hw09.ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"HW09_106306062_資管三_郭宇雋.ipynb","provenance":[{"file_id":"1hCDXAqKSTjHkhqtLcfi9kltEiJZR4GN-","timestamp":1591967357803},{"file_id":"1eOp6TqZMPsffGm6GcFrXmVPwAdaxSjMK","timestamp":1590844708377}],"collapsed_sections":[],"toc_visible":true},"kernelspec":{"name":"python3","display_name":"Python 3"},"accelerator":"GPU"},"cells":[{"cell_type":"markdown","metadata":{"id":"njbVq0jShEBz","colab_type":"text"},"source":["# Using CNN to classifiy malicious PE files into malware family (With AutoEncoder)\n","\n","Please take a careful look at the following tutorial.\n","\n","https://www.tensorflow.org/tutorials/keras/classification\n","\n","It is an example of using tensorflow/keras to build a simple neural network to classify images of clothing. The images are 28x28 NumPy arrays, with pixel values ranging from 0 to 255. The labels are an array of integers, ranging from 0 to 9.\n","\n","Please write your homework by modifying this tutorial. We will provide you the first 748 bytes of several PE files and their corresponding class labels (just like the example in the tutorial). Would you please desgin a more complicated neural network (e.g., using convolution, pooling, dense, dropout, dense, softmax, ...) to increase the classification accuracy."]},{"cell_type":"code","metadata":{"id":"LYDFyLIgkYRZ","colab_type":"code","outputId":"cf284888-9222-43bf-866c-47cae7452b77","executionInfo":{"status":"ok","timestamp":1592049283994,"user_tz":-480,"elapsed":2105,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}},"colab":{"base_uri":"https://localhost:8080/","height":88}},"source":["# TensorFlow and tf.keras\n","import tensorflow as tf\n","from tensorflow import keras\n","\n","# Helper libraries\n","import numpy as np\n","import pandas as pd\n","import matplotlib.pyplot as plt\n","\n","import seaborn as sns\n","sns.set()\n","\n","print(tf.__version__)"],"execution_count":1,"outputs":[{"output_type":"stream","text":["2.2.0\n"],"name":"stdout"},{"output_type":"stream","text":["/usr/local/lib/python3.6/dist-packages/statsmodels/tools/_testing.py:19: FutureWarning: pandas.util.testing is deprecated. Use the functions in the public API at pandas.testing instead.\n","  import pandas.util.testing as tm\n"],"name":"stderr"}]},{"cell_type":"markdown","metadata":{"id":"PDNaCNOSjDgH","colab_type":"text"},"source":["### Get pefiles"]},{"cell_type":"code","metadata":{"id":"fVDAjk_Hg_Yu","colab_type":"code","outputId":"40fd0d00-adbb-4809-d658-5a3ea7d168c8","executionInfo":{"status":"ok","timestamp":1592049286962,"user_tz":-480,"elapsed":5057,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}},"colab":{"base_uri":"https://localhost:8080/","height":326}},"source":["!wget https://drive.google.com/uc?id=1Vm3K0F8LdtvEh-8GoDtJELu5Qo179aVw -O pefiles.pickle"],"execution_count":2,"outputs":[{"output_type":"stream","text":["--2020-06-13 11:54:46--  https://drive.google.com/uc?id=1Vm3K0F8LdtvEh-8GoDtJELu5Qo179aVw\n","Resolving drive.google.com (drive.google.com)... 173.194.79.101, 173.194.79.139, 173.194.79.113, ...\n","Connecting to drive.google.com (drive.google.com)|173.194.79.101|:443... connected.\n","HTTP request sent, awaiting response... 302 Moved Temporarily\n","Location: https://doc-0c-38-docs.googleusercontent.com/docs/securesc/ha0ro937gcuc7l7deffksulhg5h7mbp1/nhkv05ve0a7pr5pinki0tkotfoubl9eq/1592049225000/17174434063503717705/*/1Vm3K0F8LdtvEh-8GoDtJELu5Qo179aVw [following]\n","Warning: wildcards not supported in HTTP.\n","--2020-06-13 11:54:46--  https://doc-0c-38-docs.googleusercontent.com/docs/securesc/ha0ro937gcuc7l7deffksulhg5h7mbp1/nhkv05ve0a7pr5pinki0tkotfoubl9eq/1592049225000/17174434063503717705/*/1Vm3K0F8LdtvEh-8GoDtJELu5Qo179aVw\n","Resolving doc-0c-38-docs.googleusercontent.com (doc-0c-38-docs.googleusercontent.com)... 108.177.119.132, 2a00:1450:4013:c00::84\n","Connecting to doc-0c-38-docs.googleusercontent.com (doc-0c-38-docs.googleusercontent.com)|108.177.119.132|:443... connected.\n","HTTP request sent, awaiting response... 200 OK\n","Length: 282207 (276K) [application/octet-stream]\n","Saving to: ‘pefiles.pickle’\n","\n","\rpefiles.pickle        0%[                    ]       0  --.-KB/s               \rpefiles.pickle      100%[===================>] 275.59K  --.-KB/s    in 0.002s  \n","\n","2020-06-13 11:54:47 (154 MB/s) - ‘pefiles.pickle’ saved [282207/282207]\n","\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"xZHA0VTPjKY1","colab_type":"code","outputId":"caf68b24-bbc5-4255-f2e7-ca878dfb090e","executionInfo":{"status":"ok","timestamp":1592049290016,"user_tz":-480,"elapsed":8042,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}},"colab":{"base_uri":"https://localhost:8080/","height":119}},"source":["!ls -al"],"execution_count":3,"outputs":[{"output_type":"stream","text":["total 292\n","drwxr-xr-x 1 root root   4096 Jun 13 11:14 .\n","drwxr-xr-x 1 root root   4096 Jun 13 11:14 ..\n","drwxr-xr-x 1 root root   4096 Jun 10 16:28 .config\n","-rw-r--r-- 1 root root 282207 Jun 13 11:54 pefiles.pickle\n","drwxr-xr-x 1 root root   4096 Jun 10 16:28 sample_data\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"-hUYU--mjpHD","colab_type":"text"},"source":["### load training data"]},{"cell_type":"code","metadata":{"id":"ZxzLOj1vjw0H","colab_type":"code","colab":{}},"source":["import pickle\n","\n","def my_load_data(p):\n","  with open(p, 'rb') as fd:\n","    tu = pickle.load(fd)\n","  return (tu[0], tu[1])"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"iRRI5NE-jMmh","colab_type":"code","colab":{}},"source":["(train_images, train_labels) = my_load_data(\"/content/pefiles.pickle\")"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"bgqyrpLOkqZx","colab_type":"code","colab":{}},"source":["# now go https://www.tensorflow.org/tutorials/keras/classification\n","# and try to understand the tutorial\n","# and write your code below.\n","# My codes is very similar to the tutorial."],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"qxZ4y3Pdk7Ig","colab_type":"code","outputId":"24e62f1f-f370-4acf-c738-b6ce2c4b80d4","executionInfo":{"status":"ok","timestamp":1592049290019,"user_tz":-480,"elapsed":7992,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}},"colab":{"base_uri":"https://localhost:8080/","height":34}},"source":["# our data includes 11 different malware families\n","class_names = ['Urausy.C', 'Elkern.B', 'Fareit', 'Hotbar', 'Gepys.A', 'Bulta!rfn', 'Zbot',\n","               'GameVance', 'Zegost.B', 'Bifrose.AE', 'Zbot!GO']\n","len_classes = len(class_names)\n","\n","print(len_classes)"],"execution_count":7,"outputs":[{"output_type":"stream","text":["11\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"V9Ncbl6IlUh1","colab_type":"text"},"source":["### Explore the data"]},{"cell_type":"code","metadata":{"id":"ZnphVAxKlVJH","colab_type":"code","outputId":"d54ed6b8-e6f5-480c-f9db-9080bfadb5fb","executionInfo":{"status":"ok","timestamp":1592049290019,"user_tz":-480,"elapsed":7949,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}},"colab":{"base_uri":"https://localhost:8080/","height":34}},"source":["# We have 356 malware pefiles, and each of them is a 28x28 byte array.\n","train_images.shape"],"execution_count":8,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(356, 28, 28)"]},"metadata":{"tags":[]},"execution_count":8}]},{"cell_type":"code","metadata":{"id":"kFVRbQCFlZdd","colab_type":"code","outputId":"3ddfc32d-8b8d-4df4-b56a-f5cdbfc80c7d","executionInfo":{"status":"ok","timestamp":1592049290020,"user_tz":-480,"elapsed":7867,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}},"colab":{"base_uri":"https://localhost:8080/","height":34}},"source":["# The lenght of the label is 356 as well.\n","len(train_labels)"],"execution_count":9,"outputs":[{"output_type":"execute_result","data":{"text/plain":["356"]},"metadata":{"tags":[]},"execution_count":9}]},{"cell_type":"code","metadata":{"id":"qiaQHLfKlmrP","colab_type":"code","outputId":"41b8f6bf-f354-442b-f10c-d40de800e099","executionInfo":{"status":"ok","timestamp":1592049290020,"user_tz":-480,"elapsed":7857,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}},"colab":{"base_uri":"https://localhost:8080/","height":374}},"source":["# Each label is an integer.\n","# The 1st malware's label is 5.\n","# It stands for that the 1st malware belongs to 'Urausy.C' family.\n","# See class_names.\n","train_labels"],"execution_count":10,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([ 5,  4,  2,  6,  3,  7,  2,  3,  0,  2,  2,  3,  7,  3,  3,  3,  2,\n","        2,  7,  9,  5,  1,  4, 10,  2,  4,  0,  6,  2,  6,  2,  3,  4,  2,\n","        2,  2,  6, 10,  2,  3,  2,  0,  2,  3,  4,  0, 10,  3,  2,  4,  2,\n","        2,  2,  9,  2,  5,  0,  5,  4,  3,  0,  2,  3,  3,  2,  3,  1,  2,\n","        0,  4,  3,  4,  3,  2,  3,  1,  4,  2,  0,  1,  6,  4,  3,  4,  2,\n","        3,  2,  5,  7,  2,  3,  3,  8,  5,  6,  4,  4,  0,  3,  6,  0,  3,\n","        7,  4,  9,  6,  2,  3,  4,  4,  1,  2, 10,  4, 10,  4,  3,  6,  8,\n","        4,  0,  0,  0,  9,  3,  4,  1,  2,  0,  2, 10,  3,  2,  3,  2,  2,\n","        4,  3,  3,  4,  2,  3,  2,  0,  8,  6,  2,  2, 10,  3,  0,  4,  5,\n","        5,  0,  3,  9,  9,  2,  0,  1,  5,  0,  6,  6,  6,  3,  9,  6, 10,\n","        3,  6,  8,  2,  4,  0,  3,  3,  6,  5,  4,  4,  4,  2,  3,  1,  1,\n","       10,  3,  3,  5,  3,  2,  6,  3,  8,  2,  2,  0,  0,  7,  2,  3,  3,\n","        1,  9,  4,  0,  3,  3,  0,  7,  4,  2,  5,  5, 10,  8,  6,  4,  0,\n","        3,  4,  4,  2,  0,  2,  2,  7,  3, 10,  3,  3,  5,  4,  6,  0,  3,\n","        2,  7,  2,  0,  1,  2,  2,  3,  3,  6,  6,  0,  5,  0,  4,  5,  2,\n","        3,  3,  2,  2,  2,  3,  2,  3,  7,  7,  2,  7,  2,  4,  3,  4,  2,\n","        2,  3,  2,  2,  1,  0,  6,  4,  3,  2,  2,  0,  6,  2, 10,  5, 10,\n","        4,  3,  4,  0,  2,  3,  2,  2,  9,  0,  2,  6,  1,  2,  0,  0,  1,\n","        6,  6,  4,  5,  4,  2,  3,  9,  5,  4,  0,  3,  2,  8,  3,  3,  2,\n","        0,  0,  3,  0,  6,  5,  0,  2,  2,  0,  0,  6,  0,  8,  3,  7,  2,\n","        6,  0,  0,  4,  3,  6,  3,  5,  3,  0,  3,  8,  3,  2,  3,  2])"]},"metadata":{"tags":[]},"execution_count":10}]},{"cell_type":"code","metadata":{"id":"Wbcbt1D0lvjQ","colab_type":"code","outputId":"fda4f4e4-c5aa-483e-9a97-c31edd7cef62","executionInfo":{"status":"ok","timestamp":1592049290023,"user_tz":-480,"elapsed":7788,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}},"colab":{"base_uri":"https://localhost:8080/","height":1000}},"source":["# An example of first PE files.\n","# It is a 28x28 integer matrix.\n","train_images[0]"],"execution_count":11,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([[ 77,  90,  80,   0,   2,   0,   0,   0,   0,   0,  15,   0, 255,\n","        255,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,  64,   0,\n","         26,   0],\n","       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0],\n","       [  0,   0,   0,   0, 124,   0,   0,   0, 186,  16,   0,  14,  31,\n","        180,   9, 205,  33, 184,   1,  76, 205,  33, 144, 144,  84, 104,\n","        105, 115],\n","       [ 32, 112, 114, 111, 103, 114,  97, 109,  32, 109, 117, 115, 116,\n","         32,  98, 101,  32, 114, 117, 110,  32, 117, 110, 100, 101, 114,\n","         32,  87],\n","       [105, 110,  51,  50,  13,  10,  36,  55,   0,   0,   0,   0,  80,\n","         69,   0,   0,  76,   1,   4,   0, 254, 114,  22,  79,   0,   0,\n","          0,   0],\n","       [  0,   0,   0,   0, 224,   0,  15,   1,  11,   1,  10,   0,   0,\n","         22,   0,   0,   0, 184,   5,   0,   0,   0,   0,   0, 180,  20,\n","          0,   0],\n","       [  0,  16,   0,   0,   0,  48,   0,   0,   0,   0,  64,   0,   0,\n","         16,   0,   0,   0,   2,   0,   0,   4,   0,   0,   0,   0,   0,\n","          0,   0],\n","       [  4,   0,   0,   0,   0,   0,   0,   0,   0, 144,  12,   0,   0,\n","          4,   0,   0,   0,   0,   0,   0,   2,   0,   0,   0,   0,   0,\n","         16,   0],\n","       [  0,  16,   0,   0,   0,   0,  16,   0,   0,  16,   0,   0,   0,\n","          0,   0,   0,  16,   0,   0,   0,  68,  30,   0,   0,  28,   0,\n","          0,   0],\n","       [176,  48,   0,   0, 120,   0,   0,   0,   0, 224,   6,   0, 110,\n","        170,   5,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0],\n","       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0],\n","       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0],\n","       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0],\n","       [  0,   0,   0,   0,   0,   0,   0,   0,  46, 116, 101, 120, 116,\n","          0,   0,   0,  64,  21,   0,   0,   0,  16,   0,   0,   0,  22,\n","          0,   0],\n","       [  0,   4,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0,   0,  32,   0,   0,  96,  46, 114, 100,  97, 116,  97,\n","          0,   0],\n","       [  0, 160,   6,   0,   0,  48,   0,   0,   0,   6,   0,   0,   0,\n","         26,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0],\n","       [ 64,   0,   0, 192,  46, 100,  97, 116,  97,   0,   0,   0,  10,\n","          4,   0,   0,   0, 208,   6,   0,   0,   6,   0,   0,   0,  32,\n","          0,   0],\n","       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,  64,\n","          0,   0, 192,  46, 114, 115, 114,  99,   0,   0,   0, 110, 170,\n","          5,   0],\n","       [  0, 224,   6,   0,   0, 172,   5,   0,   0,  38,   0,   0,   0,\n","          0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,  64,   0,\n","          0,  64],\n","       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0],\n","       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0],\n","       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0],\n","       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0],\n","       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0],\n","       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0],\n","       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0],\n","       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0],\n","       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n","          0,   0]], dtype=uint8)"]},"metadata":{"tags":[]},"execution_count":11}]},{"cell_type":"markdown","metadata":{"id":"EdwrMKuhnRPV","colab_type":"text"},"source":["### Preprocess the data"]},{"cell_type":"code","metadata":{"id":"mM8aua7tmOT7","colab_type":"code","outputId":"b918f7e9-f8b4-4052-fd18-3b51d17332c3","executionInfo":{"status":"ok","timestamp":1592049291194,"user_tz":-480,"elapsed":8942,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}},"colab":{"base_uri":"https://localhost:8080/","height":268}},"source":["plt.figure()\n","plt.imshow(train_images[0])\n","plt.colorbar()\n","plt.grid(False)\n","plt.show()\n","\n","\n","# Note that we view the first 28x28 bytes of a PE file as an image.\n","# A byte with value '0x00' has darkest color,\n","# whilw a byte with value '0xFF' has brightest color.\n","# Below is the visualization of the first PE file."],"execution_count":12,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 2 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"code","metadata":{"id":"pldW7dpcnfkh","colab_type":"code","outputId":"af532f5f-128a-44ef-ce5a-319399d3f179","executionInfo":{"status":"ok","timestamp":1592049292225,"user_tz":-480,"elapsed":9950,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}},"colab":{"base_uri":"https://localhost:8080/","height":592}},"source":["plt.figure(figsize=(10,10))\n","for i in range(36):\n","    plt.subplot(6,6,i+1)\n","    plt.xticks([])\n","    plt.yticks([])\n","    plt.grid(False)\n","    plt.imshow(train_images[i], cmap=plt.cm.binary)\n","    plt.xlabel(class_names[train_labels[i]])\n","plt.show()"],"execution_count":13,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 720x720 with 36 Axes>"]},"metadata":{"tags":[]}}]},{"cell_type":"markdown","metadata":{"id":"Snw_N4mtoSkv","colab_type":"text"},"source":["The above images are samples of different PE files. Could you visually distinguish them? Their labels are shown under each of the image.