",
      "Epoch 102/400\n",
      "18/18 [==============================] - 3s 164ms/step - loss: 0.4589 - categorical_accuracy: 0.8857\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3598 - categorical_accuracy: 0.8886 - val_loss: 0.4589 - val_categorical_accuracy: 0.8857\n",
      "Epoch 103/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.4818 - categorical_accuracy: 0.8536\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3460 - categorical_accuracy: 0.8861 - val_loss: 0.4818 - val_categorical_accuracy: 0.8536\n",
      "Epoch 104/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.4444 - categorical_accuracy: 0.8643\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3409 - categorical_accuracy: 0.8925 - val_loss: 0.4444 - val_categorical_accuracy: 0.8643\n",
      "Epoch 105/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.5180 - categorical_accuracy: 0.8714\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3120 - categorical_accuracy: 0.9039 - val_loss: 0.5180 - val_categorical_accuracy: 0.8714\n",
      "Epoch 106/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.4386 - categorical_accuracy: 0.8429\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3296 - categorical_accuracy: 0.8921 - val_loss: 0.4386 - val_categorical_accuracy: 0.8429\n",
      "Epoch 107/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5316 - categorical_accuracy: 0.8571\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3285 - categorical_accuracy: 0.8993 - val_loss: 0.5316 - val_categorical_accuracy: 0.8571\n",
      "Epoch 108/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4239 - categorical_accuracy: 0.8929\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3379 - categorical_accuracy: 0.8989 - val_loss: 0.4239 - val_categorical_accuracy: 0.8929\n",
      "Epoch 109/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4754 - categorical_accuracy: 0.8643\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3098 - categorical_accuracy: 0.9014 - val_loss: 0.4754 - val_categorical_accuracy: 0.8643\n",
      "Epoch 110/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5125 - categorical_accuracy: 0.8643\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3068 - categorical_accuracy: 0.9050 - val_loss: 0.5125 - val_categorical_accuracy: 0.8643\n",
      "Epoch 111/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4644 - categorical_accuracy: 0.8750\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3259 - categorical_accuracy: 0.9014 - val_loss: 0.4644 - val_categorical_accuracy: 0.8750\n",
      "Epoch 112/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.8119 - categorical_accuracy: 0.8321\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3351 - categorical_accuracy: 0.9029 - val_loss: 0.8119 - val_categorical_accuracy: 0.8321\n",
      "Epoch 113/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4316 - categorical_accuracy: 0.8893\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3005 - categorical_accuracy: 0.9082 - val_loss: 0.4316 - val_categorical_accuracy: 0.8893\n",
      "Epoch 114/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4443 - categorical_accuracy: 0.8857\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3129 - categorical_accuracy: 0.9043 - val_loss: 0.4443 - val_categorical_accuracy: 0.8857\n",
      "Epoch 115/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5566 - categorical_accuracy: 0.8393\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3388 - categorical_accuracy: 0.8950 - val_loss: 0.5566 - val_categorical_accuracy: 0.8393\n",
      "Epoch 116/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4005 - categorical_accuracy: 0.8786\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2985 - categorical_accuracy: 0.9089 - val_loss: 0.4005 - val_categorical_accuracy: 0.8786\n",
      "Epoch 117/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.5150 - categorical_accuracy: 0.8500\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2817 - categorical_accuracy: 0.9143 - val_loss: 0.5150 - val_categorical_accuracy: 0.8500\n",
      "Epoch 118/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5485 - categorical_accuracy: 0.8607\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2925 - categorical_accuracy: 0.9104 - val_loss: 0.5485 - val_categorical_accuracy: 0.8607\n",
      "Epoch 119/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4373 - categorical_accuracy: 0.8643\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2726 - categorical_accuracy: 0.9071 - val_loss: 0.4373 - val_categorical_accuracy: 0.8643\n",
      "Epoch 120/400\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18/18 [==============================] - 3s 163ms/step - loss: 0.3955 - categorical_accuracy: 0.8964\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 0.2879 - categorical_accuracy: 0.9104 - val_loss: 0.3955 - val_categorical_accuracy: 0.8964\n",
      "Epoch 121/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4677 - categorical_accuracy: 0.8750\n",
      "175/175 [==============================] - 89s 511ms/step - loss: 0.2993 - categorical_accuracy: 0.9093 - val_loss: 0.4677 - val_categorical_accuracy: 0.8750\n",
      "Epoch 122/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4132 - categorical_accuracy: 0.8750\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2741 - categorical_accuracy: 0.9154 - val_loss: 0.4132 - val_categorical_accuracy: 0.8750\n",
      "Epoch 123/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4534 - categorical_accuracy: 0.8929\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2738 - categorical_accuracy: 0.9193 - val_loss: 0.4534 - val_categorical_accuracy: 0.8929\n",
      "Epoch 124/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5079 - categorical_accuracy: 0.8643\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2552 - categorical_accuracy: 0.9257 - val_loss: 0.5079 - val_categorical_accuracy: 0.8643\n",
      "Epoch 125/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.3678 - categorical_accuracy: 0.9000\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 0.2521 - categorical_accuracy: 0.9229 - val_loss: 0.3678 - val_categorical_accuracy: 0.9000\n",
      "Epoch 126/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4889 - categorical_accuracy: 0.8786\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.3128 - categorical_accuracy: 0.9054 - val_loss: 0.4889 - val_categorical_accuracy: 0.8786\n",
      "Epoch 127/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4108 - categorical_accuracy: 0.8750\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2960 - categorical_accuracy: 0.9043 - val_loss: 0.4108 - val_categorical_accuracy: 0.8750\n",
      "Epoch 128/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4975 - categorical_accuracy: 0.8714\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2672 - categorical_accuracy: 0.9175 - val_loss: 0.4975 - val_categorical_accuracy: 0.8714\n",
      "Epoch 129/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4289 - categorical_accuracy: 0.8786\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2897 - categorical_accuracy: 0.9132 - val_loss: 0.4289 - val_categorical_accuracy: 0.8786\n",
      "Epoch 130/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.3702 - categorical_accuracy: 0.9000\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2705 - categorical_accuracy: 0.9143 - val_loss: 0.3702 - val_categorical_accuracy: 0.9000\n",
      "Epoch 131/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4760 - categorical_accuracy: 0.8821\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2515 - categorical_accuracy: 0.9264 - val_loss: 0.4760 - val_categorical_accuracy: 0.8821\n",
      "Epoch 132/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.4403 - categorical_accuracy: 0.8821\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2356 - categorical_accuracy: 0.9218 - val_loss: 0.4403 - val_categorical_accuracy: 0.8821\n",
      "Epoch 133/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4356 - categorical_accuracy: 0.8929\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2479 - categorical_accuracy: 0.9171 - val_loss: 0.4356 - val_categorical_accuracy: 0.8929\n",
      "Epoch 134/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.3641 - categorical_accuracy: 0.9000\n",
      "175/175 [==============================] - 91s 518ms/step - loss: 0.2784 - categorical_accuracy: 0.9189 - val_loss: 0.3641 - val_categorical_accuracy: 0.9000\n",
      "Epoch 135/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.3909 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2969 - categorical_accuracy: 0.9082 - val_loss: 0.3909 - val_categorical_accuracy: 0.9036\n",
      "Epoch 136/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.4060 - categorical_accuracy: 0.8929\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2520 - categorical_accuracy: 0.9279 - val_loss: 0.4060 - val_categorical_accuracy: 0.8929\n",
      "Epoch 137/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.4425 - categorical_accuracy: 0.8821\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2482 - categorical_accuracy: 0.9254 - val_loss: 0.4425 - val_categorical_accuracy: 0.8821\n",
      "Epoch 138/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.3512 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 0.2558 - categorical_accuracy: 0.9232 - val_loss: 0.3512 - val_categorical_accuracy: 0.9036\n",
      "Epoch 139/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4560 - categorical_accuracy: 0.8964\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2446 - categorical_accuracy: 0.9304 - val_loss: 0.4560 - val_categorical_accuracy: 0.8964\n",
      "Epoch 140/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.5199 - categorical_accuracy: 0.8786\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2682 - categorical_accuracy: 0.9150 - val_loss: 0.5199 - val_categorical_accuracy: 0.8786\n",
      "Epoch 141/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.4472 - categorical_accuracy: 0.9000\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2549 - categorical_accuracy: 0.9214 - val_loss: 0.4472 - val_categorical_accuracy: 0.9000\n",
      "Epoch 142/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.5773 - categorical_accuracy: 0.8536\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2193 - categorical_accuracy: 0.9243 - val_loss: 0.5773 - val_categorical_accuracy: 0.8536\n",
