1632 lines (1631 with data), 194.5 kB

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "e59fbb40",
   "metadata": {},
   "source": [
    "# Free GPU memory"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "cbe5999d",
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf \n",
    "physical_devices = tf.config.list_physical_devices('GPU') \n",
    "tf.config.experimental.set_memory_growth(physical_devices[0], True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "30c82b1d",
   "metadata": {},
   "source": [
    "# Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e2af1a8f",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import os\n",
    "import cv2\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "import operator\n",
    "import tensorflow as tf\n",
    "import random\n",
    "from keras.preprocessing.image import ImageDataGenerator"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f9f17229",
   "metadata": {},
   "source": [
    "# Directory"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "cca80eb2",
   "metadata": {},
   "outputs": [],
   "source": [
    "TRAIN_PATH = r'F:\\Leuk study re-designed\\ALLIDB-2\\Low imbalance\\Train - 1 to 10 ratio\\all'\n",
    "\n",
    "BATCH_SIZE=4"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bda39b68",
   "metadata": {},
   "source": [
    "# Data loader"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "26e1d227",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F:\\Leuk study re-designed\\ALLIDB-2\\Low imbalance\\Train - 1 to 10 ratio\\all\\Im030_1.tif\n",
      "F:\\Leuk study re-designed\\ALLIDB-2\\Low imbalance\\Train - 1 to 10 ratio\\all\\Im036_1.tif\n",
      "F:\\Leuk study re-designed\\ALLIDB-2\\Low imbalance\\Train - 1 to 10 ratio\\all\\Im044_1.tif\n",
      "F:\\Leuk study re-designed\\ALLIDB-2\\Low imbalance\\Train - 1 to 10 ratio\\all\\Im060_1.tif\n",
      "F:\\Leuk study re-designed\\ALLIDB-2\\Low imbalance\\Train - 1 to 10 ratio\\all\\Im074_1.tif\n",
      "F:\\Leuk study re-designed\\ALLIDB-2\\Low imbalance\\Train - 1 to 10 ratio\\all\\Im093_1.tif\n",
      "F:\\Leuk study re-designed\\ALLIDB-2\\Low imbalance\\Train - 1 to 10 ratio\\all\\Im098_1.tif\n",
      "F:\\Leuk study re-designed\\ALLIDB-2\\Low imbalance\\Train - 1 to 10 ratio\\all\\Im113_1.tif\n",
      "F:\\Leuk study re-designed\\ALLIDB-2\\Low imbalance\\Train - 1 to 10 ratio\\all\\Im120_1.tif\n",
      "F:\\Leuk study re-designed\\ALLIDB-2\\Low imbalance\\Train - 1 to 10 ratio\\all\\Im123_1.tif\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1080x1080 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import skimage\n",
    "from skimage.io import imread, imshow\n",
    "data = []\n",
    "i = 0\n",
    "CATEGORIES = r'all'\n",
    "plt.figure(figsize=(15, 15))\n",
    "\n",
    "list_img = os.listdir(TRAIN_PATH)\n",
    "\n",
    "for img in list_img:\n",
    "    img_path = os.path.join(TRAIN_PATH, img)\n",
    "    print(img_path)\n",
    "\n",
    "\n",
    "    arr = imread(img_path)\n",
    "\n",
    "    data.append(arr)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "48967a20",
   "metadata": {},
   "source": [
    "# Create train data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "22825e95",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(10, 257, 257, 3)\n",
      "uint8\n",
      "(10, 452, 452, 3)\n",
      "(10, 452, 452, 3)\n",
      "float32\n"
     ]
    }
   ],
   "source": [
    "random.shuffle(data)\n",
    "\n",
    "x_train = []\n",
    "\n",
    "\n",
    "for features in data:\n",
    "    x_train.append(features)\n",
    "    \n",
    "\n",
    "    \n",
    "x_train = np.array(x_train)\n",
    "\n",
    "print(x_train.shape)\n",
    "print(x_train.dtype)\n",
    "\n",
    "r=c=452\n",
    "import skimage\n",
    "from skimage.transform import resize\n",
    "new_x_train = np.zeros((x_train.shape[0], r, c, 3), dtype=np.uint8)\n",
    "print(new_x_train.shape)\n",
    "i=0\n",
    "for img in x_train:\n",
    "    image = resize(img,(r, c, 3), preserve_range=True)\n",
    "    new_x_train[i] = image\n",
    "    i+=1\n",
    "\n",
    "\n",
    "new_x_train = new_x_train/255.0\n",
    "new_x_train = new_x_train.astype('float32')\n",
    "print(new_x_train.shape)\n",
    "print(new_x_train.dtype)\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "37a3a302",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x2477b1902b0>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import skimage\n",
    "from skimage.io import imread, imshow\n",
    "imshow(new_x_train[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0642aec2",
   "metadata": {},
   "source": [
    "# Autoencoder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "77eafb56",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"functional_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         [(None, 452, 452, 3)]     0         \n",
      "_________________________________________________________________\n",
      "conv2d (Conv2D)              (None, 452, 452, 16)      448       \n",
      "_________________________________________________________________\n",
      "max_pooling2d (MaxPooling2D) (None, 226, 226, 16)      0         \n",
      "_________________________________________________________________\n",
      "conv2d_1 (Conv2D)            (None, 226, 226, 8)       1160      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 113, 113, 8)       0         \n",
      "_________________________________________________________________\n",
      "conv2d_2 (Conv2D)            (None, 113, 113, 8)       584       \n",
      "_________________________________________________________________\n",
      "max_pooling2d_2 (MaxPooling2 (None, 57, 57, 8)         0         \n",
      "_________________________________________________________________\n",
      "conv2d_3 (Conv2D)            (None, 57, 57, 8)         584       \n",
      "_________________________________________________________________\n",
      "up_sampling2d (UpSampling2D) (None, 114, 114, 8)       0         \n",
      "_________________________________________________________________\n",
      "conv2d_4 (Conv2D)            (None, 114, 114, 8)       584       \n",
      "_________________________________________________________________\n",
      "up_sampling2d_1 (UpSampling2 (None, 228, 228, 8)       0         \n",
      "_________________________________________________________________\n",
      "conv2d_5 (Conv2D)            (None, 226, 226, 16)      1168      \n",
      "_________________________________________________________________\n",
      "up_sampling2d_2 (UpSampling2 (None, 452, 452, 16)      0         \n",
      "_________________________________________________________________\n",
      "conv2d_6 (Conv2D)            (None, 452, 452, 3)       435       \n",
      "=================================================================\n",
      "Total params: 4,963\n",
      "Trainable params: 4,963\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "\n",
    "import keras\n",
    "from keras import layers\n",
    "\n",
    "input_img = keras.Input(shape=(452, 452, 3))\n",
    "\n",
    "x = layers.Conv2D(16, (3, 3), activation='relu', padding='same')(input_img)\n",
    "x = layers.MaxPooling2D((2, 2), padding='same')(x)\n",
    "x = layers.Conv2D(8, (3, 3), activation='relu', padding='same')(x)\n",
    "x = layers.MaxPooling2D((2, 2), padding='same')(x)\n",
    "x = layers.Conv2D(8, (3, 3), activation='relu', padding='same')(x)\n",
    "encoded = layers.MaxPooling2D((2, 2), padding='same')(x)\n",
    "\n",
    "# at this point the representation is (4, 4, 8) i.e. 128-dimensional\n",