\n","\n","Q1: Before building the model, if there is anything you can do to the data, do it right here.\n"]},{"cell_type":"code","metadata":{"id":"JISk0HUfoH6U","colab_type":"code","colab":{}},"source":["# 調整 shape\n","train_images = train_images.reshape(train_images.shape[0], 28, 28, 1)\n","# Normalization (Min-Max) (From [0,255] to [0, 1])\n","train_images = train_images / 255.0"],"execution_count":0,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"m5pwDPLApAyy","colab_type":"text"},"source":["### Build AutoEncoder model"]},{"cell_type":"code","metadata":{"id":"AEl_MHrmJ8Fk","colab_type":"code","colab":{}},"source":["input_img = keras.layers.Input(shape=(28, 28, 1))\n","\n","x = keras.layers.Conv2D(32, (3, 3), activation='relu', padding='same')(input_img)\n","x = keras.layers.MaxPooling2D((2, 2), padding='same')(x)\n","x = keras.layers.Conv2D(16, (3, 3), activation='relu', padding='same')(x)\n","x = keras.layers.MaxPooling2D((2, 2), padding='same')(x)\n","x = keras.layers.Conv2D(16, (3, 3), activation='relu', padding='same')(x)\n","encoded = keras.layers.MaxPooling2D((2, 2), padding='same')(x)\n","\n","# at this point the representation is (4, 4, 16) i.e. 256-dimensional\n","\n","x = keras.layers.Conv2D(16, (3, 3), activation='relu', padding='same')(encoded)\n","x = keras.layers.UpSampling2D((2, 2))(x)\n","x = keras.layers.Conv2D(16, (3, 3), activation='relu', padding='same')(x)\n","x = keras.layers.UpSampling2D((2, 2))(x)\n","x = keras.layers.Conv2D(32, (3, 3), activation='relu')(x)\n","x = keras.layers.UpSampling2D((2, 2))(x)\n","decoded = keras.layers.Conv2D(1, (3, 3), activation='sigmoid', padding='same')(x)\n","\n","autoencoder = keras.models.Model(input_img, decoded)"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"KZXhse4DLXAk","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":629},"outputId":"43fbf0c4-5513-49fb-80a4-348651a5a4a2","executionInfo":{"status":"ok","timestamp":1592049292564,"user_tz":-480,"elapsed":10171,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}}},"source":["autoencoder.summary()"],"execution_count":16,"outputs":[{"output_type":"stream","text":["Model: \"model\"\n","_________________________________________________________________\n","Layer (type)                 Output Shape              Param #   \n","=================================================================\n","input_1 (InputLayer)         [(None, 28, 28, 1)]       0         \n","_________________________________________________________________\n","conv2d (Conv2D)              (None, 28, 28, 32)        320       \n","_________________________________________________________________\n","max_pooling2d (MaxPooling2D) (None, 14, 14, 32)        0         \n","_________________________________________________________________\n","conv2d_1 (Conv2D)            (None, 14, 14, 16)        4624      \n","_________________________________________________________________\n","max_pooling2d_1 (MaxPooling2 (None, 7, 7, 16)          0         \n","_________________________________________________________________\n","conv2d_2 (Conv2D)            (None, 7, 7, 16)          2320      \n","_________________________________________________________________\n","max_pooling2d_2 (MaxPooling2 (None, 4, 4, 16)          0         \n","_________________________________________________________________\n","conv2d_3 (Conv2D)            (None, 4, 4, 16)          2320      \n","_________________________________________________________________\n","up_sampling2d (UpSampling2D) (None, 8, 8, 16)          0         \n","_________________________________________________________________\n","conv2d_4 (Conv2D)            (None, 8, 8, 16)          2320      \n","_________________________________________________________________\n","up_sampling2d_1 (UpSampling2 (None, 16, 16, 16)        0         \n","_________________________________________________________________\n","conv2d_5 (Conv2D)            (None, 14, 14, 32)        4640      \n","_________________________________________________________________\n","up_sampling2d_2 (UpSampling2 (None, 28, 28, 32)        0         \n","_________________________________________________________________\n","conv2d_6 (Conv2D)            (None, 28, 28, 1)         289       \n","=================================================================\n","Total params: 16,833\n","Trainable params: 16,833\n","Non-trainable params: 0\n","_________________________________________________________________\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"5eRGfT0aOslb","colab_type":"code","colab":{}},"source":["# split out the Encoder model from AutoEncoder\n","encoder = keras.models.Model(input_img, encoded)"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"OUm9Lb0oPK2F","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":391},"outputId":"6fc665e1-96ce-4572-c04e-4d6e3476dff9","executionInfo":{"status":"ok","timestamp":1592049292565,"user_tz":-480,"elapsed":10150,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}}},"source":["encoder.summary()"],"execution_count":18,"outputs":[{"output_type":"stream","text":["Model: \"model_1\"\n","_________________________________________________________________\n","Layer (type)                 Output Shape              Param #   \n","=================================================================\n","input_1 (InputLayer)         [(None, 28, 28, 1)]       0         \n","_________________________________________________________________\n","conv2d (Conv2D)              (None, 28, 28, 32)        320       \n","_________________________________________________________________\n","max_pooling2d (MaxPooling2D) (None, 14, 14, 32)        0         \n","_________________________________________________________________\n","conv2d_1 (Conv2D)            (None, 14, 14, 16)        4624      \n","_________________________________________________________________\n","max_pooling2d_1 (MaxPooling2 (None, 7, 7, 16)          0         \n","_________________________________________________________________\n","conv2d_2 (Conv2D)            (None, 7, 7, 16)          2320      \n","_________________________________________________________________\n","max_pooling2d_2 (MaxPooling2 (None, 4, 4, 16)          0         \n","=================================================================\n","Total params: 7,264\n","Trainable params: 7,264\n","Non-trainable params: 0\n","_________________________________________________________________\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"fSqIZjYoLWUn","colab_type":"code","colab":{}},"source":["autoencoder.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"yGJ_xqxiK0Nb","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":1000},"outputId":"da997ec8-f241-4d48-a4bf-e42478058976","executionInfo":{"status":"ok","timestamp":1592049334852,"user_tz":-480,"elapsed":52350,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}}},"source":["# early stop : 若是連續 10 次 epochs 在 validation data 上的 loss 都沒有改善，則會停止訓練\n","callback = tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=10)\n","autoencoder.fit(train_images, train_images,\n","                epochs=500,\n","                shuffle=True,\n","                validation_split=0.2,\n","                callbacks=[callback])"],"execution_count":20,"outputs":[{"output_type":"stream","text":["Epoch 1/500\n","9/9 [==============================] - 1s 59ms/step - loss: 0.6642 - accuracy: 0.5909 - val_loss: 0.5701 - val_accuracy: 0.7052\n","Epoch 2/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.4465 - accuracy: 0.7131 - val_loss: 0.4300 - val_accuracy: 0.7052\n","Epoch 3/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.3718 - accuracy: 0.7131 - val_loss: 0.3552 - val_accuracy: 0.7052\n","Epoch 4/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.3323 - accuracy: 0.7131 - val_loss: 0.3383 - val_accuracy: 0.7052\n","Epoch 5/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.3153 - accuracy: 0.7131 - val_loss: 0.3171 - val_accuracy: 0.7052\n","Epoch 6/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.3001 - accuracy: 0.7131 - val_loss: 0.3051 - val_accuracy: 0.7052\n","Epoch 7/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2903 - accuracy: 0.7131 - val_loss: 0.2981 - val_accuracy: 0.7052\n","Epoch 8/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2838 - accuracy: 0.7131 - val_loss: 0.2938 - val_accuracy: 0.7052\n","Epoch 9/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2798 - accuracy: 0.7131 - val_loss: 0.2912 - val_accuracy: 0.7052\n","Epoch 10/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2763 - accuracy: 0.7131 - val_loss: 0.2885 - val_accuracy: 0.7052\n","Epoch 11/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2732 - accuracy: 0.7131 - val_loss: 0.2861 - val_accuracy: 0.7052\n","Epoch 12/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2705 - accuracy: 0.7131 - val_loss: 0.2824 - val_accuracy: 0.7052\n","Epoch 13/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2673 - accuracy: 0.7131 - val_loss: 0.2806 - val_accuracy: 0.7052\n","Epoch 14/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2647 - accuracy: 0.7131 - val_loss: 0.2776 - val_accuracy: 0.7052\n","Epoch 15/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2625 - accuracy: 0.7131 - val_loss: 0.2754 - val_accuracy: 0.7052\n","Epoch 16/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2607 - accuracy: 0.7131 - val_loss: 0.2726 - val_accuracy: 0.7052\n","Epoch 17/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2587 - accuracy: 0.7131 - val_loss: 0.2701 - val_accuracy: 0.7052\n","Epoch 18/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2570 - accuracy: 0.7131 - val_loss: 0.2676 - val_accuracy: 0.7052\n","Epoch 19/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2545 - accuracy: 0.7131 - val_loss: 0.2655 - val_accuracy: 0.7051\n","Epoch 20/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2527 - accuracy: 0.7131 - val_loss: 0.2629 - val_accuracy: 0.7052\n","Epoch 21/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2511 - accuracy: 0.7130 - val_loss: 0.2602 - val_accuracy: 0.7051\n","Epoch 22/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2491 - accuracy: 0.7130 - val_loss: 0.2577 - val_accuracy: 0.7051\n","Epoch 23/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2472 - accuracy: 0.7131 - val_loss: 0.2554 - val_accuracy: 0.7052\n","Epoch 24/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2452 - accuracy: 0.7131 - val_loss: 0.2533 - val_accuracy: 0.7053\n","Epoch 25/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2435 - accuracy: 0.7131 - val_loss: 0.2510 - val_accuracy: 0.7052\n","Epoch 26/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2416 - accuracy: 0.7130 - val_loss: 0.2495 - val_accuracy: 0.7051\n","Epoch 27/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2402 - accuracy: 0.7130 - val_loss: 0.2474 - val_accuracy: 0.7051\n","Epoch 28/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2389 - accuracy: 0.7130 - val_loss: 0.2465 - val_accuracy: 0.7051\n","Epoch 29/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2380 - accuracy: 0.7130 - val_loss: 0.2462 - val_accuracy: 0.7050\n","Epoch 30/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2367 - accuracy: 0.7131 - val_loss: 0.2440 - val_accuracy: 0.7053\n","Epoch 31/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2356 - accuracy: 0.7135 - val_loss: 0.2428 - val_accuracy: 0.7063\n","Epoch 32/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2344 - accuracy: 0.7141 - val_loss: 0.2415 - val_accuracy: 0.7064\n","Epoch 33/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2332 - accuracy: 0.7146 - val_loss: 0.2402 - val_accuracy: 0.7070\n","Epoch 34/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2321 - accuracy: 0.7149 - val_loss: 0.2393 - val_accuracy: 0.7071\n","Epoch 35/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.2313 - accuracy: 0.7152 - val_loss: 0.2385 - val_accuracy: 0.7073\n","Epoch 36/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2305 - accuracy: 0.7153 - val_loss: 0.2378 - val_accuracy: 0.7073\n","Epoch 37/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2296 - accuracy: 0.7153 - val_loss: 0.2368 - val_accuracy: 0.7076\n","Epoch 38/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2288 - accuracy: 0.7154 - val_loss: 0.2359 - val_accuracy: 0.7075\n","Epoch 39/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.2276 - accuracy: 0.7155 - val_loss: 0.2352 - val_accuracy: 0.7076\n","Epoch 40/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2270 - accuracy: 0.7155 - val_loss: 0.2349 - val_accuracy: 0.7075\n","Epoch 41/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2262 - accuracy: 0.7155 - val_loss: 0.2337 - val_accuracy: 0.7077\n","Epoch 42/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2252 - accuracy: 0.7156 - val_loss: 0.2329 - val_accuracy: 0.7075\n","Epoch 43/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2247 - accuracy: 0.7156 - val_loss: 0.2326 - val_accuracy: 0.7076\n","Epoch 44/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2240 - accuracy: 0.7155 - val_loss: 0.2316 - val_accuracy: 0.7077\n","Epoch 45/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2228 - accuracy: 0.7155 - val_loss: 0.2303 - val_accuracy: 0.7077\n","Epoch 46/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2217 - accuracy: 0.7155 - val_loss: 0.2294 - val_accuracy: 0.7077\n","Epoch 47/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2209 - accuracy: 0.7155 - val_loss: 0.2288 - val_accuracy: 0.7076\n","Epoch 48/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2200 - accuracy: 0.7156 - val_loss: 0.2280 - val_accuracy: 0.7077\n","Epoch 49/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2193 - accuracy: 0.7155 - val_loss: 0.2272 - val_accuracy: 0.7077\n","Epoch 50/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2184 - accuracy: 0.7155 - val_loss: 0.2266 - val_accuracy: 0.7077\n","Epoch 51/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2176 - accuracy: 0.7155 - val_loss: 0.2260 - val_accuracy: 0.7076\n","Epoch 52/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2169 - accuracy: 0.7155 - val_loss: 0.2253 - val_accuracy: 0.7076\n","Epoch 53/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2161 - accuracy: 0.7155 - val_loss: 0.2247 - val_accuracy: 0.7076\n","Epoch 54/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2154 - accuracy: 0.7155 - val_loss: 0.2243 - val_accuracy: 0.7075\n","Epoch 55/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2149 - accuracy: 0.7155 - val_loss: 0.2242 - val_accuracy: 0.7074\n","Epoch 56/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2145 - accuracy: 0.7155 - val_loss: 0.2231 - val_accuracy: 0.7076\n","Epoch 57/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2140 - accuracy: 0.7155 - val_loss: 0.2225 - val_accuracy: 0.7076\n","Epoch 58/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2132 - accuracy: 0.7155 - val_loss: 0.2220 - val_accuracy: 0.7076\n","Epoch 59/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2126 - accuracy: 0.7155 - val_loss: 0.2217 - val_accuracy: 0.7076\n","Epoch 60/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2120 - accuracy: 0.7155 - val_loss: 0.2215 - val_accuracy: 0.7076\n","Epoch 61/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2119 - accuracy: 0.7155 - val_loss: 0.2218 - val_accuracy: 0.7074\n","Epoch 62/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2114 - accuracy: 0.7154 - val_loss: 0.2208 - val_accuracy: 0.7074\n","Epoch 63/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2111 - accuracy: 0.7155 - val_loss: 0.2204 - val_accuracy: 0.7074\n","Epoch 64/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2109 - accuracy: 0.7155 - val_loss: 0.2204 - val_accuracy: 0.7074\n","Epoch 65/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2102 - accuracy: 0.7154 - val_loss: 0.2190 - val_accuracy: 0.7075\n","Epoch 66/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2093 - accuracy: 0.7155 - val_loss: 0.2189 - val_accuracy: 0.7076\n","Epoch 67/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2090 - accuracy: 0.7155 - val_loss: 0.2185 - val_accuracy: 0.7076\n","Epoch 68/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2086 - accuracy: 0.7155 - val_loss: 0.2183 - val_accuracy: 0.7075\n","Epoch 69/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2081 - accuracy: 0.7155 - val_loss: 0.2179 - val_accuracy: 0.7076\n","Epoch 70/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2078 - accuracy: 0.7155 - val_loss: 0.2175 - val_accuracy: 0.7076\n","Epoch 71/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2074 - accuracy: 0.7155 - val_loss: 0.2175 - val_accuracy: 0.7075\n","Epoch 72/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2074 - accuracy: 0.7155 - val_loss: 0.2170 - val_accuracy: 0.7075\n","Epoch 73/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2075 - accuracy: 0.7154 - val_loss: 0.2176 - val_accuracy: 0.7076\n","Epoch 74/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2070 - accuracy: 0.7154 - val_loss: 0.2172 - val_accuracy: 0.7076\n","Epoch 75/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2062 - accuracy: 0.7155 - val_loss: 0.2163 - val_accuracy: 0.7076\n","Epoch 76/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2059 - accuracy: 0.7154 - val_loss: 0.2158 - val_accuracy: 0.7076\n","Epoch 77/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2054 - accuracy: 0.7154 - val_loss: 0.2153 - val_accuracy: 0.7075\n","Epoch 78/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2049 - accuracy: 0.7155 - val_loss: 0.2151 - val_accuracy: 0.7074\n","Epoch 79/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2046 - accuracy: 0.7155 - val_loss: 0.2151 - val_accuracy: 0.7074\n","Epoch 80/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2043 - accuracy: 0.7154 - val_loss: 0.2149 - val_accuracy: 0.7073\n","Epoch 81/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2041 - accuracy: 0.7154 - val_loss: 0.2143 - val_accuracy: 0.7073\n","Epoch 82/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2035 - accuracy: 0.7154 - val_loss: 0.2137 - val_accuracy: 0.7074\n","Epoch 83/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2032 - accuracy: 0.7154 - val_loss: 0.2138 - val_accuracy: 0.7075\n","Epoch 84/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2028 - accuracy: 0.7154 - val_loss: 0.2133 - val_accuracy: 0.7074\n","Epoch 85/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2024 - accuracy: 0.7154 - val_loss: 0.2130 - val_accuracy: 0.7075\n","Epoch 86/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2023 - accuracy: 0.7154 - val_loss: 0.2130 - val_accuracy: 0.7073\n","Epoch 87/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.2020 - accuracy: 0.7154 - val_loss: 0.2124 - val_accuracy: 0.7076\n","Epoch 88/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2021 - accuracy: 0.7154 - val_loss: 0.2133 - val_accuracy: 0.7072\n","Epoch 89/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2018 - accuracy: 0.7154 - val_loss: 0.2122 - val_accuracy: 0.7074\n","Epoch 90/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.2010 - accuracy: 0.7154 - val_loss: 0.2125 - val_accuracy: 0.7072\n","Epoch 91/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2008 - accuracy: 0.7154 - val_loss: 0.2115 - val_accuracy: 0.7074\n","Epoch 92/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2004 - accuracy: 0.7154 - val_loss: 0.2111 - val_accuracy: 0.7075\n","Epoch 93/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.2001 - accuracy: 0.7154 - val_loss: 0.2117 - val_accuracy: 0.7075\n","Epoch 94/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.2003 - accuracy: 0.7154 - val_loss: 0.2110 - val_accuracy: 0.7072\n","Epoch 95/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.2002 - accuracy: 0.7153 - val_loss: 0.2109 - val_accuracy: 0.7072\n","Epoch 96/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1996 - accuracy: 0.7153 - val_loss: 0.2105 - val_accuracy: 0.7073\n","Epoch 97/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1991 - accuracy: 0.7153 - val_loss: 0.2105 - val_accuracy: 0.7071\n","Epoch 98/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1988 - accuracy: 0.7153 - val_loss: 0.2104 - val_accuracy: 0.7071\n","Epoch 99/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1984 - accuracy: 0.7153 - val_loss: 0.2096 - val_accuracy: 0.7073\n","Epoch 100/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1982 - accuracy: 0.7153 - val_loss: 0.2094 - val_accuracy: 0.7073\n","Epoch 101/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1982 - accuracy: 0.7153 - val_loss: 0.2088 - val_accuracy: 0.7073\n","Epoch 102/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1976 - accuracy: 0.7153 - val_loss: 0.2092 - val_accuracy: 0.7073\n","Epoch 103/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1972 - accuracy: 0.7153 - val_loss: 0.2086 - val_accuracy: 0.7075\n","Epoch 104/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1969 - accuracy: 0.7153 - val_loss: 0.2084 - val_accuracy: 0.7073\n","Epoch 105/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1965 - accuracy: 0.7153 - val_loss: 0.2082 - val_accuracy: 0.7073\n","Epoch 106/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1964 - accuracy: 0.7153 - val_loss: 0.2081 - val_accuracy: 0.7074\n","Epoch 107/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1964 - accuracy: 0.7153 - val_loss: 0.2078 - val_accuracy: 0.7074\n","Epoch 108/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1958 - accuracy: 0.7153 - val_loss: 0.2079 - val_accuracy: 0.7074\n","Epoch 109/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1957 - accuracy: 0.7153 - val_loss: 0.2072 - val_accuracy: 0.7073\n","Epoch 110/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1951 - accuracy: 0.7153 - val_loss: 0.2074 - val_accuracy: 0.7071\n","Epoch 111/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1951 - accuracy: 0.7153 - val_loss: 0.2069 - val_accuracy: 0.7072\n","Epoch 112/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1948 - accuracy: 0.7153 - val_loss: 0.2063 - val_accuracy: 0.7072\n","Epoch 113/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1944 - accuracy: 0.7153 - val_loss: 0.2063 - val_accuracy: 0.7073\n","Epoch 114/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1942 - accuracy: 0.7153 - val_loss: 0.2063 - val_accuracy: 0.7072\n","Epoch 115/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1942 - accuracy: 0.7153 - val_loss: 0.2058 - val_accuracy: 0.7072\n","Epoch 116/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1937 - accuracy: 0.7152 - val_loss: 0.2056 - val_accuracy: 0.7073\n","Epoch 117/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1935 - accuracy: 0.7153 - val_loss: 0.2057 - val_accuracy: 0.7073\n","Epoch 118/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1932 - accuracy: 0.7153 - val_loss: 0.2060 - val_accuracy: 0.7076\n","Epoch 119/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1931 - accuracy: 0.7153 - val_loss: 0.2051 - val_accuracy: 0.7073\n","Epoch 120/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.1926 - accuracy: 0.7153 - val_loss: 0.2045 - val_accuracy: 0.7073\n","Epoch 121/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1928 - accuracy: 0.7153 - val_loss: 0.2049 - val_accuracy: 0.7072\n","Epoch 122/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1926 - accuracy: 0.7153 - val_loss: 0.2051 - val_accuracy: 0.7071\n","Epoch 123/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1926 - accuracy: 0.7153 - val_loss: 0.2054 - val_accuracy: 0.7075\n","Epoch 124/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1919 - accuracy: 0.7153 - val_loss: 0.2038 - val_accuracy: 0.7073\n","Epoch 125/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1913 - accuracy: 0.7153 - val_loss: 0.2040 - val_accuracy: 0.7073\n","Epoch 126/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1910 - accuracy: 0.7153 - val_loss: 0.2036 - val_accuracy: 0.7073\n","Epoch 127/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1907 - accuracy: 0.7153 - val_loss: 0.2034 - val_accuracy: 0.7073\n","Epoch 128/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1908 - accuracy: 0.7154 - val_loss: 0.2034 - val_accuracy: 0.7073\n","Epoch 129/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1914 - accuracy: 0.7153 - val_loss: 0.2037 - val_accuracy: 0.7071\n","Epoch 130/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1907 - accuracy: 0.7153 - val_loss: 0.2030 - val_accuracy: 0.7074\n","Epoch 131/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1907 - accuracy: 0.7153 - val_loss: 0.2029 - val_accuracy: 0.7073\n","Epoch 132/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1896 - accuracy: 0.7154 - val_loss: 0.2024 - val_accuracy: 0.7073\n","Epoch 133/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1895 - accuracy: 0.7153 - val_loss: 0.2021 - val_accuracy: 0.7073\n","Epoch 134/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1894 - accuracy: 0.7154 - val_loss: 0.2021 - val_accuracy: 0.7072\n","Epoch 135/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1890 - accuracy: 0.7154 - val_loss: 0.2031 - val_accuracy: 0.7070\n","Epoch 136/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1893 - accuracy: 0.7153 - val_loss: 0.2042 - val_accuracy: 0.7075\n","Epoch 137/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1891 - accuracy: 0.7154 - val_loss: 0.2019 - val_accuracy: 0.7074\n","Epoch 138/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1890 - accuracy: 0.7154 - val_loss: 0.2023 - val_accuracy: 0.7071\n","Epoch 139/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1887 - accuracy: 0.7154 - val_loss: 0.2012 - val_accuracy: 0.7073\n","Epoch 140/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1880 - accuracy: 0.7153 - val_loss: 0.2010 - val_accuracy: 0.7074\n","Epoch 141/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1876 - accuracy: 0.7154 - val_loss: 0.2007 - val_accuracy: 0.7073\n","Epoch 142/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1877 - accuracy: 0.7154 - val_loss: 0.2012 - val_accuracy: 0.7074\n","Epoch 143/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1881 - accuracy: 0.7154 - val_loss: 0.2015 - val_accuracy: 0.7072\n","Epoch 144/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1876 - accuracy: 0.7154 - val_loss: 0.2007 - val_accuracy: 0.7071\n","Epoch 145/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1871 - accuracy: 0.7154 - val_loss: 0.2003 - val_accuracy: 0.7072\n","Epoch 146/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1868 - accuracy: 0.7154 - val_loss: 0.1999 - val_accuracy: 0.7074\n","Epoch 147/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1865 - accuracy: 0.7154 - val_loss: 0.1999 - val_accuracy: 0.7074\n","Epoch 148/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1862 - accuracy: 0.7154 - val_loss: 0.1992 - val_accuracy: 0.7073\n","Epoch 149/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1860 - accuracy: 0.7154 - val_loss: 0.1996 - val_accuracy: 0.7074\n","Epoch 150/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1860 - accuracy: 0.7154 - val_loss: 0.1990 - val_accuracy: 0.7074\n","Epoch 151/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1858 - accuracy: 0.7154 - val_loss: 0.1990 - val_accuracy: 0.7073\n","Epoch 152/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1853 - accuracy: 0.7154 - val_loss: 0.1990 - val_accuracy: 0.7074\n","Epoch 153/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1851 - accuracy: 0.7154 - val_loss: 0.1989 - val_accuracy: 0.7072\n","Epoch 154/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1849 - accuracy: 0.7154 - val_loss: 0.1982 - val_accuracy: 0.7074\n","Epoch 155/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1845 - accuracy: 0.7154 - val_loss: 0.1982 - val_accuracy: 0.7074\n","Epoch 156/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1850 - accuracy: 0.7154 - val_loss: 0.1987 - val_accuracy: 0.7072\n","Epoch 157/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1848 - accuracy: 0.7154 - val_loss: 0.1990 - val_accuracy: 0.7075\n","Epoch 158/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1844 - accuracy: 0.7155 - val_loss: 0.1980 - val_accuracy: 0.7075\n","Epoch 159/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1839 - accuracy: 0.7154 - val_loss: 0.1976 - val_accuracy: 0.7074\n","Epoch 160/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1835 - accuracy: 0.7155 - val_loss: 0.1973 - val_accuracy: 0.7073\n","Epoch 161/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1835 - accuracy: 0.7155 - val_loss: 0.1969 - val_accuracy: 0.7074\n","Epoch 162/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1829 - accuracy: 0.7155 - val_loss: 0.1966 - val_accuracy: 0.7074\n","Epoch 163/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1829 - accuracy: 0.7155 - val_loss: 0.1970 - val_accuracy: 0.7072\n","Epoch 164/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1829 - accuracy: 0.7154 - val_loss: 0.1977 - val_accuracy: 0.7076\n","Epoch 165/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1828 - accuracy: 0.7155 - val_loss: 0.1969 - val_accuracy: 0.7073\n","Epoch 166/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1825 - accuracy: 0.7154 - val_loss: 0.1966 - val_accuracy: 0.7074\n","Epoch 167/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1835 - accuracy: 0.7154 - val_loss: 0.1971 - val_accuracy: 0.7075\n","Epoch 168/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1832 - accuracy: 0.7154 - val_loss: 0.1960 - val_accuracy: 0.7075\n","Epoch 169/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1825 - accuracy: 0.7155 - val_loss: 0.1960 - val_accuracy: 0.7073\n","Epoch 170/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1815 - accuracy: 0.7155 - val_loss: 0.1952 - val_accuracy: 0.7074\n","Epoch 171/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1812 - accuracy: 0.7155 - val_loss: 0.1955 - val_accuracy: 0.7073\n","Epoch 172/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1811 - accuracy: 0.7155 - val_loss: 0.1947 - val_accuracy: 0.7073\n","Epoch 173/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1807 - accuracy: 0.7154 - val_loss: 0.1945 - val_accuracy: 0.7073\n","Epoch 174/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1805 - accuracy: 0.7155 - val_loss: 0.1947 - val_accuracy: 0.7074\n","Epoch 175/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1804 - accuracy: 0.7155 - val_loss: 0.1950 - val_accuracy: 0.7070\n","Epoch 176/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1804 - accuracy: 0.7155 - val_loss: 0.1941 - val_accuracy: 0.7073\n","Epoch 177/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1800 - accuracy: 0.7155 - val_loss: 0.1939 - val_accuracy: 0.7074\n","Epoch 178/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1796 - accuracy: 0.7155 - val_loss: 0.1939 - val_accuracy: 0.7073\n","Epoch 179/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1796 - accuracy: 0.7155 - val_loss: 0.1946 - val_accuracy: 0.7076\n","Epoch 180/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1800 - accuracy: 0.7155 - val_loss: 0.1936 - val_accuracy: 0.7072\n","Epoch 181/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1796 - accuracy: 0.7155 - val_loss: 0.1941 - val_accuracy: 0.7075\n","Epoch 182/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1793 - accuracy: 0.7155 - val_loss: 0.1934 - val_accuracy: 0.7071\n","Epoch 