      "Epoch 143/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4329 - categorical_accuracy: 0.8893\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2299 - categorical_accuracy: 0.9293 - val_loss: 0.4329 - val_categorical_accuracy: 0.8893\n",
      "Epoch 144/400\n",
      "18/18 [==============================] - 3s 164ms/step - loss: 0.5544 - categorical_accuracy: 0.8929\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2533 - categorical_accuracy: 0.9261 - val_loss: 0.5544 - val_categorical_accuracy: 0.8929\n",
      "Epoch 145/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.4375 - categorical_accuracy: 0.8893\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2233 - categorical_accuracy: 0.9332 - val_loss: 0.4375 - val_categorical_accuracy: 0.8893\n",
      "Epoch 146/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.4282 - categorical_accuracy: 0.8929\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2426 - categorical_accuracy: 0.9279 - val_loss: 0.4282 - val_categorical_accuracy: 0.8929\n",
      "Epoch 147/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4924 - categorical_accuracy: 0.9000\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2208 - categorical_accuracy: 0.9314 - val_loss: 0.4924 - val_categorical_accuracy: 0.9000\n",
      "Epoch 148/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.3999 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2982 - categorical_accuracy: 0.9157 - val_loss: 0.3999 - val_categorical_accuracy: 0.9036\n",
      "Epoch 149/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.3745 - categorical_accuracy: 0.9107\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2182 - categorical_accuracy: 0.9371 - val_loss: 0.3745 - val_categorical_accuracy: 0.9107\n",
      "Epoch 150/400\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18/18 [==============================] - 3s 163ms/step - loss: 0.3953 - categorical_accuracy: 0.8964\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2480 - categorical_accuracy: 0.9254 - val_loss: 0.3953 - val_categorical_accuracy: 0.8964\n",
      "Epoch 151/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5049 - categorical_accuracy: 0.8964\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2195 - categorical_accuracy: 0.9339 - val_loss: 0.5049 - val_categorical_accuracy: 0.8964\n",
      "Epoch 152/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4168 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2098 - categorical_accuracy: 0.9339 - val_loss: 0.4168 - val_categorical_accuracy: 0.9071\n",
      "Epoch 153/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4099 - categorical_accuracy: 0.9179\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2060 - categorical_accuracy: 0.9425 - val_loss: 0.4099 - val_categorical_accuracy: 0.9179\n",
      "Epoch 154/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5362 - categorical_accuracy: 0.8643\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2335 - categorical_accuracy: 0.9257 - val_loss: 0.5362 - val_categorical_accuracy: 0.8643\n",
      "Epoch 155/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4827 - categorical_accuracy: 0.8964\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2290 - categorical_accuracy: 0.9318 - val_loss: 0.4827 - val_categorical_accuracy: 0.8964\n",
      "Epoch 156/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4416 - categorical_accuracy: 0.8929\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2060 - categorical_accuracy: 0.9393 - val_loss: 0.4416 - val_categorical_accuracy: 0.8929\n",
      "Epoch 157/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5119 - categorical_accuracy: 0.8893\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.2228 - categorical_accuracy: 0.9329 - val_loss: 0.5119 - val_categorical_accuracy: 0.8893\n",
      "Epoch 158/400\n",
      "18/18 [==============================] - 3s 164ms/step - loss: 0.4797 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.2216 - categorical_accuracy: 0.9393 - val_loss: 0.4797 - val_categorical_accuracy: 0.9036\n",
      "Epoch 159/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4965 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1922 - categorical_accuracy: 0.9471 - val_loss: 0.4965 - val_categorical_accuracy: 0.9036\n",
      "Epoch 160/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4331 - categorical_accuracy: 0.9179\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2145 - categorical_accuracy: 0.9332 - val_loss: 0.4331 - val_categorical_accuracy: 0.9179\n",
      "Epoch 161/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5134 - categorical_accuracy: 0.8821\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2240 - categorical_accuracy: 0.9393 - val_loss: 0.5134 - val_categorical_accuracy: 0.8821\n",
      "Epoch 162/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.3846 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1971 - categorical_accuracy: 0.9396 - val_loss: 0.3846 - val_categorical_accuracy: 0.9071\n",
      "Epoch 163/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.4713 - categorical_accuracy: 0.8857\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2153 - categorical_accuracy: 0.9329 - val_loss: 0.4713 - val_categorical_accuracy: 0.8857\n",
      "Epoch 164/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4560 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 89s 511ms/step - loss: 0.2245 - categorical_accuracy: 0.9354 - val_loss: 0.4560 - val_categorical_accuracy: 0.9036\n",
      "Epoch 165/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.3907 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2090 - categorical_accuracy: 0.9432 - val_loss: 0.3907 - val_categorical_accuracy: 0.9071\n",
      "Epoch 166/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4323 - categorical_accuracy: 0.8964\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1940 - categorical_accuracy: 0.9396 - val_loss: 0.4323 - val_categorical_accuracy: 0.8964\n",
      "Epoch 167/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4318 - categorical_accuracy: 0.8857\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2352 - categorical_accuracy: 0.9382 - val_loss: 0.4318 - val_categorical_accuracy: 0.8857\n",
      "Epoch 168/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4307 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1872 - categorical_accuracy: 0.9493 - val_loss: 0.4307 - val_categorical_accuracy: 0.9071\n",
      "Epoch 169/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4043 - categorical_accuracy: 0.8929\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2013 - categorical_accuracy: 0.9414 - val_loss: 0.4043 - val_categorical_accuracy: 0.8929\n",
      "Epoch 170/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4388 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1840 - categorical_accuracy: 0.9468 - val_loss: 0.4388 - val_categorical_accuracy: 0.9036\n",
      "Epoch 171/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5243 - categorical_accuracy: 0.8929\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1950 - categorical_accuracy: 0.9446 - val_loss: 0.5243 - val_categorical_accuracy: 0.8929\n",
      "Epoch 172/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4919 - categorical_accuracy: 0.8786\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2031 - categorical_accuracy: 0.9432 - val_loss: 0.4919 - val_categorical_accuracy: 0.8786\n",
      "Epoch 173/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4414 - categorical_accuracy: 0.9000\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1622 - categorical_accuracy: 0.9511 - val_loss: 0.4414 - val_categorical_accuracy: 0.9000\n",
      "Epoch 174/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.5079 - categorical_accuracy: 0.8821\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2106 - categorical_accuracy: 0.9371 - val_loss: 0.5079 - val_categorical_accuracy: 0.8821\n",
      "Epoch 175/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.4237 - categorical_accuracy: 0.8571\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1984 - categorical_accuracy: 0.9379 - val_loss: 0.4237 - val_categorical_accuracy: 0.8571\n",
      "Epoch 176/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4281 - categorical_accuracy: 0.9000\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2161 - categorical_accuracy: 0.9382 - val_loss: 0.4281 - val_categorical_accuracy: 0.9000\n",
      "Epoch 177/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4356 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1761 - categorical_accuracy: 0.9525 - val_loss: 0.4356 - val_categorical_accuracy: 0.9036\n",
      "Epoch 178/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5907 - categorical_accuracy: 0.8821\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1757 - categorical_accuracy: 0.9464 - val_loss: 0.5907 - val_categorical_accuracy: 0.8821\n",
      "Epoch 179/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.5475 - categorical_accuracy: 0.8821\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1979 - categorical_accuracy: 0.9461 - val_loss: 0.5475 - val_categorical_accuracy: 0.8821\n",
      "Epoch 180/400\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18/18 [==============================] - 3s 162ms/step - loss: 0.5120 - categorical_accuracy: 0.8964\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1803 - categorical_accuracy: 0.9443 - val_loss: 0.5120 - val_categorical_accuracy: 0.8964\n",
      "Epoch 181/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5752 - categorical_accuracy: 0.8964\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1793 - categorical_accuracy: 0.9454 - val_loss: 0.5752 - val_categorical_accuracy: 0.8964\n",
      "Epoch 182/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4468 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1863 - categorical_accuracy: 0.9504 - val_loss: 0.4468 - val_categorical_accuracy: 0.9036\n",
      "Epoch 183/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4215 - categorical_accuracy: 0.9000\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2036 - categorical_accuracy: 0.9446 - val_loss: 0.4215 - val_categorical_accuracy: 0.9000\n",
      "Epoch 184/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.3726 - categorical_accuracy: 0.9214\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1905 - categorical_accuracy: 0.9446 - val_loss: 0.3726 - val_categorical_accuracy: 0.9214\n",