    "\n",
    "x = layers.Conv2D(8, (3, 3), activation='relu', padding='same')(encoded)\n",
    "x = layers.UpSampling2D((2, 2))(x)\n",
    "x = layers.Conv2D(8, (3, 3), activation='relu', padding='same')(x)\n",
    "x = layers.UpSampling2D((2, 2))(x)\n",
    "x = layers.Conv2D(16, (3, 3), activation='relu')(x)\n",
    "x = layers.UpSampling2D((2, 2))(x)\n",
    "decoded = layers.Conv2D(3, (3, 3), activation='sigmoid', padding='same')(x)\n",
    "\n",
    "autoencoder = keras.Model(input_img, decoded)\n",
    "autoencoder.compile(optimizer='Adagrad', loss='binary_crossentropy') #metrics=['accuracy']\n",
    "autoencoder.summary()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2109799b",
   "metadata": {},
   "source": [
    "# Train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "945e2be9",
   "metadata": {},
   "outputs": [],
   "source": [
    "#import keras\n",
    "import tensorflow.keras as keras\n",
    "#from tensorflow.keras.callbacks import ModelCheckpoint\n",
    "import tensorflow as tf\n",
    "callbacks = [\n",
    "    tf.keras.callbacks.ModelCheckpoint('train_norm_autoencoder_200.h5', monitor='loss', save_best_only=True, mode='min'),\n",
    "    tf.keras.callbacks.ReduceLROnPlateau(monitor='loss', factor=0.0, verbose=1, patience=5, mode='max')]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "526e1e2a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/500\n",
      "3/3 [==============================] - 1s 174ms/step - loss: 0.6926\n",
      "Epoch 2/500\n",
      "3/3 [==============================] - 0s 46ms/step - loss: 0.6926\n",
      "Epoch 3/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 0.6926\n",
      "Epoch 4/500\n",
      "3/3 [==============================] - 0s 45ms/step - loss: 0.6926\n",
      "Epoch 5/500\n",
      "3/3 [==============================] - 0s 59ms/step - loss: 0.6926\n",
      "Epoch 6/500\n",
      "3/3 [==============================] - ETA: 0s - loss: 0.6925\n",
      "Epoch 00006: ReduceLROnPlateau reducing learning rate to 0.0.\n",
      "3/3 [==============================] - 0s 59ms/step - loss: 0.6925\n",
      "Epoch 7/500\n",
      "3/3 [==============================] - 0s 42ms/step - loss: 0.6925\n",
      "Epoch 8/500\n",
      "3/3 [==============================] - 0s 50ms/step - loss: 0.6925\n",
      "Epoch 9/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 10/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 11/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 12/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 13/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 14/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 15/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 16/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 17/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 18/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 19/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 20/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 21/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 22/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 23/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 24/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 25/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 26/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 27/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 28/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 29/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 30/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 31/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 32/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 33/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 34/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 35/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 36/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 37/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 38/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 39/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 40/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 41/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 42/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 43/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 44/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 45/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 46/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 47/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 48/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 49/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 50/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 51/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 52/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 53/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 54/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 55/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 56/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 57/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 58/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 59/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 60/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 61/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 62/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 63/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 64/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 65/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 66/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 67/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 68/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 69/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 70/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 71/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 72/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 73/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 74/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 75/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 76/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 77/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 78/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 79/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 80/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 81/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 82/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 83/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 84/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 85/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 86/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 87/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 88/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 89/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 90/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 91/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 92/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 93/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 94/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 95/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 96/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 97/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 98/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 99/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 100/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 101/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 102/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 103/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 104/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 105/500\n",