183/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1789 - accuracy: 0.7155 - val_loss: 0.1941 - val_accuracy: 0.7073\n","Epoch 184/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1788 - accuracy: 0.7155 - val_loss: 0.1930 - val_accuracy: 0.7072\n","Epoch 185/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1787 - accuracy: 0.7155 - val_loss: 0.1930 - val_accuracy: 0.7073\n","Epoch 186/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1783 - accuracy: 0.7155 - val_loss: 0.1927 - val_accuracy: 0.7072\n","Epoch 187/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1784 - accuracy: 0.7155 - val_loss: 0.1943 - val_accuracy: 0.7075\n","Epoch 188/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1794 - accuracy: 0.7154 - val_loss: 0.1925 - val_accuracy: 0.7071\n","Epoch 189/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1791 - accuracy: 0.7155 - val_loss: 0.1949 - val_accuracy: 0.7065\n","Epoch 190/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1786 - accuracy: 0.7155 - val_loss: 0.1932 - val_accuracy: 0.7069\n","Epoch 191/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1776 - accuracy: 0.7155 - val_loss: 0.1921 - val_accuracy: 0.7072\n","Epoch 192/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1771 - accuracy: 0.7155 - val_loss: 0.1925 - val_accuracy: 0.7069\n","Epoch 193/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1775 - accuracy: 0.7155 - val_loss: 0.1918 - val_accuracy: 0.7073\n","Epoch 194/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1773 - accuracy: 0.7154 - val_loss: 0.1924 - val_accuracy: 0.7074\n","Epoch 195/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1772 - accuracy: 0.7155 - val_loss: 0.1916 - val_accuracy: 0.7071\n","Epoch 196/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1778 - accuracy: 0.7155 - val_loss: 0.1924 - val_accuracy: 0.7068\n","Epoch 197/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.1770 - accuracy: 0.7155 - val_loss: 0.1918 - val_accuracy: 0.7071\n","Epoch 198/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1766 - accuracy: 0.7155 - val_loss: 0.1914 - val_accuracy: 0.7070\n","Epoch 199/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1761 - accuracy: 0.7155 - val_loss: 0.1912 - val_accuracy: 0.7071\n","Epoch 200/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1761 - accuracy: 0.7155 - val_loss: 0.1908 - val_accuracy: 0.7072\n","Epoch 201/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1758 - accuracy: 0.7155 - val_loss: 0.1914 - val_accuracy: 0.7071\n","Epoch 202/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1760 - accuracy: 0.7156 - val_loss: 0.1910 - val_accuracy: 0.7071\n","Epoch 203/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1759 - accuracy: 0.7155 - val_loss: 0.1905 - val_accuracy: 0.7072\n","Epoch 204/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1758 - accuracy: 0.7155 - val_loss: 0.1913 - val_accuracy: 0.7074\n","Epoch 205/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1752 - accuracy: 0.7155 - val_loss: 0.1903 - val_accuracy: 0.7073\n","Epoch 206/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1751 - accuracy: 0.7155 - val_loss: 0.1912 - val_accuracy: 0.7069\n","Epoch 207/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1753 - accuracy: 0.7155 - val_loss: 0.1902 - val_accuracy: 0.7072\n","Epoch 208/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1746 - accuracy: 0.7155 - val_loss: 0.1901 - val_accuracy: 0.7072\n","Epoch 209/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1745 - accuracy: 0.7155 - val_loss: 0.1898 - val_accuracy: 0.7073\n","Epoch 210/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1744 - accuracy: 0.7155 - val_loss: 0.1900 - val_accuracy: 0.7073\n","Epoch 211/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1745 - accuracy: 0.7155 - val_loss: 0.1900 - val_accuracy: 0.7071\n","Epoch 212/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1741 - accuracy: 0.7155 - val_loss: 0.1898 - val_accuracy: 0.7074\n","Epoch 213/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1741 - accuracy: 0.7155 - val_loss: 0.1896 - val_accuracy: 0.7071\n","Epoch 214/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1740 - accuracy: 0.7155 - val_loss: 0.1894 - val_accuracy: 0.7071\n","Epoch 215/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1737 - accuracy: 0.7155 - val_loss: 0.1897 - val_accuracy: 0.7069\n","Epoch 216/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1739 - accuracy: 0.7155 - val_loss: 0.1901 - val_accuracy: 0.7072\n","Epoch 217/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1737 - accuracy: 0.7155 - val_loss: 0.1891 - val_accuracy: 0.7071\n","Epoch 218/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1733 - accuracy: 0.7155 - val_loss: 0.1892 - val_accuracy: 0.7071\n","Epoch 219/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1734 - accuracy: 0.7155 - val_loss: 0.1893 - val_accuracy: 0.7069\n","Epoch 220/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1734 - accuracy: 0.7155 - val_loss: 0.1891 - val_accuracy: 0.7073\n","Epoch 221/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1730 - accuracy: 0.7155 - val_loss: 0.1888 - val_accuracy: 0.7072\n","Epoch 222/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1729 - accuracy: 0.7155 - val_loss: 0.1888 - val_accuracy: 0.7071\n","Epoch 223/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1726 - accuracy: 0.7155 - val_loss: 0.1885 - val_accuracy: 0.7072\n","Epoch 224/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1725 - accuracy: 0.7155 - val_loss: 0.1887 - val_accuracy: 0.7073\n","Epoch 225/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1737 - accuracy: 0.7154 - val_loss: 0.1892 - val_accuracy: 0.7072\n","Epoch 226/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1724 - accuracy: 0.7155 - val_loss: 0.1883 - val_accuracy: 0.7071\n","Epoch 227/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1721 - accuracy: 0.7155 - val_loss: 0.1883 - val_accuracy: 0.7071\n","Epoch 228/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1727 - accuracy: 0.7155 - val_loss: 0.1885 - val_accuracy: 0.7070\n","Epoch 229/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.1731 - accuracy: 0.7155 - val_loss: 0.1884 - val_accuracy: 0.7069\n","Epoch 230/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1721 - accuracy: 0.7155 - val_loss: 0.1879 - val_accuracy: 0.7071\n","Epoch 231/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1716 - accuracy: 0.7155 - val_loss: 0.1881 - val_accuracy: 0.7072\n","Epoch 232/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1719 - accuracy: 0.7155 - val_loss: 0.1878 - val_accuracy: 0.7071\n","Epoch 233/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1716 - accuracy: 0.7155 - val_loss: 0.1878 - val_accuracy: 0.7070\n","Epoch 234/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1720 - accuracy: 0.7155 - val_loss: 0.1887 - val_accuracy: 0.7073\n","Epoch 235/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1715 - accuracy: 0.7155 - val_loss: 0.1875 - val_accuracy: 0.7071\n","Epoch 236/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1711 - accuracy: 0.7156 - val_loss: 0.1875 - val_accuracy: 0.7071\n","Epoch 237/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1709 - accuracy: 0.7155 - val_loss: 0.1874 - val_accuracy: 0.7071\n","Epoch 238/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1710 - accuracy: 0.7155 - val_loss: 0.1874 - val_accuracy: 0.7072\n","Epoch 239/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1707 - accuracy: 0.7156 - val_loss: 0.1872 - val_accuracy: 0.7070\n","Epoch 240/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1707 - accuracy: 0.7155 - val_loss: 0.1873 - val_accuracy: 0.7072\n","Epoch 241/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1707 - accuracy: 0.7156 - val_loss: 0.1872 - val_accuracy: 0.7070\n","Epoch 242/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1706 - accuracy: 0.7155 - val_loss: 0.1876 - val_accuracy: 0.7067\n","Epoch 243/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.1717 - accuracy: 0.7155 - val_loss: 0.1869 - val_accuracy: 0.7071\n","Epoch 244/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1709 - accuracy: 0.7155 - val_loss: 0.1903 - val_accuracy: 0.7059\n","Epoch 245/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1722 - accuracy: 0.7154 - val_loss: 0.1873 - val_accuracy: 0.7072\n","Epoch 246/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1708 - accuracy: 0.7155 - val_loss: 0.1883 - val_accuracy: 0.7073\n","Epoch 247/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1709 - accuracy: 0.7155 - val_loss: 0.1892 - val_accuracy: 0.7074\n","Epoch 248/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1703 - accuracy: 0.7156 - val_loss: 0.1875 - val_accuracy: 0.7073\n","Epoch 249/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1700 - accuracy: 0.7155 - val_loss: 0.1866 - val_accuracy: 0.7069\n","Epoch 250/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.1697 - accuracy: 0.7156 - val_loss: 0.1865 - val_accuracy: 0.7070\n","Epoch 251/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1694 - accuracy: 0.7155 - val_loss: 0.1864 - val_accuracy: 0.7070\n","Epoch 252/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1691 - accuracy: 0.7156 - val_loss: 0.1863 - val_accuracy: 0.7069\n","Epoch 253/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1689 - accuracy: 0.7155 - val_loss: 0.1862 - val_accuracy: 0.7071\n","Epoch 254/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1689 - accuracy: 0.7156 - val_loss: 0.1860 - val_accuracy: 0.7068\n","Epoch 255/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1690 - accuracy: 0.7155 - val_loss: 0.1863 - val_accuracy: 0.7067\n","Epoch 256/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1694 - accuracy: 0.7155 - val_loss: 0.1873 - val_accuracy: 0.7072\n","Epoch 257/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1692 - accuracy: 0.7156 - val_loss: 0.1855 - val_accuracy: 0.7070\n","Epoch 258/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1685 - accuracy: 0.7155 - val_loss: 0.1856 - val_accuracy: 0.7070\n","Epoch 259/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1682 - accuracy: 0.7156 - val_loss: 0.1854 - val_accuracy: 0.7070\n","Epoch 260/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1680 - accuracy: 0.7156 - val_loss: 0.1854 - val_accuracy: 0.7068\n","Epoch 261/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1679 - accuracy: 0.7156 - val_loss: 0.1854 - val_accuracy: 0.7069\n","Epoch 262/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1678 - accuracy: 0.7156 - val_loss: 0.1855 - val_accuracy: 0.7070\n","Epoch 263/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1676 - accuracy: 0.7156 - val_loss: 0.1855 - val_accuracy: 0.7071\n","Epoch 264/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1676 - accuracy: 0.7156 - val_loss: 0.1855 - val_accuracy: 0.7072\n","Epoch 265/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1676 - accuracy: 0.7156 - val_loss: 0.1858 - val_accuracy: 0.7066\n","Epoch 266/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1674 - accuracy: 0.7156 - val_loss: 0.1848 - val_accuracy: 0.7070\n","Epoch 267/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1673 - accuracy: 0.7156 - val_loss: 0.1850 - val_accuracy: 0.7068\n","Epoch 268/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1678 - accuracy: 0.7155 - val_loss: 0.1850 - val_accuracy: 0.7069\n","Epoch 269/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.1672 - accuracy: 0.7156 - val_loss: 0.1848 - val_accuracy: 0.7069\n","Epoch 270/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1668 - accuracy: 0.7156 - val_loss: 0.1845 - val_accuracy: 0.7072\n","Epoch 271/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1670 - accuracy: 0.7155 - val_loss: 0.1844 - val_accuracy: 0.7071\n","Epoch 272/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1669 - accuracy: 0.7155 - val_loss: 0.1852 - val_accuracy: 0.7068\n","Epoch 273/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1667 - accuracy: 0.7156 - val_loss: 0.1843 - val_accuracy: 0.7070\n","Epoch 274/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1666 - accuracy: 0.7156 - val_loss: 0.1846 - val_accuracy: 0.7069\n","Epoch 275/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1665 - accuracy: 0.7156 - val_loss: 0.1849 - val_accuracy: 0.7073\n","Epoch 276/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1676 - accuracy: 0.7155 - val_loss: 0.1843 - val_accuracy: 0.7068\n","Epoch 277/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.1668 - accuracy: 0.7156 - val_loss: 0.1842 - val_accuracy: 0.7068\n","Epoch 278/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1663 - accuracy: 0.7156 - val_loss: 0.1845 - val_accuracy: 0.7073\n","Epoch 279/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1665 - accuracy: 0.7156 - val_loss: 0.1843 - val_accuracy: 0.7069\n","Epoch 280/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1662 - accuracy: 0.7156 - val_loss: 0.1838 - val_accuracy: 0.7071\n","Epoch 281/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1657 - accuracy: 0.7156 - val_loss: 0.1839 - val_accuracy: 0.7069\n","Epoch 282/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1657 - accuracy: 0.7156 - val_loss: 0.1841 - val_accuracy: 0.7071\n","Epoch 283/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1655 - accuracy: 0.7156 - val_loss: 0.1836 - val_accuracy: 0.7071\n","Epoch 284/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1654 - accuracy: 0.7156 - val_loss: 0.1836 - val_accuracy: 0.7068\n","Epoch 285/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1653 - accuracy: 0.7156 - val_loss: 0.1838 - val_accuracy: 0.7069\n","Epoch 286/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1659 - accuracy: 0.7156 - val_loss: 0.1839 - val_accuracy: 0.7068\n","Epoch 287/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1652 - accuracy: 0.7156 - val_loss: 0.1837 - val_accuracy: 0.7068\n","Epoch 288/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1655 - accuracy: 0.7156 - val_loss: 0.1846 - val_accuracy: 0.7064\n","Epoch 289/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1661 - accuracy: 0.7156 - val_loss: 0.1850 - val_accuracy: 0.7072\n","Epoch 290/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1660 - accuracy: 0.7156 - val_loss: 0.1832 - val_accuracy: 0.7069\n","Epoch 291/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1652 - accuracy: 0.7156 - val_loss: 0.1835 - val_accuracy: 0.7071\n","Epoch 292/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1649 - accuracy: 0.7156 - val_loss: 0.1832 - val_accuracy: 0.7068\n","Epoch 293/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1649 - accuracy: 0.7156 - val_loss: 0.1834 - val_accuracy: 0.7072\n","Epoch 294/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1647 - accuracy: 0.7156 - val_loss: 0.1829 - val_accuracy: 0.7069\n","Epoch 295/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1643 - accuracy: 0.7156 - val_loss: 0.1829 - val_accuracy: 0.7070\n","Epoch 296/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1646 - accuracy: 0.7156 - val_loss: 0.1834 - val_accuracy: 0.7072\n","Epoch 297/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1647 - accuracy: 0.7156 - val_loss: 0.1828 - val_accuracy: 0.7069\n","Epoch 298/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1651 - accuracy: 0.7156 - val_loss: 0.1829 - val_accuracy: 0.7071\n","Epoch 299/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1657 - accuracy: 0.7155 - val_loss: 0.1833 - val_accuracy: 0.7072\n","Epoch 300/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1644 - accuracy: 0.7157 - val_loss: 0.1824 - val_accuracy: 0.7071\n","Epoch 301/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1640 - accuracy: 0.7157 - val_loss: 0.1826 - val_accuracy: 0.7068\n","Epoch 302/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1639 - accuracy: 0.7156 - val_loss: 0.1823 - val_accuracy: 0.7070\n","Epoch 303/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.1637 - accuracy: 0.7157 - val_loss: 0.1823 - val_accuracy: 0.7070\n","Epoch 304/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1636 - accuracy: 0.7157 - val_loss: 0.1825 - val_accuracy: 0.7068\n","Epoch 305/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1637 - accuracy: 0.7156 - val_loss: 0.1821 - val_accuracy: 0.7071\n","Epoch 306/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1635 - accuracy: 0.7156 - val_loss: 0.1823 - val_accuracy: 0.7068\n","Epoch 307/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1638 - accuracy: 0.7156 - val_loss: 0.1822 - val_accuracy: 0.7070\n","Epoch 308/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1643 - accuracy: 0.7156 - val_loss: 0.1824 - val_accuracy: 0.7069\n","Epoch 309/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1644 - accuracy: 0.7156 - val_loss: 0.1830 - val_accuracy: 0.7073\n","Epoch 310/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1638 - accuracy: 0.7156 - val_loss: 0.1821 - val_accuracy: 0.7069\n","Epoch 311/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1634 - accuracy: 0.7157 - val_loss: 0.1820 - val_accuracy: 0.7070\n","Epoch 312/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1633 - accuracy: 0.7157 - val_loss: 0.1821 - val_accuracy: 0.7068\n","Epoch 313/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1633 - accuracy: 0.7156 - val_loss: 0.1824 - val_accuracy: 0.7071\n","Epoch 314/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1632 - accuracy: 0.7157 - val_loss: 0.1818 - val_accuracy: 0.7069\n","Epoch 315/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1636 - accuracy: 0.7156 - val_loss: 0.1817 - val_accuracy: 0.7069\n","Epoch 316/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1629 - accuracy: 0.7157 - val_loss: 0.1818 - val_accuracy: 0.7070\n","Epoch 317/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1628 - accuracy: 0.7157 - val_loss: 0.1820 - val_accuracy: 0.7072\n","Epoch 318/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1631 - accuracy: 0.7156 - val_loss: 0.1822 - val_accuracy: 0.7067\n","Epoch 319/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1628 - accuracy: 0.7156 - val_loss: 0.1822 - val_accuracy: 0.7072\n","Epoch 320/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1627 - accuracy: 0.7157 - val_loss: 0.1815 - val_accuracy: 0.7069\n","Epoch 321/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1626 - accuracy: 0.7156 - val_loss: 0.1815 - val_accuracy: 0.7071\n","Epoch 322/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1625 - accuracy: 0.7157 - val_loss: 0.1817 - val_accuracy: 0.7068\n","Epoch 323/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1624 - accuracy: 0.7157 - val_loss: 0.1818 - val_accuracy: 0.7071\n","Epoch 324/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1624 - accuracy: 0.7156 - val_loss: 0.1814 - val_accuracy: 0.7069\n","Epoch 325/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1623 - accuracy: 0.7157 - val_loss: 0.1816 - val_accuracy: 0.7071\n","Epoch 326/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1624 - accuracy: 0.7157 - val_loss: 0.1821 - val_accuracy: 0.7072\n","Epoch 327/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1626 - accuracy: 0.7156 - val_loss: 0.1824 - val_accuracy: 0.7065\n","Epoch 328/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1627 - accuracy: 0.7156 - val_loss: 0.1826 - val_accuracy: 0.7073\n","Epoch 329/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1625 - accuracy: 0.7156 - val_loss: 0.1814 - val_accuracy: 0.7067\n","Epoch 330/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1621 - accuracy: 0.7157 - val_loss: 0.1815 - val_accuracy: 0.7071\n","Epoch 331/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1622 - accuracy: 0.7156 - val_loss: 0.1810 - val_accuracy: 0.7071\n","Epoch 332/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1624 - accuracy: 0.7156 - val_loss: 0.1811 - val_accuracy: 0.7068\n","Epoch 333/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1620 - accuracy: 0.7157 - val_loss: 0.1810 - val_accuracy: 0.7069\n","Epoch 334/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1619 - accuracy: 0.7157 - val_loss: 0.1809 - val_accuracy: 0.7071\n","Epoch 335/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1624 - accuracy: 0.7156 - val_loss: 0.1826 - val_accuracy: 0.7062\n","Epoch 336/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1629 - accuracy: 0.7156 - val_loss: 0.1814 - val_accuracy: 0.7072\n","Epoch 337/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1622 - accuracy: 0.7157 - val_loss: 0.1814 - val_accuracy: 0.7065\n","Epoch 338/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1617 - accuracy: 0.7156 - val_loss: 0.1809 - val_accuracy: 0.7071\n","Epoch 339/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1615 - accuracy: 0.7157 - val_loss: 0.1809 - val_accuracy: 0.7069\n","Epoch 340/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1614 - accuracy: 0.7156 - val_loss: 0.1808 - val_accuracy: 0.7068\n","Epoch 341/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1614 - accuracy: 0.7157 - val_loss: 0.1816 - val_accuracy: 0.7072\n","Epoch 342/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1615 - accuracy: 0.7157 - val_loss: 0.1806 - val_accuracy: 0.7069\n","Epoch 343/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1615 - accuracy: 0.7157 - val_loss: 0.1807 - val_accuracy: 0.7067\n","Epoch 344/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1611 - accuracy: 0.7157 - val_loss: 0.1804 - val_accuracy: 0.7069\n","Epoch 345/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1612 - accuracy: 0.7156 - val_loss: 0.1806 - val_accuracy: 0.7070\n","Epoch 346/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1615 - accuracy: 0.7157 - val_loss: 0.1808 - val_accuracy: 0.7067\n","Epoch 347/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1618 - accuracy: 0.7156 - val_loss: 0.1809 - val_accuracy: 0.7067\n","Epoch 348/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1618 - accuracy: 0.7157 - val_loss: 0.1817 - val_accuracy: 0.7072\n","Epoch 349/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1622 - accuracy: 0.7156 - val_loss: 0.1812 - val_accuracy: 0.7071\n","Epoch 350/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1620 - accuracy: 0.7156 - val_loss: 0.1807 - val_accuracy: 0.7065\n","Epoch 351/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1613 - accuracy: 0.7156 - val_loss: 0.1803 - val_accuracy: 0.7069\n","Epoch 352/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1609 - accuracy: 0.7157 - val_loss: 0.1807 - val_accuracy: 0.7071\n","Epoch 353/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1608 - accuracy: 0.7157 - val_loss: 0.1803 - val_accuracy: 0.7069\n","Epoch 354/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1607 - accuracy: 0.7157 - val_loss: 0.1805 - val_accuracy: 0.7070\n","Epoch 355/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1610 - accuracy: 0.7156 - val_loss: 0.1808 - val_accuracy: 0.7072\n","Epoch 356/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1614 - accuracy: 0.7157 - val_loss: 0.1802 - val_accuracy: 0.7068\n","Epoch 357/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1606 - accuracy: 0.7157 - val_loss: 0.1807 - val_accuracy: 0.7066\n","Epoch 358/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1608 - accuracy: 0.7157 - val_loss: 0.1799 - val_accuracy: 0.7071\n","Epoch 359/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1609 - accuracy: 0.7156 - val_loss: 0.1811 - val_accuracy: 0.7065\n","Epoch 360/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1611 - accuracy: 0.7156 - val_loss: 0.1803 - val_accuracy: 0.7070\n","Epoch 361/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1603 - accuracy: 0.7156 - val_loss: 0.1800 - val_accuracy: 0.7070\n","Epoch 362/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1603 - accuracy: 0.7157 - val_loss: 0.1803 - val_accuracy: 0.7068\n","Epoch 363/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1603 - accuracy: 0.7157 - val_loss: 0.1798 - val_accuracy: 0.7069\n","Epoch 364/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1600 - accuracy: 0.7157 - val_loss: 0.1801 - val_accuracy: 0.7068\n","Epoch 365/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1601 - accuracy: 0.7156 - val_loss: 0.1798 - val_accuracy: 0.7070\n","Epoch 366/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1600 - accuracy: 0.7157 - val_loss: 0.1800 - val_accuracy: 0.7068\n","Epoch 367/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1604 - accuracy: 0.7156 - val_loss: 0.1803 - val_accuracy: 0.7071\n","Epoch 368/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1600 - accuracy: 0.7157 - val_loss: 0.1802 - val_accuracy: 0.7071\n","Epoch 369/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1602 - accuracy: 0.7156 - val_loss: 0.1802 - val_accuracy: 0.7067\n","Epoch 370/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1606 - accuracy: 0.7157 - val_loss: 0.1800 - val_accuracy: 0.7068\n","Epoch 371/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1602 - accuracy: 0.7157 - val_loss: 0.1799 - val_accuracy: 0.7070\n","Epoch 372/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1602 - accuracy: 0.7156 - val_loss: 0.1798 - val_accuracy: 0.7069\n","Epoch 373/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1602 - accuracy: 0.7157 - val_loss: 0.1800 - val_accuracy: 0.7070\n","Epoch 374/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1599 - accuracy: 0.7156 - val_loss: 0.1795 - val_accuracy: 0.7068\n","Epoch 375/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1596 - accuracy: 0.7157 - val_loss: 0.1801 - val_accuracy: 0.7070\n","Epoch 376/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1602 - accuracy: 0.7157 - val_loss: 0.1794 - val_accuracy: 0.7069\n","Epoch 377/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1599 - accuracy: 0.7157 - val_loss: 0.1796 - val_accuracy: 0.7068\n","Epoch 378/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1597 - accuracy: 0.7157 - val_loss: 0.1802 - val_accuracy: 0.7070\n","Epoch 379/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1597 - accuracy: 0.7157 - val_loss: 0.1795 - val_accuracy: 0.7070\n","Epoch 380/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1595 - accuracy: 0.7157 - val_loss: 0.1792 - val_accuracy: 0.7069\n","Epoch 381/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1594 - accuracy: 0.7157 - val_loss: 0.1794 - val_accuracy: 0.7069\n","Epoch 382/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1593 - accuracy: 0.7157 - val_loss: 0.1793 - val_accuracy: 0.7069\n","Epoch 383/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1598 - accuracy: 0.7156 - val_loss: 0.1794 - val_accuracy: 0.7069\n","Epoch 384/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1599 - accuracy: 0.7157 - val_loss: 0.1802 - val_accuracy: 0.7066\n","Epoch 385/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1596 - accuracy: 0.7156 - val_loss: 0.1796 - val_accuracy: 0.7069\n","Epoch 386/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1595 - accuracy: 0.7156 - val_loss: 0.1792 - val_accuracy: 0.7069\n","Epoch 387/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1596 - accuracy: 0.7156 - val_loss: 0.1793 - val_accuracy: 0.7068\n","Epoch 388/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1590 - accuracy: 0.7157 - val_loss: 0.1795 - val_accuracy: 0.7070\n","Epoch 389/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1590 - accuracy: 0.7157 - val_loss: 0.1793 - val_accuracy: 0.7069\n","Epoch 390/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1589 - accuracy: 0.7157 - val_loss: 0.1791 - val_accuracy: 0.7070\n","Epoch 391/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1588 - accuracy: 0.7157 - val_loss: 0.1791 - val_accuracy: 0.7068\n","Epoch 392/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1589 - accuracy: 0.7157 - val_loss: 0.1793 - val_accuracy: 0.7067\n","Epoch 393/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1592 - accuracy: 0.7157 - val_loss: 0.1794 - val_accuracy: 0.7067\n","Epoch 394/500\n","9/9 [==============================] - 0s 22ms/step - loss: 0.1595 - accuracy: 0.7156 - val_loss: 0.1801 - val_accuracy: 0.7071\n","Epoch 395/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1592 - accuracy: 0.7157 - val_loss: 0.1795 - val_accuracy: 0.7066\n","Epoch 396/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1590 - accuracy: 0.7156 - val_loss: 0.1791 - val_accuracy: 0.7067\n","Epoch 397/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1589 - accuracy: 0.7157 - val_loss: 0.1788 - val_accuracy: 0.7069\n","Epoch 398/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1589 - accuracy: 0.7157 - val_loss: 0.1798 - val_accuracy: 0.7065\n","Epoch 399/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1592 - accuracy: 0.7156 - val_loss: 0.1794 - val_accuracy: 0.7069\n","Epoch 400/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1589 - accuracy: 0.7156 - val_loss: 0.1788 - val_accuracy: 0.7069\n","Epoch 401/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.1587 - accuracy: 0.7157 - val_loss: 0.1790 - val_accuracy: 0.7067\n","Epoch 402/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1585 - accuracy: 0.7157 - val_loss: 0.1788 - val_accuracy: 0.7069\n","Epoch 403/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1585 - accuracy: 0.7157 - val_loss: 0.1789 - val_accuracy: 0.7068\n","Epoch 404/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1584 - accuracy: 0.7157 - val_loss: 0.1787 - val_accuracy: 0.7069\n","Epoch 405/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1583 - accuracy: 0.7157 - val_loss: 0.1789 - val_accuracy: 0.7068\n","Epoch 406/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1584 - accuracy: 0.7157 - val_loss: 0.1788 - val_accuracy: 0.7068\n","Epoch 407/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.1589 - accuracy: 0.7156 - val_loss: 0.1812 - val_accuracy: 0.7071\n","Epoch 408/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1604 - accuracy: 0.7156 - val_loss: 0.1798 - val_accuracy: 0.7066\n","Epoch 409/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1596 - accuracy: 0.7156 - val_loss: 0.1798 - val_accuracy: 0.7071\n","Epoch 410/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1590 - accuracy: 0.7157 - val_loss: 0.1788 - val_accuracy: 0.7067\n","Epoch 411/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1586 - accuracy: 0.7157 - val_loss: 0.1787 - val_accuracy: 0.7069\n","Epoch 412/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1590 - accuracy: 0.7157 - val_loss: 0.1789 - val_accuracy: 0.7068\n","Epoch 413/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1584 - accuracy: 0.7157 - val_loss: 0.1789 - val_accuracy: 0.7066\n","Epoch 414/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1584 - accuracy: 0.7157 - val_loss: 0.1786 - val_accuracy: 0.7069\n","Epoch 415/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1583 - accuracy: 0.7156 - val_loss: 0.1792 - val_accuracy: 0.7069\n","Epoch 416/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1581 - accuracy: 0.7157 - val_loss: 0.1784 - val_accuracy: 0.7068\n","Epoch 417/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1579 - accuracy: 0.7157 - val_loss: 0.1792 - val_accuracy: 0.7065\n","Epoch 418/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1585 - accuracy: 0.7157 - val_loss: 0.1807 - val_accuracy: 0.7072\n","Epoch 419/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1594 - accuracy: 0.7156 - val_loss: 0.1808 - val_accuracy: 0.7062\n","Epoch 420/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1590 - accuracy: 0.7156 - val_loss: 0.1787 - val_accuracy: 0.7068\n","Epoch 421/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1584 - accuracy: 0.7157 - val_loss: 0.1788 - val_accuracy: 0.7068\n","Epoch 422/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1580 - accuracy: 0.7157 - val_loss: 0.1783 - val_accuracy: 0.7067\n","Epoch 423/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1580 - accuracy: 0.7156 - val_loss: 0.1784 - val_accuracy: 0.7068\n","Epoch 424/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1579 - accuracy: 0.7157 - val_loss: 0.1787 - val_accuracy: 0.7071\n","Epoch 425/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1580 - accuracy: 0.7157 - val_loss: 0.1794 - val_accuracy: 0.7064\n","Epoch 426/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1580 - accuracy: 0.7157 - val_loss: 0.1785 - val_accuracy: 0.7070\n","Epoch 427/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1579 - accuracy: 0.7157 - val_loss: 0.1782 - val_accuracy: 0.7068\n","Epoch 428/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.1578 - accuracy: 0.7157 - val_loss: 0.1781 - val_accuracy: 0.7068\n","Epoch 429/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1575 - accuracy: 0.7157 - val_loss: 0.1782 - val_accuracy: 0.7069\n","Epoch 430/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1575 - accuracy: 0.7157 - val_loss: 0.1782 - val_accuracy: 0.7067\n","Epoch 431/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1575 - accuracy: 0.7157 - val_loss: 0.1781 - val_accuracy: 0.7068\n","Epoch 432/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1575 - accuracy: 0.7157 - val_loss: 0.1779 - val_accuracy: 0.7069\n","Epoch 433/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1574 - accuracy: 0.7157 - val_loss: 0.1781 - val_accuracy: 0.7067\n","Epoch 434/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1581 - accuracy: 0.7156 - val_loss: 0.1805 - val_accuracy: 0.7072\n","Epoch 435/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1595 - accuracy: 0.7156 - val_loss: 0.1792 - val_accuracy: 0.7064\n","Epoch 436/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1577 - accuracy: 0.7157 - val_loss: 0.1780 - val_accuracy: 0.7068\n","Epoch 437/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1574 - accuracy: 0.7157 - val_loss: 0.1781 - val_accuracy: 0.7065\n","Epoch 438/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1574 - accuracy: 0.7157 - val_loss: 0.1778 - val_accuracy: 0.7068\n","Epoch 439/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1571 - accuracy: 0.7157 - val_loss: 0.1778 - val_accuracy: 0.7069\n","Epoch 440/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1571 - accuracy: 0.7157 - val_loss: 0.1778 - val_accuracy: 0.7068\n","Epoch 441/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.1571 - accuracy: 0.7157 - val_loss: 0.1779 - val_accuracy: 0.7070\n","Epoch 442/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1571 - accuracy: 0.7157 - val_loss: 0.1777 - val_accuracy: 0.7068\n","Epoch 443/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1570 - accuracy: 0.7157 - val_loss: 0.1779 - val_accuracy: 0.7067\n","Epoch 444/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1571 - accuracy: 0.7157 - val_loss: 0.1778 - val_accuracy: 0.7069\n","Epoch 445/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1572 - accuracy: 0.7157 - val_loss: 0.1782 - val_accuracy: 0.7069\n","Epoch 446/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1573 - accuracy: 0.7157 - val_loss: 0.1781 - val_accuracy: 0.7066\n","Epoch 447/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1572 - accuracy: 0.7157 - val_loss: 0.1776 - val_accuracy: 0.7068\n","Epoch 448/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1572 - accuracy: 0.7157 - val_loss: 0.1778 - val_accuracy: 0.7068\n","Epoch 449/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1569 - accuracy: 0.7157 - val_loss: 0.1777 - val_accuracy: 0.7068\n","Epoch 450/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1570 - accuracy: 0.7157 - val_loss: 0.1778 - val_accuracy: 0.7069\n","Epoch 451/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1570 - accuracy: 0.7157 - val_loss: 0.1780 - val_accuracy: 0.7067\n","Epoch 452/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1570 - accuracy: 0.7156 - val_loss: 0.1777 - val_accuracy: 0.7070\n","Epoch 453/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1573 - accuracy: 0.7157 - val_loss: 0.1775 - val_accuracy: 0.7069\n","Epoch 454/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1568 - accuracy: 0.7157 - val_loss: 0.1775 - val_accuracy: 0.7067\n","Epoch 455/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1567 - accuracy: 0.7157 - val_loss: 0.1779 - val_accuracy: 0.7067\n","Epoch 456/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1567 - accuracy: 0.7157 - val_loss: 0.1777 - val_accuracy: 0.7067\n","Epoch 457/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1566 - accuracy: 0.7157 - val_loss: 0.1777 - val_accuracy: 0.7068\n","Epoch 458/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1569 - accuracy: 0.7157 - val_loss: 0.1777 - val_accuracy: 0.7070\n","Epoch 459/500\n","9/9 [==============================] - 0s 10ms/step - loss: 0.1571 - accuracy: 0.7157 - val_loss: 0.1775 - val_accuracy: 0.7069\n","Epoch 460/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1567 - accuracy: 0.7157 - val_loss: 0.1778 - val_accuracy: 0.7066\n","Epoch 461/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1567 - accuracy: 0.7157 - val_loss: 0.1786 - val_accuracy: 0.7070\n","Epoch 462/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1576 - accuracy: 0.7156 - val_loss: 0.1783 - val_accuracy: 0.7065\n","Epoch 463/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1565 - accuracy: 0.7157 - val_loss: 0.1775 - val_accuracy: 0.7069\n","Epoch 464/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1564 - accuracy: 0.7157 - val_loss: 0.1775 - val_accuracy: 0.7068\n","Epoch 465/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1565 - accuracy: 0.7157 - val_loss: 0.1774 - val_accuracy: 0.7068\n","Epoch 466/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1568 - accuracy: 0.7157 - val_loss: 0.1776 - val_accuracy: 0.7069\n","Epoch 467/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1568 - accuracy: 0.7157 - val_loss: 0.1779 - val_accuracy: 0.7065\n","Epoch 468/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1568 - accuracy: 0.7157 - val_loss: 0.1781 - val_accuracy: 0.7069\n","Epoch 469/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1563 - accuracy: 0.7157 - val_loss: 0.1773 - val_accuracy: 0.7069\n","Epoch 470/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1564 - accuracy: 0.7157 - val_loss: 0.1776 - val_accuracy: 0.7066\n","Epoch 471/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1564 - accuracy: 0.7157 - val_loss: 0.1775 - val_accuracy: 0.7070\n","Epoch 472/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1561 - accuracy: 0.7157 - val_loss: 0.1773 - val_accuracy: 0.7069\n","Epoch 473/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1560 - accuracy: 0.7157 - val_loss: 0.1772 - val_accuracy: 0.7069\n","Epoch 474/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1561 - accuracy: 0.7157 - val_loss: 0.1772 - val_accuracy: 0.7068\n","Epoch 475/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1560 - accuracy: 0.7157 - val_loss: 0.1771 - val_accuracy: 0.7069\n","Epoch 476/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1560 - accuracy: 0.7157 - val_loss: 0.1776 - val_accuracy: 0.7066\n","Epoch 477/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1560 - accuracy: 0.7157 - val_loss: 0.1771 - val_accuracy: 0.7068\n","Epoch 478/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.1558 - accuracy: 0.7157 - val_loss: 0.1772 - val_accuracy: 0.7068\n","Epoch 479/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.1559 - accuracy: 0.7157 - val_loss: 0.1780 - val_accuracy: 0.7071\n","Epoch 480/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1562 - accuracy: 0.7157 - val_loss: 0.1786 - val_accuracy: 0.7064\n","Epoch 481/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1574 - accuracy: 0.7157 - val_loss: 0.1800 - val_accuracy: 0.7071\n","Epoch 482/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1573 - accuracy: 0.7157 - val_loss: 0.1780 - val_accuracy: 0.7064\n","Epoch 483/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.1563 - accuracy: 0.7157 - val_loss: 0.1770 - val_accuracy: 0.7068\n","Epoch 484/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1562 - accuracy: 0.7157 - val_loss: 0.1775 - val_accuracy: 0.7071\n","Epoch 485/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1563 - accuracy: 0.7157 - val_loss: 0.1785 - val_accuracy: 0.7063\n","Epoch 486/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1569 - accuracy: 0.7157 - val_loss: 0.1774 - val_accuracy: 0.7070\n","Epoch 487/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1561 - accuracy: 0.7157 - val_loss: 0.1777 - val_accuracy: 0.7071\n","Epoch 488/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1563 - accuracy: 0.7157 - val_loss: 0.1773 - val_accuracy: 0.7066\n","Epoch 489/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1559 - accuracy: 0.7157 - val_loss: 0.1773 - val_accuracy: 0.7066\n","Epoch 490/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1559 - accuracy: 0.7156 - val_loss: 0.1771 - val_accuracy: 0.7069\n","Epoch 491/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1555 - accuracy: 0.7157 - val_loss: 0.1770 - val_accuracy: 0.7070\n","Epoch 492/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.1555 - accuracy: 0.7157 - val_loss: 0.1769 - val_accuracy: 0.7068\n","Epoch 493/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1558 - accuracy: 0.7157 - val_loss: 0.1784 - val_accuracy: 0.7071\n","Epoch 494/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1560 - accuracy: 0.7157 - val_loss: 0.1772 - val_accuracy: 0.7067\n","Epoch 495/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1557 - accuracy: 0.7157 - val_loss: 0.1769 - val_accuracy: 0.7069\n","Epoch 496/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1555 - accuracy: 0.7157 - val_loss: 0.1770 - val_accuracy: 0.7069\n","Epoch 497/500\n","9/9 [==============================] - 0s 9ms/step - loss: 0.1555 - accuracy: 0.7157 - val_loss: 0.1775 - val_accuracy: 0.7069\n","Epoch 498/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1556 - accuracy: 0.7157 - val_loss: 0.1768 - val_accuracy: 0.7068\n","Epoch 499/500\n","9/9 [==============================] - 0s 8ms/step - loss: 0.1556 - accuracy: 0.7157 - val_loss: 0.1769 - val_accuracy: 0.7068\n","Epoch 500/500\n","9/9 [==============================] - 0s 7ms/step - loss: 0.1557 - accuracy: 0.7157 - val_loss: 0.1767 - val_accuracy: 0.7068\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["<tensorflow.python.keras.callbacks.History at 0x7f6b1decf828>"]},"metadata":{"tags":[]},"execution_count":20}]},{"cell_type":"code","metadata":{"id":"mvlBFUK7FQwT","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":244},"outputId":"75bec96d-059a-41a5-e417-fa41743a77af","executionInfo":{"status":"ok","timestamp":1592049335612,"user_tz":-480,"elapsed":53090,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}}},"source":["# original images vs. images after encoded and decoded\n","decoded_images = autoencoder.predict(train_images)\n","\n","n = 10\n","plt.figure(figsize=(20, 4))\n","for i in range(1, n):\n","    # display original\n","    ax = plt.subplot(2, n, i)\n","    plt.imshow(train_images[i].reshape(28, 28))\n","    plt.gray()\n","    ax.get_xaxis().set_visible(False)\n","    ax.get_yaxis().set_visible(False)\n","\n","    # display reconstruction\n","    ax = plt.subplot(2, n, i + n)\n","    plt.imshow(decoded_images[i].reshape(28, 28))\n","    plt.gray()\n","    ax.get_xaxis().set_visible(False)\n","    ax.get_yaxis().set_visible(False)\n","plt.show()"],"execution_count":21,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 1440x288 with 18 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"code","metadata":{"id":"856x4_D_DocA","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":34},"outputId":"5b29277f-6b57-4c5f-82c0-34486cc56a12","executionInfo":{"status":"ok","timestamp":1592049335614,"user_tz":-480,"elapsed":53079,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}}},"source":["# encoded images (middle layer)\n","encoded_images = encoder.predict(train_images)\n","encoded_images = encoded_images.reshape(356, 16, 16)\n","encoded_images.shape"],"execution_count":22,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(356, 16, 16)"]},"metadata":{"tags":[]},"execution_count":22}]},{"cell_type":"code","metadata":{"id":"8oFfCj38EK-8","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":122},"outputId":"a5b553e3-2c0a-4289-e55e-da54998f8692","executionInfo":{"status":"ok","timestamp":1592049335615,"user_tz":-480,"elapsed":53068,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}}},"source":["# check encoded images\n","n = 10\n","plt.figure(figsize=(30, 20))\n","for i in range(1, n):\n","    ax = plt.subplot(1, n, i)\n","    plt.imshow(encoded_images[i])\n","    plt.gray()\n","    ax.get_xaxis().set_visible(False)\n","    ax.get_yaxis().set_visible(False)\n","plt.show()"],"execution_count":23,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 2160x1440 with 9 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"dEi36UFUHJkA","colab_type":"text"},"source":["### Build CNN model with encoded image as input"]},{"cell_type":"code","metadata":{"id":"NLB0CukEpDoz","colab_type":"code","colab":{}},"source":["model = tf.keras.models.Sequential([\n","    tf.keras.layers.Conv2D(32, (3,3), activation='relu', input_shape=(16, 16, 1), padding=\"same\"),\n","    tf.keras.layers.Conv2D(64, (3,3), activation='relu', padding=\"same\"),\n","    tf.keras.layers.Conv2D(64, (3,3), activation='relu', padding=\"same\"),\n","    tf.keras.layers.MaxPooling2D(2,2),\n","    tf.keras.layers.Flatten(),\n","    tf.keras.layers.Dense(256, activation='relu'),\n","    tf.keras.layers.Dropout(0.4),\n","    tf.keras.layers.Dense(len_classes, activation='softmax')\n","])"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"miGsAuqMELz6","colab_type":"code","outputId":"fb92cb2a-e59f-4929-f635-78a58374471f","executionInfo":{"status":"ok","timestamp":1592049335617,"user_tz":-480,"elapsed":53053,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}},"colab":{"base_uri":"https://localhost:8080/","height":425}},"source":["model.summary()"],"execution_count":25,"outputs":[{"output_type":"stream","text":["Model: \"sequential\"\n","_________________________________________________________________\n","Layer (type)                 Output Shape              Param #   \n","=================================================================\n","conv2d_7 (Conv2D)            (None, 16, 16, 32)        320       \n","_________________________________________________________________\n","conv2d_8 (Conv2D)            (None, 16, 16, 64)        18496     \n","_________________________________________________________________\n","conv2d_9 (Conv2D)            (None, 16, 16, 64)        36928     \n","_________________________________________________________________\n","max_pooling2d_3 (MaxPooling2 (None, 8, 8, 64)          0         \n","_________________________________________________________________\n","flatten (Flatten)            (None, 4096)              0         \n","_________________________________________________________________\n","dense (Dense)                (None, 256)               1048832   \n","_________________________________________________________________\n","dropout (Dropout)            (None, 256)               0         \n","_________________________________________________________________\n","dense_1 (Dense)              (None, 11)                2827      \n","=================================================================\n","Total params: 1,107,403\n","Trainable params: 1,107,403\n","Non-trainable params: 0\n","_________________________________________________________________\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"OlamJI6vpbaC","colab_type":"code","colab":{}},"source":["model.compile(optimizer=keras.optimizers.Adam(learning_rate=5e-5),\n","              loss=tf.keras.losses.SparseCategoricalCrossentropy(),\n","              metrics=['accuracy'])"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"bhypWI5sINKT","colab_type":"code","colab":{}},"source":["# reshape encoded images as 16*16\n","encoded_images = encoded_images.reshape(encoded_images.shape[0], 16, 16, 1)"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"Kk5jNTIvp2Hc","colab_type":"code","outputId":"12cd0fc1-71fc-498d-f44f-9c4ca4d31baa","executionInfo":{"status":"ok","timestamp":1592049360614,"user_tz":-480,"elapsed":77616,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}},"colab":{"base_uri":"https://localhost:8080/","height":1000}},"source":["# early stop : 若是連續 100 次 epochs 在 validation data 上的 loss 都沒有改善，則會停止訓練\n","callback = tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=100)\n","\n","# 訓練模型 (Training data 與 validation data 各為 70% 與 30%。