      "Epoch 185/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.4827 - categorical_accuracy: 0.8964\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1578 - categorical_accuracy: 0.9543 - val_loss: 0.4827 - val_categorical_accuracy: 0.8964\n",
      "Epoch 186/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.4221 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1621 - categorical_accuracy: 0.9504 - val_loss: 0.4221 - val_categorical_accuracy: 0.9071\n",
      "Epoch 187/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.3462 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 0.1600 - categorical_accuracy: 0.9554 - val_loss: 0.3462 - val_categorical_accuracy: 0.9071\n",
      "Epoch 188/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5349 - categorical_accuracy: 0.8929\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1554 - categorical_accuracy: 0.9614 - val_loss: 0.5349 - val_categorical_accuracy: 0.8929\n",
      "Epoch 189/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4789 - categorical_accuracy: 0.8786\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2314 - categorical_accuracy: 0.9314 - val_loss: 0.4789 - val_categorical_accuracy: 0.8786\n",
      "Epoch 190/400\n",
      "18/18 [==============================] - 3s 164ms/step - loss: 0.3913 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2066 - categorical_accuracy: 0.9407 - val_loss: 0.3913 - val_categorical_accuracy: 0.9071\n",
      "Epoch 191/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5516 - categorical_accuracy: 0.8714\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1719 - categorical_accuracy: 0.9493 - val_loss: 0.5516 - val_categorical_accuracy: 0.8714\n",
      "Epoch 192/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.5064 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1698 - categorical_accuracy: 0.9496 - val_loss: 0.5064 - val_categorical_accuracy: 0.9036\n",
      "Epoch 193/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.5121 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1561 - categorical_accuracy: 0.9568 - val_loss: 0.5121 - val_categorical_accuracy: 0.9036\n",
      "Epoch 194/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.5482 - categorical_accuracy: 0.8607\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1748 - categorical_accuracy: 0.9432 - val_loss: 0.5482 - val_categorical_accuracy: 0.8607\n",
      "Epoch 195/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.3941 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.2021 - categorical_accuracy: 0.9457 - val_loss: 0.3941 - val_categorical_accuracy: 0.9036\n",
      "Epoch 196/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4170 - categorical_accuracy: 0.9214\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1732 - categorical_accuracy: 0.9571 - val_loss: 0.4170 - val_categorical_accuracy: 0.9214\n",
      "Epoch 197/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.3565 - categorical_accuracy: 0.9321\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1649 - categorical_accuracy: 0.9529 - val_loss: 0.3565 - val_categorical_accuracy: 0.9321\n",
      "Epoch 198/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4815 - categorical_accuracy: 0.8964\n",
      "175/175 [==============================] - 90s 511ms/step - loss: 0.1666 - categorical_accuracy: 0.9593 - val_loss: 0.4815 - val_categorical_accuracy: 0.8964\n",
      "Epoch 199/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5285 - categorical_accuracy: 0.8893\n",
      "175/175 [==============================] - 89s 511ms/step - loss: 0.1406 - categorical_accuracy: 0.9625 - val_loss: 0.5285 - val_categorical_accuracy: 0.8893\n",
      "Epoch 200/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4261 - categorical_accuracy: 0.9179\n",
      "175/175 [==============================] - 90s 511ms/step - loss: 0.1662 - categorical_accuracy: 0.9546 - val_loss: 0.4261 - val_categorical_accuracy: 0.9179\n",
      "Epoch 201/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.3931 - categorical_accuracy: 0.9250\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1570 - categorical_accuracy: 0.9600 - val_loss: 0.3931 - val_categorical_accuracy: 0.9250\n",
      "Epoch 202/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.4228 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1832 - categorical_accuracy: 0.9489 - val_loss: 0.4228 - val_categorical_accuracy: 0.9071\n",
      "Epoch 203/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4100 - categorical_accuracy: 0.8964\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1427 - categorical_accuracy: 0.9589 - val_loss: 0.4100 - val_categorical_accuracy: 0.8964\n",
      "Epoch 204/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.6396 - categorical_accuracy: 0.8679\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1811 - categorical_accuracy: 0.9514 - val_loss: 0.6396 - val_categorical_accuracy: 0.8679\n",
      "Epoch 205/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5025 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1714 - categorical_accuracy: 0.9539 - val_loss: 0.5025 - val_categorical_accuracy: 0.9036\n",
      "Epoch 206/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.4723 - categorical_accuracy: 0.8893\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1605 - categorical_accuracy: 0.9539 - val_loss: 0.4723 - val_categorical_accuracy: 0.8893\n",
      "Epoch 207/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.6866 - categorical_accuracy: 0.8714\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1864 - categorical_accuracy: 0.9482 - val_loss: 0.6866 - val_categorical_accuracy: 0.8714\n",
      "Epoch 208/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.3378 - categorical_accuracy: 0.9179\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 0.1703 - categorical_accuracy: 0.9546 - val_loss: 0.3378 - val_categorical_accuracy: 0.9179\n",
      "Epoch 209/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5254 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1576 - categorical_accuracy: 0.9568 - val_loss: 0.5254 - val_categorical_accuracy: 0.9036\n",
      "Epoch 210/400\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4408 - categorical_accuracy: 0.9250\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1661 - categorical_accuracy: 0.9543 - val_loss: 0.4408 - val_categorical_accuracy: 0.9250\n",
      "Epoch 211/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5762 - categorical_accuracy: 0.8750\n",
      "175/175 [==============================] - 90s 511ms/step - loss: 0.1437 - categorical_accuracy: 0.9621 - val_loss: 0.5762 - val_categorical_accuracy: 0.8750\n",
      "Epoch 212/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5657 - categorical_accuracy: 0.8857\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1585 - categorical_accuracy: 0.9586 - val_loss: 0.5657 - val_categorical_accuracy: 0.8857\n",
      "Epoch 213/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4368 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1465 - categorical_accuracy: 0.9629 - val_loss: 0.4368 - val_categorical_accuracy: 0.9071\n",
      "Epoch 214/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.4645 - categorical_accuracy: 0.8679\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1634 - categorical_accuracy: 0.9568 - val_loss: 0.4645 - val_categorical_accuracy: 0.8679\n",
      "Epoch 215/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4149 - categorical_accuracy: 0.9214\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1437 - categorical_accuracy: 0.9579 - val_loss: 0.4149 - val_categorical_accuracy: 0.9214\n",
      "Epoch 216/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.6156 - categorical_accuracy: 0.8929\n",
      "175/175 [==============================] - 89s 511ms/step - loss: 0.1540 - categorical_accuracy: 0.9532 - val_loss: 0.6156 - val_categorical_accuracy: 0.8929\n",
      "Epoch 217/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5483 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1347 - categorical_accuracy: 0.9607 - val_loss: 0.5483 - val_categorical_accuracy: 0.9036\n",
      "Epoch 218/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.5539 - categorical_accuracy: 0.8821\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1733 - categorical_accuracy: 0.9557 - val_loss: 0.5539 - val_categorical_accuracy: 0.8821\n",
      "Epoch 219/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.5067 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1702 - categorical_accuracy: 0.9571 - val_loss: 0.5067 - val_categorical_accuracy: 0.9036\n",
      "Epoch 220/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5803 - categorical_accuracy: 0.9107\n",
      "175/175 [==============================] - 90s 511ms/step - loss: 0.1326 - categorical_accuracy: 0.9686 - val_loss: 0.5803 - val_categorical_accuracy: 0.9107\n",
      "Epoch 221/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.3935 - categorical_accuracy: 0.8964\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1433 - categorical_accuracy: 0.9629 - val_loss: 0.3935 - val_categorical_accuracy: 0.8964\n",
      "Epoch 222/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4574 - categorical_accuracy: 0.9107\n",
      "175/175 [==============================] - 89s 511ms/step - loss: 0.1529 - categorical_accuracy: 0.9604 - val_loss: 0.4574 - val_categorical_accuracy: 0.9107\n",
      "Epoch 223/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4718 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 89s 511ms/step - loss: 0.1680 - categorical_accuracy: 0.9539 - val_loss: 0.4718 - val_categorical_accuracy: 0.9071\n",
      "Epoch 224/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.7664 - categorical_accuracy: 0.8893\n",
      "175/175 [==============================] - 90s 511ms/step - loss: 0.1138 - categorical_accuracy: 0.9661 - val_loss: 0.7664 - val_categorical_accuracy: 0.8893\n",
      "Epoch 225/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5631 - categorical_accuracy: 0.8929\n",
      "175/175 [==============================] - 89s 511ms/step - loss: 0.1618 - categorical_accuracy: 0.9568 - val_loss: 0.5631 - val_categorical_accuracy: 0.8929\n",
      "Epoch 226/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.4921 - categorical_accuracy: 0.9107\n",
      "175/175 [==============================] - 90s 511ms/step - loss: 0.1660 - categorical_accuracy: 0.9525 - val_loss: 0.4921 - val_categorical_accuracy: 0.9107\n",