      "3/3 [==============================] - 0s 64ms/step - loss: 0.6925\n",
      "Epoch 106/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 107/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 108/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 109/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 110/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 111/500\n",
      "3/3 [==============================] - 0s 26ms/step - loss: 0.6925\n",
      "Epoch 112/500\n",
      "3/3 [==============================] - 0s 26ms/step - loss: 0.6925\n",
      "Epoch 113/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 114/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 115/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 116/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 117/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 118/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 119/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 120/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 121/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 122/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 123/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 124/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 125/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 126/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 127/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 128/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 129/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 130/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 131/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 132/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 133/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 134/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 135/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 136/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 137/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 138/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 139/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 140/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 141/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 142/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 143/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 144/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 145/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 146/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 147/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 148/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 149/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 150/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 151/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 152/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 153/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 154/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 155/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 156/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 157/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 158/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 159/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 160/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 161/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 162/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 163/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 164/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 165/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 166/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 167/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 168/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 169/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 170/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 171/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 172/500\n",
      "3/3 [==============================] - 0s 26ms/step - loss: 0.6925\n",
      "Epoch 173/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 174/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 175/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 176/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 177/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 178/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 179/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 180/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 181/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 182/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 183/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 184/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 185/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 186/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 187/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 188/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 189/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 190/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 191/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 192/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 193/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 194/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 195/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 196/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 197/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 198/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 199/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 200/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 201/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 202/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 203/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 204/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 205/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 206/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 207/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 208/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 209/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 210/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 211/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 212/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 213/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 214/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 215/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 216/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 217/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 218/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 219/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 220/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 221/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 222/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 223/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 224/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 225/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 226/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 227/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 228/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 229/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 230/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 231/500\n",