此外，透過 callbacks 呼叫 early stop)\n","history = model.fit(encoded_images, train_labels, validation_split=0.3, epochs=1000, callbacks=[callback])"],"execution_count":28,"outputs":[{"output_type":"stream","text":["Epoch 1/1000\n","8/8 [==============================] - 0s 36ms/step - loss: 2.2222 - accuracy: 0.2048 - val_loss: 2.0075 - val_accuracy: 0.3832\n","Epoch 2/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 2.0705 - accuracy: 0.3052 - val_loss: 1.9598 - val_accuracy: 0.3832\n","Epoch 3/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 2.0276 - accuracy: 0.3494 - val_loss: 1.8780 - val_accuracy: 0.3832\n","Epoch 4/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 1.9318 - accuracy: 0.3896 - val_loss: 1.8265 - val_accuracy: 0.3832\n","Epoch 5/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 1.8519 - accuracy: 0.4217 - val_loss: 1.7904 - val_accuracy: 0.3832\n","Epoch 6/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 1.8001 - accuracy: 0.4418 - val_loss: 1.7554 - val_accuracy: 0.3832\n","Epoch 7/1000\n","8/8 [==============================] - 0s 9ms/step - loss: 1.7448 - accuracy: 0.4257 - val_loss: 1.7041 - val_accuracy: 0.4112\n","Epoch 8/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 1.7209 - accuracy: 0.4137 - val_loss: 1.6760 - val_accuracy: 0.4393\n","Epoch 9/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 1.6598 - accuracy: 0.4418 - val_loss: 1.6200 - val_accuracy: 0.4299\n","Epoch 10/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 1.6468 - accuracy: 0.4538 - val_loss: 1.6023 - val_accuracy: 0.4299\n","Epoch 11/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 1.5986 - accuracy: 0.4618 - val_loss: 1.5744 - val_accuracy: 0.4673\n","Epoch 12/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 1.5236 - accuracy: 0.4900 - val_loss: 1.5221 - val_accuracy: 0.4393\n","Epoch 13/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 1.4941 - accuracy: 0.4739 - val_loss: 1.4899 - val_accuracy: 0.5047\n","Epoch 14/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 1.4701 - accuracy: 0.5100 - val_loss: 1.4792 - val_accuracy: 0.4766\n","Epoch 15/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 1.4395 - accuracy: 0.5422 - val_loss: 1.4502 - val_accuracy: 0.5140\n","Epoch 16/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 1.4251 - accuracy: 0.5261 - val_loss: 1.4239 - val_accuracy: 0.5514\n","Epoch 17/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 1.3510 - accuracy: 0.5341 - val_loss: 1.3940 - val_accuracy: 0.5421\n","Epoch 18/1000\n","8/8 [==============================] - 0s 9ms/step - loss: 1.3357 - accuracy: 0.5783 - val_loss: 1.3839 - val_accuracy: 0.5234\n","Epoch 19/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 1.3001 - accuracy: 0.5462 - val_loss: 1.3422 - val_accuracy: 0.5514\n","Epoch 20/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 1.2592 - accuracy: 0.6024 - val_loss: 1.3382 - val_accuracy: 0.5234\n","Epoch 21/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 1.2195 - accuracy: 0.6104 - val_loss: 1.3022 - val_accuracy: 0.5327\n","Epoch 22/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 1.2353 - accuracy: 0.5984 - val_loss: 1.2807 - val_accuracy: 0.5607\n","Epoch 23/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 1.1746 - accuracy: 0.6546 - val_loss: 1.2742 - val_accuracy: 0.5514\n","Epoch 24/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 1.1602 - accuracy: 0.6104 - val_loss: 1.2524 - val_accuracy: 0.5421\n","Epoch 25/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 1.1452 - accuracy: 0.6466 - val_loss: 1.2550 - val_accuracy: 0.5327\n","Epoch 26/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 1.1969 - accuracy: 0.6024 - val_loss: 1.2134 - val_accuracy: 0.5514\n","Epoch 27/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 1.1410 - accuracy: 0.6426 - val_loss: 1.2123 - val_accuracy: 0.5888\n","Epoch 28/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 1.0773 - accuracy: 0.6265 - val_loss: 1.1873 - val_accuracy: 0.5421\n","Epoch 29/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 1.0735 - accuracy: 0.6546 - val_loss: 1.1825 - val_accuracy: 0.5607\n","Epoch 30/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 1.1137 - accuracy: 0.6345 - val_loss: 1.1796 - val_accuracy: 0.6262\n","Epoch 31/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 1.0658 - accuracy: 0.6586 - val_loss: 1.1549 - val_accuracy: 0.5701\n","Epoch 32/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 1.0607 - accuracy: 0.6345 - val_loss: 1.1445 - val_accuracy: 0.5794\n","Epoch 33/1000\n","8/8 [==============================] - 0s 10ms/step - loss: 1.0147 - accuracy: 0.6747 - val_loss: 1.1321 - val_accuracy: 0.5981\n","Epoch 34/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 1.0284 - accuracy: 0.6908 - val_loss: 1.1302 - val_accuracy: 0.5888\n","Epoch 35/1000\n","8/8 [==============================] - 0s 9ms/step - loss: 1.0414 - accuracy: 0.6586 - val_loss: 1.0781 - val_accuracy: 0.6542\n","Epoch 36/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 1.0120 - accuracy: 0.6506 - val_loss: 1.1078 - val_accuracy: 0.6168\n","Epoch 37/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.9933 - accuracy: 0.6948 - val_loss: 1.1142 - val_accuracy: 0.6262\n","Epoch 38/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.9310 - accuracy: 0.6867 - val_loss: 1.0610 - val_accuracy: 0.6729\n","Epoch 39/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.9657 - accuracy: 0.6948 - val_loss: 1.0657 - val_accuracy: 0.6542\n","Epoch 40/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.9232 - accuracy: 0.7108 - val_loss: 1.0644 - val_accuracy: 0.6262\n","Epoch 41/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.9889 - accuracy: 0.6667 - val_loss: 1.0434 - val_accuracy: 0.6542\n","Epoch 42/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.8987 - accuracy: 0.7510 - val_loss: 1.0231 - val_accuracy: 0.6636\n","Epoch 43/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.8740 - accuracy: 0.7229 - val_loss: 1.0172 - val_accuracy: 0.6729\n","Epoch 44/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.9112 - accuracy: 0.7189 - val_loss: 1.0266 - val_accuracy: 0.6542\n","Epoch 45/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.8771 - accuracy: 0.7269 - val_loss: 0.9959 - val_accuracy: 0.7103\n","Epoch 46/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.8646 - accuracy: 0.7470 - val_loss: 1.0024 - val_accuracy: 0.6822\n","Epoch 47/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.8884 - accuracy: 0.7028 - val_loss: 0.9925 - val_accuracy: 0.7009\n","Epoch 48/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.8701 - accuracy: 0.6988 - val_loss: 0.9797 - val_accuracy: 0.7009\n","Epoch 49/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.8404 - accuracy: 0.7470 - val_loss: 0.9678 - val_accuracy: 0.6729\n","Epoch 50/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.8016 - accuracy: 0.7671 - val_loss: 0.9790 - val_accuracy: 0.6729\n","Epoch 51/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.8339 - accuracy: 0.7510 - val_loss: 0.9577 - val_accuracy: 0.6636\n","Epoch 52/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.8358 - accuracy: 0.7068 - val_loss: 0.9629 - val_accuracy: 0.7009\n","Epoch 53/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.7739 - accuracy: 0.7871 - val_loss: 0.9572 - val_accuracy: 0.7009\n","Epoch 54/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.7965 - accuracy: 0.7590 - val_loss: 0.9392 - val_accuracy: 0.7103\n","Epoch 55/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.7570 - accuracy: 0.7751 - val_loss: 0.9571 - val_accuracy: 0.6729\n","Epoch 56/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.7678 - accuracy: 0.7711 - val_loss: 0.9292 - val_accuracy: 0.7009\n","Epoch 57/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.7512 - accuracy: 0.7831 - val_loss: 0.9286 - val_accuracy: 0.7196\n","Epoch 58/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.7452 - accuracy: 0.7510 - val_loss: 0.9021 - val_accuracy: 0.7196\n","Epoch 59/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.7893 - accuracy: 0.7430 - val_loss: 0.9226 - val_accuracy: 0.6822\n","Epoch 60/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.7673 - accuracy: 0.7631 - val_loss: 0.9126 - val_accuracy: 0.7196\n","Epoch 61/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.7681 - accuracy: 0.7590 - val_loss: 0.8989 - val_accuracy: 0.7664\n","Epoch 62/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.7574 - accuracy: 0.7510 - val_loss: 0.9152 - val_accuracy: 0.7290\n","Epoch 63/1000\n","8/8 [==============================] - 0s 9ms/step - loss: 0.7286 - accuracy: 0.7831 - val_loss: 0.8856 - val_accuracy: 0.7103\n","Epoch 64/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.7098 - accuracy: 0.8032 - val_loss: 0.8966 - val_accuracy: 0.7196\n","Epoch 65/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.6872 - accuracy: 0.7751 - val_loss: 0.8876 - val_accuracy: 0.7290\n","Epoch 66/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.7250 - accuracy: 0.7952 - val_loss: 0.8717 - val_accuracy: 0.7290\n","Epoch 67/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.6904 - accuracy: 0.7871 - val_loss: 0.8588 - val_accuracy: 0.7757\n","Epoch 68/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.7285 - accuracy: 0.7912 - val_loss: 0.8695 - val_accuracy: 0.7383\n","Epoch 69/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.7145 - accuracy: 0.7590 - val_loss: 0.8592 - val_accuracy: 0.7477\n","Epoch 70/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.7277 - accuracy: 0.7871 - val_loss: 0.8605 - val_accuracy: 0.7290\n","Epoch 71/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.6750 - accuracy: 0.7912 - val_loss: 0.8857 - val_accuracy: 0.6916\n","Epoch 72/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.7018 - accuracy: 0.7711 - val_loss: 0.8712 - val_accuracy: 0.7477\n","Epoch 73/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.6969 - accuracy: 0.7952 - val_loss: 0.8372 - val_accuracy: 0.7477\n","Epoch 74/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.6534 - accuracy: 0.7992 - val_loss: 0.8432 - val_accuracy: 0.7570\n","Epoch 75/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.6843 - accuracy: 0.7992 - val_loss: 0.8344 - val_accuracy: 0.7757\n","Epoch 76/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.6548 - accuracy: 0.7992 - val_loss: 0.8417 - val_accuracy: 0.7196\n","Epoch 77/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.6752 - accuracy: 0.7831 - val_loss: 0.8282 - val_accuracy: 0.7757\n","Epoch 78/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.6414 - accuracy: 0.7711 - val_loss: 0.8267 - val_accuracy: 0.7477\n","Epoch 79/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.6823 - accuracy: 0.7992 - val_loss: 0.8017 - val_accuracy: 0.7850\n","Epoch 80/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.6208 - accuracy: 0.7952 - val_loss: 0.8341 - val_accuracy: 0.7383\n","Epoch 81/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.6480 - accuracy: 0.8193 - val_loss: 0.8201 - val_accuracy: 0.7664\n","Epoch 82/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.6408 - accuracy: 0.7952 - val_loss: 0.8149 - val_accuracy: 0.7757\n","Epoch 83/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.6235 - accuracy: 0.8273 - val_loss: 0.8132 - val_accuracy: 0.7664\n","Epoch 84/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.6085 - accuracy: 0.8193 - val_loss: 0.7968 - val_accuracy: 0.7570\n","Epoch 85/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.5995 - accuracy: 0.8273 - val_loss: 0.7986 - val_accuracy: 0.7477\n","Epoch 86/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.5826 - accuracy: 0.8233 - val_loss: 0.8044 - val_accuracy: 0.7570\n","Epoch 87/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.6165 - accuracy: 0.8112 - val_loss: 0.7895 - val_accuracy: 0.7757\n","Epoch 88/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.6351 - accuracy: 0.8153 - val_loss: 0.7858 - val_accuracy: 0.7757\n","Epoch 89/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.5791 - accuracy: 0.8153 - val_loss: 0.8119 - val_accuracy: 0.7290\n","Epoch 90/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.6152 - accuracy: 0.8313 - val_loss: 0.7947 - val_accuracy: 0.7850\n","Epoch 91/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.6016 - accuracy: 0.8353 - val_loss: 0.7827 - val_accuracy: 0.7850\n","Epoch 92/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.5863 - accuracy: 0.7992 - val_loss: 0.7784 - val_accuracy: 0.8037\n","Epoch 93/1000\n","8/8 [==============================] - 0s 9ms/step - loss: 0.6325 - accuracy: 0.8233 - val_loss: 0.7864 - val_accuracy: 0.7664\n","Epoch 94/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.5744 - accuracy: 0.8394 - val_loss: 0.7893 - val_accuracy: 0.7383\n","Epoch 95/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.5769 - accuracy: 0.8193 - val_loss: 0.7770 - val_accuracy: 0.8131\n","Epoch 96/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.5591 - accuracy: 0.8514 - val_loss: 0.7610 - val_accuracy: 0.8037\n","Epoch 97/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.5824 - accuracy: 0.8353 - val_loss: 0.7540 - val_accuracy: 0.7570\n","Epoch 98/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.6017 - accuracy: 0.8193 - val_loss: 0.7615 - val_accuracy: 0.7757\n","Epoch 99/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.5309 - accuracy: 0.8434 - val_loss: 0.7518 - val_accuracy: 0.7944\n","Epoch 100/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.5836 - accuracy: 0.7912 - val_loss: 0.7510 - val_accuracy: 0.7850\n","Epoch 101/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.5403 - accuracy: 0.8554 - val_loss: 0.7726 - val_accuracy: 0.7570\n","Epoch 102/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.5511 - accuracy: 0.8514 - val_loss: 0.7678 - val_accuracy: 0.7850\n","Epoch 103/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.5081 - accuracy: 0.8514 - val_loss: 0.7694 - val_accuracy: 0.7757\n","Epoch 104/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.5348 - accuracy: 0.8273 - val_loss: 0.7409 - val_accuracy: 0.7944\n","Epoch 105/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.5480 - accuracy: 0.8353 - val_loss: 0.7534 - val_accuracy: 0.8131\n","Epoch 106/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.5769 - accuracy: 0.8153 - val_loss: 0.7612 - val_accuracy: 0.7477\n","Epoch 107/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.5580 - accuracy: 0.8153 - val_loss: 0.7396 - val_accuracy: 0.8037\n","Epoch 108/1000\n","8/8 [==============================] - 