      "Epoch 227/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4156 - categorical_accuracy: 0.8929\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1320 - categorical_accuracy: 0.9643 - val_loss: 0.4156 - val_categorical_accuracy: 0.8929\n",
      "Epoch 228/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.6738 - categorical_accuracy: 0.9000\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1699 - categorical_accuracy: 0.9561 - val_loss: 0.6738 - val_categorical_accuracy: 0.9000\n",
      "Epoch 229/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5200 - categorical_accuracy: 0.9143\n",
      "175/175 [==============================] - 89s 511ms/step - loss: 0.1670 - categorical_accuracy: 0.9543 - val_loss: 0.5200 - val_categorical_accuracy: 0.9143\n",
      "Epoch 230/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.5268 - categorical_accuracy: 0.8893\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1741 - categorical_accuracy: 0.9500 - val_loss: 0.5268 - val_categorical_accuracy: 0.8893\n",
      "Epoch 231/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4263 - categorical_accuracy: 0.9286\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1400 - categorical_accuracy: 0.9650 - val_loss: 0.4263 - val_categorical_accuracy: 0.9286\n",
      "Epoch 232/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.3417 - categorical_accuracy: 0.9214\n",
      "175/175 [==============================] - 89s 511ms/step - loss: 0.1687 - categorical_accuracy: 0.9521 - val_loss: 0.3417 - val_categorical_accuracy: 0.9214\n",
      "Epoch 233/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4056 - categorical_accuracy: 0.9393\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1176 - categorical_accuracy: 0.9704 - val_loss: 0.4056 - val_categorical_accuracy: 0.9393\n",
      "Epoch 234/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.3878 - categorical_accuracy: 0.9214\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1451 - categorical_accuracy: 0.9654 - val_loss: 0.3878 - val_categorical_accuracy: 0.9214\n",
      "Epoch 235/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.4589 - categorical_accuracy: 0.9179\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1203 - categorical_accuracy: 0.9668 - val_loss: 0.4589 - val_categorical_accuracy: 0.9179\n",
      "Epoch 236/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.8201 - categorical_accuracy: 0.8286\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 0.1534 - categorical_accuracy: 0.9579 - val_loss: 0.8201 - val_categorical_accuracy: 0.8286\n",
      "Epoch 237/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.4974 - categorical_accuracy: 0.8857\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 0.1686 - categorical_accuracy: 0.9554 - val_loss: 0.4974 - val_categorical_accuracy: 0.8857\n",
      "Epoch 238/400\n",
      "18/18 [==============================] - 3s 168ms/step - loss: 0.5363 - categorical_accuracy: 0.9000\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.1303 - categorical_accuracy: 0.9611 - val_loss: 0.5363 - val_categorical_accuracy: 0.9000\n",
      "Epoch 239/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.5822 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 0.1167 - categorical_accuracy: 0.9657 - val_loss: 0.5822 - val_categorical_accuracy: 0.9036\n",
      "Epoch 240/400\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4686 - categorical_accuracy: 0.9321\n",
      "175/175 [==============================] - 90s 517ms/step - loss: 0.1464 - categorical_accuracy: 0.9600 - val_loss: 0.4686 - val_categorical_accuracy: 0.9321\n",
      "Epoch 241/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.3647 - categorical_accuracy: 0.9321\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 0.1649 - categorical_accuracy: 0.9518 - val_loss: 0.3647 - val_categorical_accuracy: 0.9321\n",
      "Epoch 242/400\n",
      "18/18 [==============================] - 3s 168ms/step - loss: 0.3670 - categorical_accuracy: 0.8786\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.1415 - categorical_accuracy: 0.9636 - val_loss: 0.3670 - val_categorical_accuracy: 0.8786\n",
      "Epoch 243/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4373 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 0.1123 - categorical_accuracy: 0.9704 - val_loss: 0.4373 - val_categorical_accuracy: 0.9071\n",
      "Epoch 244/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.4699 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 0.1364 - categorical_accuracy: 0.9636 - val_loss: 0.4699 - val_categorical_accuracy: 0.9071\n",
      "Epoch 245/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4828 - categorical_accuracy: 0.9179\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 0.1814 - categorical_accuracy: 0.9525 - val_loss: 0.4828 - val_categorical_accuracy: 0.9179\n",
      "Epoch 246/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.3932 - categorical_accuracy: 0.9250\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.1211 - categorical_accuracy: 0.9650 - val_loss: 0.3932 - val_categorical_accuracy: 0.9250\n",
      "Epoch 247/400\n",
      "18/18 [==============================] - 3s 168ms/step - loss: 0.4875 - categorical_accuracy: 0.9000\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 0.1271 - categorical_accuracy: 0.9664 - val_loss: 0.4875 - val_categorical_accuracy: 0.9000\n",
      "Epoch 248/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.5658 - categorical_accuracy: 0.8929\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 0.1146 - categorical_accuracy: 0.9704 - val_loss: 0.5658 - val_categorical_accuracy: 0.8929\n",
      "Epoch 249/400\n",
      "18/18 [==============================] - 3s 168ms/step - loss: 0.4180 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 0.1211 - categorical_accuracy: 0.9707 - val_loss: 0.4180 - val_categorical_accuracy: 0.9071\n",
      "Epoch 250/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.4695 - categorical_accuracy: 0.8929\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.1396 - categorical_accuracy: 0.9646 - val_loss: 0.4695 - val_categorical_accuracy: 0.8929\n",
      "Epoch 251/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.4837 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.1251 - categorical_accuracy: 0.9650 - val_loss: 0.4837 - val_categorical_accuracy: 0.9071\n",
      "Epoch 252/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4546 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 0.1479 - categorical_accuracy: 0.9643 - val_loss: 0.4546 - val_categorical_accuracy: 0.9036\n",
      "Epoch 253/400\n",
      "18/18 [==============================] - 3s 169ms/step - loss: 0.4449 - categorical_accuracy: 0.9143: 2s - loss: 0.169\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 0.1177 - categorical_accuracy: 0.9675 - val_loss: 0.4449 - val_categorical_accuracy: 0.9143\n",
      "Epoch 254/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.6276 - categorical_accuracy: 0.8821\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.1067 - categorical_accuracy: 0.9700 - val_loss: 0.6276 - val_categorical_accuracy: 0.8821\n",
      "Epoch 255/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.6342 - categorical_accuracy: 0.8964\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.1120 - categorical_accuracy: 0.9675 - val_loss: 0.6342 - val_categorical_accuracy: 0.8964\n",
      "Epoch 256/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.3542 - categorical_accuracy: 0.9179\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1375 - categorical_accuracy: 0.9596 - val_loss: 0.3542 - val_categorical_accuracy: 0.9179\n",
      "Epoch 257/400\n",
      "18/18 [==============================] - 3s 168ms/step - loss: 0.4459 - categorical_accuracy: 0.9179\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.1589 - categorical_accuracy: 0.9575 - val_loss: 0.4459 - val_categorical_accuracy: 0.9179\n",
      "Epoch 258/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.4642 - categorical_accuracy: 0.8857\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1776 - categorical_accuracy: 0.9482 - val_loss: 0.4642 - val_categorical_accuracy: 0.8857\n",
      "Epoch 259/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.4773 - categorical_accuracy: 0.9214\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.1155 - categorical_accuracy: 0.9686 - val_loss: 0.4773 - val_categorical_accuracy: 0.9214\n",
      "Epoch 260/400\n",
      "18/18 [==============================] - 3s 168ms/step - loss: 0.3389 - categorical_accuracy: 0.9286\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 0.1099 - categorical_accuracy: 0.9721 - val_loss: 0.3389 - val_categorical_accuracy: 0.9286\n",
      "Epoch 261/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.4408 - categorical_accuracy: 0.9107\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 0.1274 - categorical_accuracy: 0.9625 - val_loss: 0.4408 - val_categorical_accuracy: 0.9107\n",
      "Epoch 262/400\n",
      "18/18 [==============================] - 3s 168ms/step - loss: 0.4938 - categorical_accuracy: 0.9321\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 0.1164 - categorical_accuracy: 0.9700 - val_loss: 0.4938 - val_categorical_accuracy: 0.9321\n",
      "Epoch 263/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4191 - categorical_accuracy: 0.9250\n",
      "175/175 [==============================] - 90s 517ms/step - loss: 0.1111 - categorical_accuracy: 0.9718 - val_loss: 0.4191 - val_categorical_accuracy: 0.9250\n",
      "Epoch 264/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.3536 - categorical_accuracy: 0.9286\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 0.1174 - categorical_accuracy: 0.9679 - val_loss: 0.3536 - val_categorical_accuracy: 0.9286\n",
      "Epoch 265/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.3609 - categorical_accuracy: 0.9107\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.0992 - categorical_accuracy: 0.9736 - val_loss: 0.3609 - val_categorical_accuracy: 0.9107\n",
      "Epoch 266/400\n",
      "18/18 [==============================] - 3s 168ms/step - loss: 0.4894 - categorical_accuracy: 0.9286\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 0.1052 - categorical_accuracy: 0.9746 - val_loss: 0.4894 - val_categorical_accuracy: 0.9286\n",
      "Epoch 267/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.5999 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.1029 - categorical_accuracy: 0.9700 - val_loss: 0.5999 - val_categorical_accuracy: 0.9071\n",