      "3/3 [==============================] - 0s 27ms/step - loss: 0.6925\n",
      "Epoch 232/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 233/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 234/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 235/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 236/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 237/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 238/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 239/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 240/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 241/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 242/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 243/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 244/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 245/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 246/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 247/500\n",
      "3/3 [==============================] - 0s 26ms/step - loss: 0.6925\n",
      "Epoch 248/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 249/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 250/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 251/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 252/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 253/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 254/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 255/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 256/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 257/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 258/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 259/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 260/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 261/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 262/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 263/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 264/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 265/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 266/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 267/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 268/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 269/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 270/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 271/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 272/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 273/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 274/500\n",
      "3/3 [==============================] - 0s 26ms/step - loss: 0.6925\n",
      "Epoch 275/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 276/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 277/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 278/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 279/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 280/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 281/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 282/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 283/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 284/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 285/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 286/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 287/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 288/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 289/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 290/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 291/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 292/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 293/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 294/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 295/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 296/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 297/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 298/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 299/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 300/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 301/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 302/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 303/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 304/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 305/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 306/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 307/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 308/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 309/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 310/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 311/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 312/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 313/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 314/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 315/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 316/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 317/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 318/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 319/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 320/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 321/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 322/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 323/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 324/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 325/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 326/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 327/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 328/500\n",
      "3/3 [==============================] - 0s 26ms/step - loss: 0.6925\n",
      "Epoch 329/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 330/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 331/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 332/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 333/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 334/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 335/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 336/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 337/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 338/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 339/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 340/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 341/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 342/500\n",
      "3/3 [==============================] - 0s 26ms/step - loss: 0.6925\n",
      "Epoch 343/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 344/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 345/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 346/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 347/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 348/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 349/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 350/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 351/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 352/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 353/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 354/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 355/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 356/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 357/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 358/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 359/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 360/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 361/500\n",
      "3/3 [==============================] - 0s 26ms/step - loss: 0.6925\n",
      "Epoch 362/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 363/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 364/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 365/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 366/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 367/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 368/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 369/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 370/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 371/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 372/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 373/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 374/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 375/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 376/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 377/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 378/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 379/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 380/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 381/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 382/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 383/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 384/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 385/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 386/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 387/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 388/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 389/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 390/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 391/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 392/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 393/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 394/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 395/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 396/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 397/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 398/500\n",
      "3/3 [==============================] - 0s 26ms/step - loss: 0.6925\n",
      "Epoch 399/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 400/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 401/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 402/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 403/500\n",
      "3/3 [==============================] - 0s 25ms/step - loss: 0.6925\n",
      "Epoch 404/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 405/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 406/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 407/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 408/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 409/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 410/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 411/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 412/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 413/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 414/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 415/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 416/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 417/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 418/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 419/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 420/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 421/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925: 0s - loss: 0.692\n",
      "Epoch 422/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 423/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 424/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 425/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 426/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 427/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 428/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 429/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 430/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 431/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 432/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 433/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 434/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 435/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 436/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 437/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 438/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 439/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 440/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 441/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 442/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 443/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 444/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 445/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 446/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 447/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 448/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 449/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 450/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 451/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 452/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 453/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 454/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 455/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 456/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 457/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 458/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 459/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 460/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 461/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 462/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 463/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 464/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 465/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 466/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 467/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 468/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 469/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 470/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 471/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 472/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 473/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 474/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 475/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 476/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 477/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 478/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 479/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 480/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 481/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 482/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 483/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 484/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 485/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 486/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 487/500\n",
      "3/3 [==============================] - 0s 23ms/step - loss: 0.6925\n",