      "Epoch 268/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.6913 - categorical_accuracy: 0.8786\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 0.1269 - categorical_accuracy: 0.9629 - val_loss: 0.6913 - val_categorical_accuracy: 0.8786\n",
      "Epoch 269/400\n",
      "18/18 [==============================] - 3s 168ms/step - loss: 0.5699 - categorical_accuracy: 0.8893\n",
      "175/175 [==============================] - 90s 516ms/step - loss: 0.1173 - categorical_accuracy: 0.9664 - val_loss: 0.5699 - val_categorical_accuracy: 0.8893\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 270/400\n",
      "18/18 [==============================] - 3s 164ms/step - loss: 0.4283 - categorical_accuracy: 0.9214\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 0.1220 - categorical_accuracy: 0.9693 - val_loss: 0.4283 - val_categorical_accuracy: 0.9214\n",
      "Epoch 271/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.5083 - categorical_accuracy: 0.9107\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1484 - categorical_accuracy: 0.9621 - val_loss: 0.5083 - val_categorical_accuracy: 0.9107\n",
      "Epoch 272/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.6394 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.1008 - categorical_accuracy: 0.9746 - val_loss: 0.6394 - val_categorical_accuracy: 0.9036\n",
      "Epoch 273/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4951 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1028 - categorical_accuracy: 0.9743 - val_loss: 0.4951 - val_categorical_accuracy: 0.9036\n",
      "Epoch 274/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.4200 - categorical_accuracy: 0.9107\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1067 - categorical_accuracy: 0.9693 - val_loss: 0.4200 - val_categorical_accuracy: 0.9107\n",
      "Epoch 275/400\n",
      "18/18 [==============================] - 3s 169ms/step - loss: 0.5491 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 0.1268 - categorical_accuracy: 0.9689 - val_loss: 0.5491 - val_categorical_accuracy: 0.9036\n",
      "Epoch 276/400\n",
      "18/18 [==============================] - 3s 164ms/step - loss: 0.7020 - categorical_accuracy: 0.8929\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 0.0769 - categorical_accuracy: 0.9782 - val_loss: 0.7020 - val_categorical_accuracy: 0.8929\n",
      "Epoch 277/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.8891 - categorical_accuracy: 0.8714\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1328 - categorical_accuracy: 0.9621 - val_loss: 0.8891 - val_categorical_accuracy: 0.8714\n",
      "Epoch 278/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4141 - categorical_accuracy: 0.9000\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.1629 - categorical_accuracy: 0.9561 - val_loss: 0.4141 - val_categorical_accuracy: 0.9000\n",
      "Epoch 279/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.5173 - categorical_accuracy: 0.9179\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.1067 - categorical_accuracy: 0.9689 - val_loss: 0.5173 - val_categorical_accuracy: 0.9179\n",
      "Epoch 280/400\n",
      "18/18 [==============================] - 3s 170ms/step - loss: 0.7591 - categorical_accuracy: 0.8643\n",
      "175/175 [==============================] - 91s 517ms/step - loss: 0.1724 - categorical_accuracy: 0.9564 - val_loss: 0.7591 - val_categorical_accuracy: 0.8643\n",
      "Epoch 281/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.5709 - categorical_accuracy: 0.9000\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1099 - categorical_accuracy: 0.9693 - val_loss: 0.5709 - val_categorical_accuracy: 0.9000\n",
      "Epoch 282/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.5434 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.0832 - categorical_accuracy: 0.9775 - val_loss: 0.5434 - val_categorical_accuracy: 0.9036\n",
      "Epoch 283/400\n",
      "18/18 [==============================] - 3s 169ms/step - loss: 0.6396 - categorical_accuracy: 0.8964\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.1086 - categorical_accuracy: 0.9700 - val_loss: 0.6396 - val_categorical_accuracy: 0.8964\n",
      "Epoch 284/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.5522 - categorical_accuracy: 0.8964\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.1359 - categorical_accuracy: 0.9636 - val_loss: 0.5522 - val_categorical_accuracy: 0.8964\n",
      "Epoch 285/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.4012 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.1979 - categorical_accuracy: 0.9486 - val_loss: 0.4012 - val_categorical_accuracy: 0.9036\n",
      "Epoch 286/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.5724 - categorical_accuracy: 0.9143\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.0949 - categorical_accuracy: 0.9764 - val_loss: 0.5724 - val_categorical_accuracy: 0.9143\n",
      "Epoch 287/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.4836 - categorical_accuracy: 0.9107\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0947 - categorical_accuracy: 0.9779 - val_loss: 0.4836 - val_categorical_accuracy: 0.9107\n",
      "Epoch 288/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.3836 - categorical_accuracy: 0.9250\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.0978 - categorical_accuracy: 0.9761 - val_loss: 0.3836 - val_categorical_accuracy: 0.9250\n",
      "Epoch 289/400\n",
      "18/18 [==============================] - 3s 168ms/step - loss: 0.4221 - categorical_accuracy: 0.9143\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0825 - categorical_accuracy: 0.9804 - val_loss: 0.4221 - val_categorical_accuracy: 0.9143\n",
      "Epoch 290/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.4522 - categorical_accuracy: 0.9179\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.0975 - categorical_accuracy: 0.9732 - val_loss: 0.4522 - val_categorical_accuracy: 0.9179\n",
      "Epoch 291/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.5519 - categorical_accuracy: 0.9214\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0889 - categorical_accuracy: 0.9757 - val_loss: 0.5519 - val_categorical_accuracy: 0.9214\n",
      "Epoch 292/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.5007 - categorical_accuracy: 0.9250\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0988 - categorical_accuracy: 0.9754 - val_loss: 0.5007 - val_categorical_accuracy: 0.9250\n",
      "Epoch 293/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.5063 - categorical_accuracy: 0.9143\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1216 - categorical_accuracy: 0.9696 - val_loss: 0.5063 - val_categorical_accuracy: 0.9143\n",
      "Epoch 294/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.4402 - categorical_accuracy: 0.9214\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1236 - categorical_accuracy: 0.9646 - val_loss: 0.4402 - val_categorical_accuracy: 0.9214\n",
      "Epoch 295/400\n",
      "18/18 [==============================] - 3s 164ms/step - loss: 0.6118 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1256 - categorical_accuracy: 0.9682 - val_loss: 0.6118 - val_categorical_accuracy: 0.9036\n",
      "Epoch 296/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.5487 - categorical_accuracy: 0.9214\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0822 - categorical_accuracy: 0.9764 - val_loss: 0.5487 - val_categorical_accuracy: 0.9214\n",
      "Epoch 297/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.4992 - categorical_accuracy: 0.9107\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0895 - categorical_accuracy: 0.9764 - val_loss: 0.4992 - val_categorical_accuracy: 0.9107\n",
      "Epoch 298/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.5155 - categorical_accuracy: 0.8964\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1172 - categorical_accuracy: 0.9711 - val_loss: 0.5155 - val_categorical_accuracy: 0.8964\n",
      "Epoch 299/400\n",
      "18/18 [==============================] - 3s 164ms/step - loss: 0.5379 - categorical_accuracy: 0.9000\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0981 - categorical_accuracy: 0.9746 - val_loss: 0.5379 - val_categorical_accuracy: 0.9000\n",
      "Epoch 300/400\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18/18 [==============================] - 3s 164ms/step - loss: 0.4381 - categorical_accuracy: 0.9000\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1238 - categorical_accuracy: 0.9693 - val_loss: 0.4381 - val_categorical_accuracy: 0.9000\n",
      "Epoch 301/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.5821 - categorical_accuracy: 0.8750\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1237 - categorical_accuracy: 0.9639 - val_loss: 0.5821 - val_categorical_accuracy: 0.8750\n",
      "Epoch 302/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.5599 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.0762 - categorical_accuracy: 0.9814 - val_loss: 0.5599 - val_categorical_accuracy: 0.9036\n",
      "Epoch 303/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.5829 - categorical_accuracy: 0.8929\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1130 - categorical_accuracy: 0.9707 - val_loss: 0.5829 - val_categorical_accuracy: 0.8929\n",
      "Epoch 304/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.5754 - categorical_accuracy: 0.8893\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1120 - categorical_accuracy: 0.9711 - val_loss: 0.5754 - val_categorical_accuracy: 0.8893\n",
      "Epoch 305/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.5755 - categorical_accuracy: 0.8821\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0946 - categorical_accuracy: 0.9721 - val_loss: 0.5755 - val_categorical_accuracy: 0.8821\n",
      "Epoch 306/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.4089 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1516 - categorical_accuracy: 0.9600 - val_loss: 0.4089 - val_categorical_accuracy: 0.9036\n",
      "Epoch 307/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.6431 - categorical_accuracy: 0.8786\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1501 - categorical_accuracy: 0.9650 - val_loss: 0.6431 - val_categorical_accuracy: 0.8786\n",
      "Epoch 308/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4346 - categorical_accuracy: 0.9286\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1205 - categorical_accuracy: 0.9689 - val_loss: 0.4346 - val_categorical_accuracy: 0.9286\n",
      "Epoch 309/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.4587 - categorical_accuracy: 0.9250\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.0702 - categorical_accuracy: 0.9821 - val_loss: 0.4587 - val_categorical_accuracy: 0.9250\n",
      "Epoch 310/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.6299 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0887 - categorical_accuracy: 0.9793 - val_loss: 0.6299 - val_categorical_accuracy: 0.9036\n",
      "Epoch 311/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.4499 - categorical_accuracy: 0.9250\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0911 - categorical_accuracy: 0.9779 - val_loss: 0.4499 - val_categorical_accuracy: 0.9250\n",
      "Epoch 312/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4153 - categorical_accuracy: 0.9286\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1026 - categorical_accuracy: 0.9739 - val_loss: 0.4153 - val_categorical_accuracy: 0.9286\n",
      "Epoch 313/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.3924 - categorical_accuracy: 0.9107\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0861 - categorical_accuracy: 0.9818 - val_loss: 0.3924 - val_categorical_accuracy: 0.9107\n",
      "Epoch 314/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.5818 - categorical_accuracy: 0.9107\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1138 - categorical_accuracy: 0.9650 - val_loss: 0.5818 - val_categorical_accuracy: 0.9107\n",
      "Epoch 315/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4722 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1174 - categorical_accuracy: 0.9686 - val_loss: 0.4722 - val_categorical_accuracy: 0.9071\n",
      "Epoch 316/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4974 - categorical_accuracy: 0.9250\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1065 - categorical_accuracy: 0.9704 - val_loss: 0.4974 - val_categorical_accuracy: 0.9250\n",
      "Epoch 317/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.5945 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.0923 - categorical_accuracy: 0.9757 - val_loss: 0.5945 - val_categorical_accuracy: 0.9071\n",
      "Epoch 318/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.4695 - categorical_accuracy: 0.9179\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0974 - categorical_accuracy: 0.9721 - val_loss: 0.4695 - val_categorical_accuracy: 0.9179\n",
      "Epoch 319/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4798 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.0957 - categorical_accuracy: 0.9754 - val_loss: 0.4798 - val_categorical_accuracy: 0.9071\n",
      "Epoch 320/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.6251 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0855 - categorical_accuracy: 0.9818 - val_loss: 0.6251 - val_categorical_accuracy: 0.9071\n",
      "Epoch 321/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.5329 - categorical_accuracy: 0.9143\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1491 - categorical_accuracy: 0.9600 - val_loss: 0.5329 - val_categorical_accuracy: 0.9143\n",
      "Epoch 322/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.6958 - categorical_accuracy: 0.9000\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1132 - categorical_accuracy: 0.9707 - val_loss: 0.6958 - val_categorical_accuracy: 0.9000\n",
      "Epoch 323/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.6692 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1085 - categorical_accuracy: 0.9686 - val_loss: 0.6692 - val_categorical_accuracy: 0.9036\n",
      "Epoch 324/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.5066 - categorical_accuracy: 0.9321\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1094 - categorical_accuracy: 0.9764 - val_loss: 0.5066 - val_categorical_accuracy: 0.9321\n",
      "Epoch 325/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.5225 - categorical_accuracy: 0.9143\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1317 - categorical_accuracy: 0.9668 - val_loss: 0.5225 - val_categorical_accuracy: 0.9143\n",
      "Epoch 326/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.3898 - categorical_accuracy: 0.9214\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1137 - categorical_accuracy: 0.9675 - val_loss: 0.3898 - val_categorical_accuracy: 0.9214\n",
      "Epoch 327/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4468 - categorical_accuracy: 0.9250\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1014 - categorical_accuracy: 0.9786 - val_loss: 0.4468 - val_categorical_accuracy: 0.9250\n",
      "Epoch 328/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.6024 - categorical_accuracy: 0.9179\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0948 - categorical_accuracy: 0.9743 - val_loss: 0.6024 - val_categorical_accuracy: 0.9179\n",
      "Epoch 329/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.4498 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0931 - categorical_accuracy: 0.9750 - val_loss: 0.4498 - val_categorical_accuracy: 0.9071\n",
      "Epoch 330/400\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4208 - categorical_accuracy: 0.9321\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1103 - categorical_accuracy: 0.9689 - val_loss: 0.4208 - val_categorical_accuracy: 0.9321\n",
      "Epoch 331/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.3691 - categorical_accuracy: 0.9107\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1068 - categorical_accuracy: 0.9750 - val_loss: 0.3691 - val_categorical_accuracy: 0.9107\n",
      "Epoch 332/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.3545 - categorical_accuracy: 0.9143\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0996 - categorical_accuracy: 0.9764 - val_loss: 0.3545 - val_categorical_accuracy: 0.9143\n",
      "Epoch 333/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.3613 - categorical_accuracy: 0.9214\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.0949 - categorical_accuracy: 0.9764 - val_loss: 0.3613 - val_categorical_accuracy: 0.9214\n",
      "Epoch 334/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.6384 - categorical_accuracy: 0.8893\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1524 - categorical_accuracy: 0.9625 - val_loss: 0.6384 - val_categorical_accuracy: 0.8893\n",
      "Epoch 335/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.3651 - categorical_accuracy: 0.9393\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1019 - categorical_accuracy: 0.9789 - val_loss: 0.3651 - val_categorical_accuracy: 0.9393\n",
      "Epoch 336/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4863 - categorical_accuracy: 0.9214\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.0654 - categorical_accuracy: 0.9839 - val_loss: 0.4863 - val_categorical_accuracy: 0.9214\n",
      "Epoch 337/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.5353 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1047 - categorical_accuracy: 0.9771 - val_loss: 0.5353 - val_categorical_accuracy: 0.9071\n",
      "Epoch 338/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4326 - categorical_accuracy: 0.9357\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1304 - categorical_accuracy: 0.9646 - val_loss: 0.4326 - val_categorical_accuracy: 0.9357\n",
      "Epoch 339/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.5234 - categorical_accuracy: 0.9179\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.0856 - categorical_accuracy: 0.9796 - val_loss: 0.5234 - val_categorical_accuracy: 0.9179\n",
      "Epoch 340/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.5063 - categorical_accuracy: 0.9000\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.0694 - categorical_accuracy: 0.9796 - val_loss: 0.5063 - val_categorical_accuracy: 0.9000\n",
      "Epoch 341/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.5025 - categorical_accuracy: 0.9214\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.0947 - categorical_accuracy: 0.9711 - val_loss: 0.5025 - val_categorical_accuracy: 0.9214\n",
      "Epoch 342/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.5784 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1004 - categorical_accuracy: 0.9718 - val_loss: 0.5784 - val_categorical_accuracy: 0.9071\n",
      "Epoch 343/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.6271 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0681 - categorical_accuracy: 0.9832 - val_loss: 0.6271 - val_categorical_accuracy: 0.9036\n",
      "Epoch 344/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.5484 - categorical_accuracy: 0.9143\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0964 - categorical_accuracy: 0.9775 - val_loss: 0.5484 - val_categorical_accuracy: 0.9143\n",
      "Epoch 345/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.6275 - categorical_accuracy: 0.9107\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0865 - categorical_accuracy: 0.9754 - val_loss: 0.6275 - val_categorical_accuracy: 0.9107\n",
      "Epoch 346/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.6020 - categorical_accuracy: 0.9107\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0957 - categorical_accuracy: 0.9743 - val_loss: 0.6020 - val_categorical_accuracy: 0.9107\n",
      "Epoch 347/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.3557 - categorical_accuracy: 0.9107\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1311 - categorical_accuracy: 0.9671 - val_loss: 0.3557 - val_categorical_accuracy: 0.9107\n",
      "Epoch 348/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.4390 - categorical_accuracy: 0.9321\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0822 - categorical_accuracy: 0.9775 - val_loss: 0.4390 - val_categorical_accuracy: 0.9321\n",
      "Epoch 349/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4110 - categorical_accuracy: 0.9214\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1158 - categorical_accuracy: 0.9750 - val_loss: 0.4110 - val_categorical_accuracy: 0.9214\n",
      "Epoch 350/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.7931 - categorical_accuracy: 0.8821\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0728 - categorical_accuracy: 0.9807 - val_loss: 0.7931 - val_categorical_accuracy: 0.8821\n",
      "Epoch 351/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4393 - categorical_accuracy: 0.9214\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0730 - categorical_accuracy: 0.9800 - val_loss: 0.4393 - val_categorical_accuracy: 0.9214\n",