      "Epoch 488/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 489/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 490/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 491/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 492/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 493/500\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.6925\n",
      "Epoch 494/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 495/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 496/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 497/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 498/500\n",
      "3/3 [==============================] - 0s 22ms/step - loss: 0.6925\n",
      "Epoch 499/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n",
      "Epoch 500/500\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.6925\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x2470ab3d4f0>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "\n",
    "autoencoder.fit(new_x_train, new_x_train,\n",
    "                epochs=500,\n",
    "                batch_size=4,\n",
    "                shuffle=True,\n",
    "                callbacks = callbacks)\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1c056610",
   "metadata": {},
   "source": [
    "# Save model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "da39390c",
   "metadata": {},
   "outputs": [],
   "source": [
    "autoencoder.save('./final_hem_autoencoder.h5')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "435d93e3",
   "metadata": {},
   "source": [
    "# Prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "a6b37edd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'tensorflow.python.keras.engine.functional.Functional'>\n",
      "(338, 452, 452, 3)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "TiffWriter: data are stored as RGB with contiguous samples. Specify the 'photometric' parameter to silence this warning\n",
      "TiffWriter: data are stored as RGB with contiguous samples. Specify the 'photometric' parameter to silence this warning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'numpy.ndarray'>\n",
      "(452, 452, 3)\n",
      "Im030_1\n",
      "F:/Leuk study re-designed/ALLIDB-2/Low imbalance/Autoencoder Oversample/all/Im030_1_syn.tif\n",
      "<class 'numpy.ndarray'>\n",
      "(452, 452, 3)\n",
      "Im036_1\n",
      "F:/Leuk study re-designed/ALLIDB-2/Low imbalance/Autoencoder Oversample/all/Im036_1_syn.tif\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "TiffWriter: data are stored as RGB with contiguous samples. Specify the 'photometric' parameter to silence this warning\n",
      "TiffWriter: data are stored as RGB with contiguous samples. Specify the 'photometric' parameter to silence this warning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'numpy.ndarray'>\n",
      "(452, 452, 3)\n",
      "Im044_1\n",
      "F:/Leuk study re-designed/ALLIDB-2/Low imbalance/Autoencoder Oversample/all/Im044_1_syn.tif\n",
      "<class 'numpy.ndarray'>\n",
      "(452, 452, 3)\n",
      "Im060_1\n",
      "F:/Leuk study re-designed/ALLIDB-2/Low imbalance/Autoencoder Oversample/all/Im060_1_syn.tif\n",
      "<class 'numpy.ndarray'>\n",
      "(452, 452, 3)\n",
      "Im074_1\n",
      "F:/Leuk study re-designed/ALLIDB-2/Low imbalance/Autoencoder Oversample/all/Im074_1_syn.tif\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "TiffWriter: data are stored as RGB with contiguous samples. Specify the 'photometric' parameter to silence this warning\n",
      "TiffWriter: data are stored as RGB with contiguous samples. Specify the 'photometric' parameter to silence this warning\n",
      "TiffWriter: data are stored as RGB with contiguous samples. Specify the 'photometric' parameter to silence this warning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'numpy.ndarray'>\n",
      "(452, 452, 3)\n",
      "Im093_1\n",
      "F:/Leuk study re-designed/ALLIDB-2/Low imbalance/Autoencoder Oversample/all/Im093_1_syn.tif\n",
      "<class 'numpy.ndarray'>\n",
      "(452, 452, 3)\n",
      "Im098_1\n",
      "F:/Leuk study re-designed/ALLIDB-2/Low imbalance/Autoencoder Oversample/all/Im098_1_syn.tif\n",
      "<class 'numpy.ndarray'>\n",
      "(452, 452, 3)\n",
      "Im113_1\n",
      "F:/Leuk study re-designed/ALLIDB-2/Low imbalance/Autoencoder Oversample/all/Im113_1_syn.tif\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "TiffWriter: data are stored as RGB with contiguous samples. Specify the 'photometric' parameter to silence this warning\n",
      "TiffWriter: data are stored as RGB with contiguous samples. Specify the 'photometric' parameter to silence this warning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'numpy.ndarray'>\n",
      "(452, 452, 3)\n",
      "Im120_1\n",
      "F:/Leuk study re-designed/ALLIDB-2/Low imbalance/Autoencoder Oversample/all/Im120_1_syn.tif\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "TiffWriter: data are stored as RGB with contiguous samples. Specify the 'photometric' parameter to silence this warning\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'numpy.ndarray'>\n",
      "(452, 452, 3)\n",
      "Im123_1\n",
      "F:/Leuk study re-designed/ALLIDB-2/Low imbalance/Autoencoder Oversample/all/Im123_1_syn.tif\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'\\nx_test = new_x_train[50]\\nimshow(new_x_train[50])\\nx_test = np.expand_dims(x_test, axis=0)\\nx_test.shape\\n\\nnx_test = x_test[0,:,:,:]\\nnx_test.shape\\nimshow(nx_test)\\n'"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "''''''\n",
    "import tensorflow as tf\n",
    "from skimage.io import imsave, imshow\n",
    "from skimage.transform import resize\n",
    "model = tf.keras.models.load_model('./final_all_autoencoder.h5')\n",
    "\n",
    "img_names = os.listdir(r'F:\\Leuk study re-designed\\ALLIDB-2\\Low imbalance\\Train - 1 to 10 ratio\\all')\n",
    "img_names.sort()\n",
    "print(type(model))\n",
    "\n",
    "\n",
    "\n",
    "PATH = r'F:\\Leuk study re-designed\\ALLIDB-2\\Low imbalance\\Train - 1 to 10 ratio\\all'\n",
    "import skimage \n",
    "from skimage.io import imread\n",
    "\n",
    "set_test = np.zeros((338,452,452,3), dtype=np.uint8) #hem batch 648,  #all batch 1,219\n",
    "for x in range(len(img_names)):\n",
    "    img = imread(os.path.join(PATH, img_names[x]))\n",
    "    img = resize(img,(new_x_train.shape[1], new_x_train.shape[2], new_x_train.shape[3]), preserve_range=True)\n",
    "    set_test[x] = img\n",
    "\n",
    "set_test = set_test/255.0\n",
    "set_test = set_test.astype('float32')\n",
    "print(set_test.shape)\n",
    "    \n",
    "    \n",
    "    \n",
    "test_preds = np.zeros((new_x_train.shape[0], new_x_train.shape[1], new_x_train.shape[2], new_x_train.shape[3]), dtype=np.float32)\n",
    "\n",
    "\n",
    "for x in range(new_x_train.shape[0]):\n",
    "    x_test = set_test[x]\n",
    "\n",
    "    x_test = np.expand_dims(x_test, axis=0)\n",
    "\n",
    "    test_preds[x] = model.predict(x_test)\n",
    "    print(type(test_preds[x]))\n",
    "    print(test_preds[x].shape)\n",
    "    \n",
    "  \n",
    "    print(img_names[x][:-4])\n",
    "    path = 'F:/Leuk study re-designed/ALLIDB-2/Low imbalance/Autoencoder Oversample/all' + '/' + img_names[x][:-4] + '_syn.tif'\n",
    "    print(path)\n",
    "    imsave(path, resize(test_preds[x].astype('float32'), (450, 450, 3), preserve_range=True))\n",
    "\n",
    "   \n",
    "    \n",
    "\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "leukemia",
   "language": "python",
   "name": "leukemia"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}