      "Epoch 352/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.6208 - categorical_accuracy: 0.9250\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0709 - categorical_accuracy: 0.9825 - val_loss: 0.6208 - val_categorical_accuracy: 0.9250\n",
      "Epoch 353/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.6541 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0889 - categorical_accuracy: 0.9757 - val_loss: 0.6541 - val_categorical_accuracy: 0.9036\n",
      "Epoch 354/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.5907 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1242 - categorical_accuracy: 0.9689 - val_loss: 0.5907 - val_categorical_accuracy: 0.9036\n",
      "Epoch 355/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.3978 - categorical_accuracy: 0.9107\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1032 - categorical_accuracy: 0.9779 - val_loss: 0.3978 - val_categorical_accuracy: 0.9107\n",
      "Epoch 356/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.5646 - categorical_accuracy: 0.8643\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1711 - categorical_accuracy: 0.9589 - val_loss: 0.5646 - val_categorical_accuracy: 0.8643\n",
      "Epoch 357/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.5356 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.0916 - categorical_accuracy: 0.9779 - val_loss: 0.5356 - val_categorical_accuracy: 0.9071\n",
      "Epoch 358/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.5386 - categorical_accuracy: 0.9107\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0845 - categorical_accuracy: 0.9768 - val_loss: 0.5386 - val_categorical_accuracy: 0.9107\n",
      "Epoch 359/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.8226 - categorical_accuracy: 0.8821\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0891 - categorical_accuracy: 0.9757 - val_loss: 0.8226 - val_categorical_accuracy: 0.8821\n",
      "Epoch 360/400\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18/18 [==============================] - 3s 166ms/step - loss: 0.5739 - categorical_accuracy: 0.8821\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0917 - categorical_accuracy: 0.9743 - val_loss: 0.5739 - val_categorical_accuracy: 0.8821\n",
      "Epoch 361/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.3394 - categorical_accuracy: 0.9250\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1115 - categorical_accuracy: 0.9725 - val_loss: 0.3394 - val_categorical_accuracy: 0.9250\n",
      "Epoch 362/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.2910 - categorical_accuracy: 0.9250\n",
      "175/175 [==============================] - 90s 517ms/step - loss: 0.1353 - categorical_accuracy: 0.9636 - val_loss: 0.2910 - val_categorical_accuracy: 0.9250\n",
      "Epoch 363/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4541 - categorical_accuracy: 0.9179\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0774 - categorical_accuracy: 0.9811 - val_loss: 0.4541 - val_categorical_accuracy: 0.9179\n",
      "Epoch 364/400\n",
      "18/18 [==============================] - 3s 164ms/step - loss: 0.5319 - categorical_accuracy: 0.9107\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0628 - categorical_accuracy: 0.9864 - val_loss: 0.5319 - val_categorical_accuracy: 0.9107\n",
      "Epoch 365/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.5166 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0798 - categorical_accuracy: 0.9804 - val_loss: 0.5166 - val_categorical_accuracy: 0.9071\n",
      "Epoch 366/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.4277 - categorical_accuracy: 0.9250\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0920 - categorical_accuracy: 0.9764 - val_loss: 0.4277 - val_categorical_accuracy: 0.9250\n",
      "Epoch 367/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.5831 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0795 - categorical_accuracy: 0.9814 - val_loss: 0.5831 - val_categorical_accuracy: 0.9071\n",
      "Epoch 368/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.5745 - categorical_accuracy: 0.9071\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1097 - categorical_accuracy: 0.9732 - val_loss: 0.5745 - val_categorical_accuracy: 0.9071\n",
      "Epoch 369/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.3993 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0900 - categorical_accuracy: 0.9786 - val_loss: 0.3993 - val_categorical_accuracy: 0.9036\n",
      "Epoch 370/400\n",
      "18/18 [==============================] - 3s 165ms/step - loss: 0.7215 - categorical_accuracy: 0.8964\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0749 - categorical_accuracy: 0.9782 - val_loss: 0.7215 - val_categorical_accuracy: 0.8964\n",
      "Epoch 371/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.3758 - categorical_accuracy: 0.9321\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1138 - categorical_accuracy: 0.9689 - val_loss: 0.3758 - val_categorical_accuracy: 0.9321\n",
      "Epoch 372/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.4622 - categorical_accuracy: 0.9250\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1368 - categorical_accuracy: 0.9671 - val_loss: 0.4622 - val_categorical_accuracy: 0.9250\n",
      "Epoch 373/400\n",
      "18/18 [==============================] - 3s 168ms/step - loss: 0.4477 - categorical_accuracy: 0.9393\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.0803 - categorical_accuracy: 0.9800 - val_loss: 0.4477 - val_categorical_accuracy: 0.9393\n",
      "Epoch 374/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4873 - categorical_accuracy: 0.9143\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.0550 - categorical_accuracy: 0.9864 - val_loss: 0.4873 - val_categorical_accuracy: 0.9143\n",
      "Epoch 375/400\n",
      "18/18 [==============================] - 3s 169ms/step - loss: 0.3055 - categorical_accuracy: 0.9464\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.0802 - categorical_accuracy: 0.9804 - val_loss: 0.3055 - val_categorical_accuracy: 0.9464\n",
      "Epoch 376/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.3017 - categorical_accuracy: 0.9429\n",
      "175/175 [==============================] - 90s 515ms/step - loss: 0.1054 - categorical_accuracy: 0.9743 - val_loss: 0.3017 - val_categorical_accuracy: 0.9429\n",
      "Epoch 377/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4164 - categorical_accuracy: 0.9250\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1019 - categorical_accuracy: 0.9693 - val_loss: 0.4164 - val_categorical_accuracy: 0.9250\n",
      "Epoch 378/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.5738 - categorical_accuracy: 0.9179\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1007 - categorical_accuracy: 0.9754 - val_loss: 0.5738 - val_categorical_accuracy: 0.9179\n",
      "Epoch 379/400\n",
      "18/18 [==============================] - 3s 168ms/step - loss: 0.3325 - categorical_accuracy: 0.9464\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.1284 - categorical_accuracy: 0.9689 - val_loss: 0.3325 - val_categorical_accuracy: 0.9464\n",
      "Epoch 380/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.3707 - categorical_accuracy: 0.9321\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.0767 - categorical_accuracy: 0.9850 - val_loss: 0.3707 - val_categorical_accuracy: 0.9321\n",
      "Epoch 381/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.4947 - categorical_accuracy: 0.9214\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0815 - categorical_accuracy: 0.9789 - val_loss: 0.4947 - val_categorical_accuracy: 0.9214\n",
      "Epoch 382/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.4487 - categorical_accuracy: 0.9357\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0607 - categorical_accuracy: 0.9850 - val_loss: 0.4487 - val_categorical_accuracy: 0.9357\n",
      "Epoch 383/400\n",
      "18/18 [==============================] - 3s 167ms/step - loss: 0.4327 - categorical_accuracy: 0.9250\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1320 - categorical_accuracy: 0.9679 - val_loss: 0.4327 - val_categorical_accuracy: 0.9250\n",
      "Epoch 384/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.5023 - categorical_accuracy: 0.8929\n",
      "175/175 [==============================] - 90s 513ms/step - loss: 0.0943 - categorical_accuracy: 0.9739 - val_loss: 0.5023 - val_categorical_accuracy: 0.8929\n",
      "Epoch 385/400\n",
      "18/18 [==============================] - 3s 164ms/step - loss: 0.4864 - categorical_accuracy: 0.9036\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.1213 - categorical_accuracy: 0.9707 - val_loss: 0.4864 - val_categorical_accuracy: 0.9036\n",
      "Epoch 386/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.4091 - categorical_accuracy: 0.9321\n",
      "175/175 [==============================] - 90s 512ms/step - loss: 0.0811 - categorical_accuracy: 0.9796 - val_loss: 0.4091 - val_categorical_accuracy: 0.9321\n",
      "Epoch 387/400\n",
      "18/18 [==============================] - 3s 166ms/step - loss: 0.2577 - categorical_accuracy: 0.9250\n",
      "175/175 [==============================] - 91s 517ms/step - loss: 0.1249 - categorical_accuracy: 0.9721 - val_loss: 0.2577 - val_categorical_accuracy: 0.9250\n",
      "Epoch 388/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.2504 - categorical_accuracy: 0.9429\n",
      "175/175 [==============================] - 90s 514ms/step - loss: 0.0679 - categorical_accuracy: 0.9839 - val_loss: 0.2504 - val_categorical_accuracy: 0.9429\n",
      "Epoch 389/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.4251 - categorical_accuracy: 0.9107\n",
      "175/175 [==============================] - 89s 510ms/step - loss: 0.0852 - categorical_accuracy: 0.9775 - val_loss: 0.4251 - val_categorical_accuracy: 0.9107\n",
      "Epoch 390/400\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18/18 [==============================] - 3s 163ms/step - loss: 0.3579 - categorical_accuracy: 0.9393\n",
      "175/175 [==============================] - 89s 510ms/step - loss: 0.0720 - categorical_accuracy: 0.9836 - val_loss: 0.3579 - val_categorical_accuracy: 0.9393\n",
      "Epoch 391/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.3610 - categorical_accuracy: 0.9321\n",
      "175/175 [==============================] - 89s 510ms/step - loss: 0.1144 - categorical_accuracy: 0.9707 - val_loss: 0.3610 - val_categorical_accuracy: 0.9321\n",
      "Epoch 392/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.3835 - categorical_accuracy: 0.9250\n",
      "175/175 [==============================] - 89s 510ms/step - loss: 0.0943 - categorical_accuracy: 0.9757 - val_loss: 0.3835 - val_categorical_accuracy: 0.9250\n",
      "Epoch 393/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.4635 - categorical_accuracy: 0.9143\n",
      "175/175 [==============================] - 89s 510ms/step - loss: 0.0932 - categorical_accuracy: 0.9732 - val_loss: 0.4635 - val_categorical_accuracy: 0.9143\n",
      "Epoch 394/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.5422 - categorical_accuracy: 0.9214\n",
      "175/175 [==============================] - 89s 510ms/step - loss: 0.0731 - categorical_accuracy: 0.9839 - val_loss: 0.5422 - val_categorical_accuracy: 0.9214\n",
      "Epoch 395/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.2922 - categorical_accuracy: 0.9286\n",
      "175/175 [==============================] - 89s 510ms/step - loss: 0.0971 - categorical_accuracy: 0.9725 - val_loss: 0.2922 - val_categorical_accuracy: 0.9286\n",
      "Epoch 396/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.3486 - categorical_accuracy: 0.9250\n",
      "175/175 [==============================] - 89s 510ms/step - loss: 0.1230 - categorical_accuracy: 0.9743 - val_loss: 0.3486 - val_categorical_accuracy: 0.9250\n",
      "Epoch 397/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.3489 - categorical_accuracy: 0.9214\n",
      "175/175 [==============================] - 89s 510ms/step - loss: 0.1068 - categorical_accuracy: 0.9754 - val_loss: 0.3489 - val_categorical_accuracy: 0.9214\n",
      "Epoch 398/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.6735 - categorical_accuracy: 0.8929\n",
      "175/175 [==============================] - 89s 510ms/step - loss: 0.0838 - categorical_accuracy: 0.9761 - val_loss: 0.6735 - val_categorical_accuracy: 0.8929\n",
      "Epoch 399/400\n",
      "18/18 [==============================] - 3s 162ms/step - loss: 0.4706 - categorical_accuracy: 0.9214\n",
      "175/175 [==============================] - 89s 510ms/step - loss: 0.0941 - categorical_accuracy: 0.9779 - val_loss: 0.4706 - val_categorical_accuracy: 0.9214\n",
      "Epoch 400/400\n",
      "18/18 [==============================] - 3s 163ms/step - loss: 0.3644 - categorical_accuracy: 0.9250\n",
      "175/175 [==============================] - 89s 510ms/step - loss: 0.0804 - categorical_accuracy: 0.9786 - val_loss: 0.3644 - val_categorical_accuracy: 0.9250\n"
     ]
    }
   ],
   "source": [
    "import math\n",
    "\n",
    "history = indusNetModel.fit_generator(\n",
    "      train_generator,\n",
    "      steps_per_epoch=math.ceil(number_of_training_samples//batch_size),\n",
    "      epochs=400,\n",
    "      callbacks=callbacks_list,\n",
    "      validation_data=validation_generator,\n",
    "      validation_steps=math.ceil(number_of_validation_samples//batch_size))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "acc=history.history['categorical_accuracy']\n",
    "val_acc=history.history['val_categorical_accuracy']\n",
    "loss=history.history['loss']\n",
    "val_loss=history.history['val_loss']\n",
    "epochs=range(1,len(acc)+1)\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.plot(epochs, acc, 'bo', label='Training acc')\n",
    "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend()\n",
    "plt.figure()\n",
    "\n",
    "plt.plot(epochs, loss, 'bo', label='Training loss')\n",
    "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.python.keras.models import load_model\n",
    "\n",
    "best_model = load_model('IndusNet_7_classes_100_100_1st Experiment.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "35/35 [==============================] - 6s 169ms/step - loss: 0.5791 - categorical_accuracy: 0.9161\n",
      "Test-set accuracy: 91.61%\n"
     ]
    }
   ],
   "source": [
    "from tensorflow.python.keras.models import load_model\n",
    "\n",
    "steps_test=int(number_of_test_samples/batch_size)\n",
    "result = best_model.evaluate_generator(test_generator, steps=steps_test,verbose=1)\n",
    "print(\"Test-set accuracy: {0:.2%}\".format(result[1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total number of test examples:  (560, 100, 100, 3)\n",
      "Total number of labels: (560, 1)\n"
     ]
    }
   ],
   "source": [
    "# Preparing test data\n",
    "import glob\n",
    "from pathlib import Path\n",
    "from tensorflow.python.keras.utils import to_categorical\n",
    "import cv2\n",
    "\n",
    "apple_cases_dir = Path('D:\\\\TensorFlow Programs\\\\fruit_image_classification_Indus Fruit Dataset\\\\Indus Fruit Dataset_Resized Images_100_100\\\\Test\\\\apple')\n",
    "banana_cases_dir = Path('D:\\\\TensorFlow Programs\\\\fruit_image_classification_Indus Fruit Dataset\\\\Indus Fruit Dataset_Resized Images_100_100\\\\Test\\\\banana')\n",
    "orange_cases_dir = Path('D:\\\\TensorFlow Programs\\\\fruit_image_classification_Indus Fruit Dataset\\\\Indus Fruit Dataset_Resized Images_100_100\\\\Test\\\\orange')\n",
    "watermelon_cases_dir = Path('D:\\\\TensorFlow Programs\\\\fruit_image_classification_Indus Fruit Dataset\\\\Indus Fruit Dataset_Resized Images_100_100\\\\Test\\\\watermelon')\n",
    "mango_cases_dir = Path('D:\\\\TensorFlow Programs\\\\fruit_image_classification_Indus Fruit Dataset\\\\Indus Fruit Dataset_Resized Images_100_100\\\\Test\\\\mango')\n",
    "grape_cases_dir = Path('D:\\\\TensorFlow Programs\\\\fruit_image_classification_Indus Fruit Dataset\\\\Indus Fruit Dataset_Resized Images_100_100\\\\Test\\\\grape')\n",
    "strawberry_cases_dir = Path('D:\\\\TensorFlow Programs\\\\fruit_image_classification_Indus Fruit Dataset\\\\Indus Fruit Dataset_Resized Images_100_100\\\\Test\\\\strawberry')\n",
    "\n",
    "apple_cases = apple_cases_dir.glob('*.jpg')\n",
    "banana_cases = banana_cases_dir.glob('*.jpg')\n",
    "orange_cases = orange_cases_dir.glob('*.jpg')\n",
    "watermelon_cases = watermelon_cases_dir.glob('*.jpg')\n",
    "mango_cases = mango_cases_dir.glob('*.jpg')\n",
    "grape_cases = grape_cases_dir.glob('*.jpg')\n",
    "strawberry_cases = strawberry_cases_dir.glob('*.jpg')\n",
    "\n",
    "test_data = []\n",
    "test_labels = []\n",
    "\n",
    "for img in apple_cases:\n",
    "    img = cv2.imread(str(img))\n",
    "    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)\n",
    "    img = img.astype(np.float32)/255.\n",
    "    label = [0]\n",
    "    test_data.append(img)\n",
    "    test_labels.append(label)\n",
    "\n",
    "for img in banana_cases:\n",
    "    img = cv2.imread(str(img))\n",
    "    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)\n",
    "    img = img.astype(np.float32)/255.\n",
    "    label = [1]\n",
    "    test_data.append(img)\n",
    "    test_labels.append(label)\n",
    "\n",
    "for img in grape_cases:\n",
    "    img = cv2.imread(str(img))\n",
    "    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)\n",
    "    img = img.astype(np.float32)/255.\n",
    "    label = [2]\n",
    "    test_data.append(img)\n",
    "    test_labels.append(label)\n",
    "    \n",
    "for img in mango_cases:\n",
    "    img = cv2.imread(str(img))\n",
    "    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)\n",
    "    img = img.astype(np.float32)/255.\n",
    "    label = [3]\n",
    "    test_data.append(img)\n",
    "    test_labels.append(label)\n",
    "\n",
    "for img in orange_cases:\n",
    "    img = cv2.imread(str(img))\n",
    "    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)\n",
    "    img = img.astype(np.float32)/255.\n",
    "    label = [4]\n",
    "    test_data.append(img)\n",
    "    test_labels.append(label)\n",
    "\n",
    "for img in strawberry_cases:\n",
    "    img = cv2.imread(str(img))\n",
    "    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)\n",
    "    img = img.astype(np.float32)/255.\n",
    "    label = [5]\n",
    "    test_data.append(img)\n",
    "    test_labels.append(label)\n",
    "\n",
    "for img in watermelon_cases:\n",
    "    img = cv2.imread(str(img))\n",
    "    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)\n",
    "    img = img.astype(np.float32)/255.\n",
    "    label = [6]\n",
    "    test_data.append(img)\n",
    "    test_labels.append(label)\n",
    "\n",
    "test_data = np.array(test_data)\n",
    "test_labels = np.array(test_labels)\n",
    "test_labels_one_hot_encoding=to_categorical(test_labels)\n",
    "\n",
    "print(\"Total number of test examples: \", test_data.shape)\n",
    "print(\"Total number of labels:\", test_labels.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "560/560 [==============================] - 10s 18ms/sample - loss: 0.5791 - categorical_accuracy: 0.9161\n",
      "Loss on test set:  0.5790533946694008\n",
      "Accuracy on test set:  0.9160714\n"
     ]
    }
   ],
   "source": [
    "# Evaluation on test dataset\n",
    "\n",
    "test_loss, test_score = best_model.evaluate(test_data, test_labels_one_hot_encoding, batch_size=8)\n",
    "print(\"Loss on test set: \", test_loss)\n",
    "print(\"Accuracy on test set: \", test_score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "560/560 [==============================] - 7s 12ms/sample\n",
      "(560, 1)\n",
      "(560, 1)\n"
     ]
    }
   ],
   "source": [
    "# Get predictions\n",
    "preds = best_model.predict(test_data, batch_size=16,verbose=1)\n",
    "preds=np.around(preds)\n",
    "orig_test_labels=test_labels\n",
    "predicted_labels=np.argmax(preds,axis=1)\n",
    "predicted_labels=np.reshape(predicted_labels,(560,1))\n",
    "print(predicted_labels.shape)\n",
    "print(orig_test_labels.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "# Get the confusion matrix\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from mlxtend.plotting import plot_confusion_matrix\n",
    "\n",
    "cm  = confusion_matrix(orig_test_labels, predicted_labels)\n",
    "plt.figure()\n",
    "plot_confusion_matrix(cm,figsize=(12,10), hide_ticks=True,cmap=plt.cm.Blues)\n",
    "plt.xticks(range(7), ['apple', 'banana','grape','mango','orange','strawberry','watermelon'], fontsize=12)\n",
    "plt.yticks(range(7), ['apple', 'banana','grape','mango','orange','strawberry','watermelon'], fontsize=12)\n",
    "plt.show()"
   ]
  }
 ],
 "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": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}