0s 9ms/step - loss: 0.5113 - accuracy: 0.8554 - val_loss: 0.7348 - val_accuracy: 0.7850\n","Epoch 109/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.5252 - accuracy: 0.8353 - val_loss: 0.7415 - val_accuracy: 0.7850\n","Epoch 110/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.5403 - accuracy: 0.8233 - val_loss: 0.7284 - val_accuracy: 0.7944\n","Epoch 111/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.5535 - accuracy: 0.8233 - val_loss: 0.7366 - val_accuracy: 0.7850\n","Epoch 112/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.5077 - accuracy: 0.8635 - val_loss: 0.7403 - val_accuracy: 0.7944\n","Epoch 113/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.5413 - accuracy: 0.8514 - val_loss: 0.7234 - val_accuracy: 0.8131\n","Epoch 114/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4862 - accuracy: 0.8755 - val_loss: 0.7217 - val_accuracy: 0.7850\n","Epoch 115/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.5299 - accuracy: 0.8353 - val_loss: 0.7312 - val_accuracy: 0.7850\n","Epoch 116/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4739 - accuracy: 0.8755 - val_loss: 0.7318 - val_accuracy: 0.7850\n","Epoch 117/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.5138 - accuracy: 0.8635 - val_loss: 0.7238 - val_accuracy: 0.7850\n","Epoch 118/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4727 - accuracy: 0.8635 - val_loss: 0.7166 - val_accuracy: 0.7944\n","Epoch 119/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4965 - accuracy: 0.8353 - val_loss: 0.7204 - val_accuracy: 0.8037\n","Epoch 120/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.4702 - accuracy: 0.8675 - val_loss: 0.7086 - val_accuracy: 0.8037\n","Epoch 121/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4574 - accuracy: 0.8635 - val_loss: 0.7069 - val_accuracy: 0.8037\n","Epoch 122/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.4907 - accuracy: 0.8554 - val_loss: 0.7076 - val_accuracy: 0.7944\n","Epoch 123/1000\n","8/8 [==============================] - 0s 9ms/step - loss: 0.5008 - accuracy: 0.8795 - val_loss: 0.6906 - val_accuracy: 0.8131\n","Epoch 124/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.4868 - accuracy: 0.8715 - val_loss: 0.6969 - val_accuracy: 0.8037\n","Epoch 125/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.4556 - accuracy: 0.8835 - val_loss: 0.7061 - val_accuracy: 0.7944\n","Epoch 126/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.4687 - accuracy: 0.8835 - val_loss: 0.7002 - val_accuracy: 0.8131\n","Epoch 127/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.4592 - accuracy: 0.8635 - val_loss: 0.7055 - val_accuracy: 0.8037\n","Epoch 128/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.4640 - accuracy: 0.8795 - val_loss: 0.6848 - val_accuracy: 0.8037\n","Epoch 129/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4393 - accuracy: 0.8795 - val_loss: 0.6921 - val_accuracy: 0.8037\n","Epoch 130/1000\n","8/8 [==============================] - 0s 9ms/step - loss: 0.4460 - accuracy: 0.8635 - val_loss: 0.6986 - val_accuracy: 0.8037\n","Epoch 131/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.4779 - accuracy: 0.8635 - val_loss: 0.7000 - val_accuracy: 0.8224\n","Epoch 132/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4609 - accuracy: 0.8876 - val_loss: 0.6901 - val_accuracy: 0.8131\n","Epoch 133/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4660 - accuracy: 0.8675 - val_loss: 0.6808 - val_accuracy: 0.8037\n","Epoch 134/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4426 - accuracy: 0.8474 - val_loss: 0.6737 - val_accuracy: 0.8224\n","Epoch 135/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4164 - accuracy: 0.8916 - val_loss: 0.7070 - val_accuracy: 0.8037\n","Epoch 136/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4439 - accuracy: 0.8755 - val_loss: 0.6903 - val_accuracy: 0.8131\n","Epoch 137/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4538 - accuracy: 0.8635 - val_loss: 0.6712 - val_accuracy: 0.8224\n","Epoch 138/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.4096 - accuracy: 0.8835 - val_loss: 0.6743 - val_accuracy: 0.8037\n","Epoch 139/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.4388 - accuracy: 0.8835 - val_loss: 0.6795 - val_accuracy: 0.8037\n","Epoch 140/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4286 - accuracy: 0.8755 - val_loss: 0.6851 - val_accuracy: 0.7944\n","Epoch 141/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4022 - accuracy: 0.8916 - val_loss: 0.6820 - val_accuracy: 0.8131\n","Epoch 142/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3944 - accuracy: 0.8755 - val_loss: 0.6795 - val_accuracy: 0.8224\n","Epoch 143/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.3932 - accuracy: 0.9116 - val_loss: 0.6831 - val_accuracy: 0.8037\n","Epoch 144/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3870 - accuracy: 0.8876 - val_loss: 0.6803 - val_accuracy: 0.8131\n","Epoch 145/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4215 - accuracy: 0.8916 - val_loss: 0.6764 - val_accuracy: 0.8224\n","Epoch 146/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4203 - accuracy: 0.8916 - val_loss: 0.6774 - val_accuracy: 0.8131\n","Epoch 147/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4030 - accuracy: 0.9076 - val_loss: 0.6819 - val_accuracy: 0.8131\n","Epoch 148/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4103 - accuracy: 0.8956 - val_loss: 0.6676 - val_accuracy: 0.8131\n","Epoch 149/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.3915 - accuracy: 0.8755 - val_loss: 0.6729 - val_accuracy: 0.8224\n","Epoch 150/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4178 - accuracy: 0.9036 - val_loss: 0.6717 - val_accuracy: 0.8131\n","Epoch 151/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4006 - accuracy: 0.8996 - val_loss: 0.6744 - val_accuracy: 0.8131\n","Epoch 152/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3847 - accuracy: 0.8916 - val_loss: 0.6579 - val_accuracy: 0.8131\n","Epoch 153/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.4266 - accuracy: 0.8715 - val_loss: 0.6593 - val_accuracy: 0.8224\n","Epoch 154/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.3958 - accuracy: 0.8916 - val_loss: 0.6718 - val_accuracy: 0.8037\n","Epoch 155/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.4161 - accuracy: 0.8715 - val_loss: 0.6857 - val_accuracy: 0.8131\n","Epoch 156/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.4112 - accuracy: 0.8835 - val_loss: 0.6617 - val_accuracy: 0.8131\n","Epoch 157/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.3944 - accuracy: 0.8876 - val_loss: 0.6602 - val_accuracy: 0.8037\n","Epoch 158/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.3599 - accuracy: 0.8996 - val_loss: 0.6462 - val_accuracy: 0.8131\n","Epoch 159/1000\n","8/8 [==============================] - 0s 9ms/step - loss: 0.3951 - accuracy: 0.8795 - val_loss: 0.6735 - val_accuracy: 0.8037\n","Epoch 160/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.3972 - accuracy: 0.8715 - val_loss: 0.6844 - val_accuracy: 0.8037\n","Epoch 161/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4118 - accuracy: 0.8916 - val_loss: 0.6595 - val_accuracy: 0.8224\n","Epoch 162/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3977 - accuracy: 0.8795 - val_loss: 0.6475 - val_accuracy: 0.8224\n","Epoch 163/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.4080 - accuracy: 0.8715 - val_loss: 0.6719 - val_accuracy: 0.8224\n","Epoch 164/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3828 - accuracy: 0.8835 - val_loss: 0.6573 - val_accuracy: 0.8224\n","Epoch 165/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3395 - accuracy: 0.8956 - val_loss: 0.6497 - val_accuracy: 0.8131\n","Epoch 166/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3896 - accuracy: 0.8916 - val_loss: 0.6577 - val_accuracy: 0.8224\n","Epoch 167/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3422 - accuracy: 0.8996 - val_loss: 0.6522 - val_accuracy: 0.8224\n","Epoch 168/1000\n","8/8 [==============================] - 0s 9ms/step - loss: 0.3934 - accuracy: 0.8916 - val_loss: 0.6394 - val_accuracy: 0.8224\n","Epoch 169/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3907 - accuracy: 0.8876 - val_loss: 0.6501 - val_accuracy: 0.8037\n","Epoch 170/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3385 - accuracy: 0.9116 - val_loss: 0.6582 - val_accuracy: 0.8224\n","Epoch 171/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3856 - accuracy: 0.8916 - val_loss: 0.6480 - val_accuracy: 0.8131\n","Epoch 172/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3713 - accuracy: 0.8996 - val_loss: 0.6662 - val_accuracy: 0.8131\n","Epoch 173/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.3641 - accuracy: 0.8795 - val_loss: 0.6676 - val_accuracy: 0.8131\n","Epoch 174/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3358 - accuracy: 0.9237 - val_loss: 0.6520 - val_accuracy: 0.8224\n","Epoch 175/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3310 - accuracy: 0.9036 - val_loss: 0.6379 - val_accuracy: 0.8131\n","Epoch 176/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3478 - accuracy: 0.8956 - val_loss: 0.6548 - val_accuracy: 0.8131\n","Epoch 177/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3390 - accuracy: 0.9076 - val_loss: 0.6513 - val_accuracy: 0.8131\n","Epoch 178/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.3494 - accuracy: 0.8996 - val_loss: 0.6473 - val_accuracy: 0.8131\n","Epoch 179/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3280 - accuracy: 0.9157 - val_loss: 0.6382 - val_accuracy: 0.8131\n","Epoch 180/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3249 - accuracy: 0.8996 - val_loss: 0.6474 - val_accuracy: 0.8224\n","Epoch 181/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3100 - accuracy: 0.9116 - val_loss: 0.6393 - val_accuracy: 0.8318\n","Epoch 182/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3468 - accuracy: 0.8956 - val_loss: 0.6415 - val_accuracy: 0.8318\n","Epoch 183/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.3325 - accuracy: 0.9277 - val_loss: 0.6522 - val_accuracy: 0.8411\n","Epoch 184/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3626 - accuracy: 0.8956 - val_loss: 0.6429 - val_accuracy: 0.8318\n","Epoch 185/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3254 - accuracy: 0.9237 - val_loss: 0.6305 - val_accuracy: 0.8224\n","Epoch 186/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3538 - accuracy: 0.9116 - val_loss: 0.6470 - val_accuracy: 0.8131\n","Epoch 187/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.3089 - accuracy: 0.9237 - val_loss: 0.6579 - val_accuracy: 0.8131\n","Epoch 188/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3161 - accuracy: 0.9076 - val_loss: 0.6495 - val_accuracy: 0.8224\n","Epoch 189/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3042 - accuracy: 0.9076 - val_loss: 0.6248 - val_accuracy: 0.8224\n","Epoch 190/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3371 - accuracy: 0.9076 - val_loss: 0.6354 - val_accuracy: 0.8224\n","Epoch 191/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3056 - accuracy: 0.9197 - val_loss: 0.6392 - val_accuracy: 0.8318\n","Epoch 192/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2896 - accuracy: 0.9197 - val_loss: 0.6280 - val_accuracy: 0.8224\n","Epoch 193/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2882 - accuracy: 0.9116 - val_loss: 0.6283 - val_accuracy: 0.8318\n","Epoch 194/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2767 - accuracy: 0.9357 - val_loss: 0.6351 - val_accuracy: 0.8131\n","Epoch 195/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3178 - accuracy: 0.9116 - val_loss: 0.6424 - val_accuracy: 0.8131\n","Epoch 196/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3077 - accuracy: 0.9076 - val_loss: 0.6487 - val_accuracy: 0.8318\n","Epoch 197/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.3092 - accuracy: 0.9277 - val_loss: 0.6425 - val_accuracy: 0.8224\n","Epoch 198/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2576 - accuracy: 0.9357 - val_loss: 0.6519 - val_accuracy: 0.8224\n","Epoch 199/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2813 - accuracy: 0.9197 - val_loss: 0.6594 - val_accuracy: 0.8224\n","Epoch 200/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3102 - accuracy: 0.9237 - val_loss: 0.6545 - val_accuracy: 0.8224\n","Epoch 201/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2940 - accuracy: 0.9237 - val_loss: 0.6334 - val_accuracy: 0.8131\n","Epoch 202/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2865 - accuracy: 0.9197 - val_loss: 0.6489 - val_accuracy: 0.8224\n","Epoch 203/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2907 - accuracy: 0.9157 - val_loss: 0.6444 - val_accuracy: 0.8224\n","Epoch 204/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2946 - accuracy: 0.9157 - val_loss: 0.6378 - val_accuracy: 0.8318\n","Epoch 205/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2739 - accuracy: 0.9438 - val_loss: 0.6317 - val_accuracy: 0.8224\n","Epoch 206/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2878 - accuracy: 0.9116 - val_loss: 0.6218 - val_accuracy: 0.8318\n","Epoch 207/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2964 - accuracy: 0.9157 - val_loss: 0.6483 - val_accuracy: 0.8224\n","Epoch 208/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2907 - accuracy: 0.9197 - val_loss: 0.6419 - val_accuracy: 0.8037\n","Epoch 209/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.3101 - accuracy: 0.9157 - val_loss: 0.6308 - val_accuracy: 0.8224\n","Epoch 210/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2915 - accuracy: 0.9277 - val_loss: 0.6489 - val_accuracy: 0.8037\n","Epoch 211/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2928 - accuracy: 0.9277 - val_loss: 0.6496 - val_accuracy: 0.8224\n","Epoch 212/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2848 - accuracy: 0.9317 - val_loss: 0.6339 - val_accuracy: 0.8131\n","Epoch 213/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2954 - accuracy: 0.9197 - val_loss: 0.6328 - val_accuracy: 0.8131\n","Epoch 214/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2694 - accuracy: 0.9317 - val_loss: 0.6396 - val_accuracy: 0.8318\n","Epoch 215/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2772 - accuracy: 0.9237 - val_loss: 0.6240 - val_accuracy: 0.8318\n","Epoch 216/1000\n","8/8 [==============================] - 0s 9ms/step - loss: 0.2917 - accuracy: 0.9237 - val_loss: 0.6324 - val_accuracy: 0.8131\n","Epoch 217/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2852 - accuracy: 0.9237 - val_loss: 0.6228 - val_accuracy: 0.8131\n","Epoch 218/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2375 - accuracy: 0.9398 - val_loss: 0.6144 - val_accuracy: 0.8318\n","Epoch 219/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2717 - accuracy: 0.9237 - val_loss: 0.6392 - val_accuracy: 0.8224\n","Epoch 220/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2853 - accuracy: 0.9197 - val_loss: 0.6189 - val_accuracy: 0.8224\n","Epoch 221/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2481 - accuracy: 0.9357 - val_loss: 0.6294 - val_accuracy: 0.8318\n","Epoch 222/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2320 - accuracy: 0.9438 - val_loss: 0.6349 - val_accuracy: 0.8318\n","Epoch 223/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2661 - accuracy: 0.9036 - val_loss: 0.6138 - val_accuracy: 0.8411\n","Epoch 224/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2465 - accuracy: 0.9277 - val_loss: 0.6293 - val_accuracy: 0.8131\n","Epoch 225/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2588 - accuracy: 0.9317 - val_loss: 0.6284 - val_accuracy: 0.8318\n","Epoch 226/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2531 - accuracy: 0.9277 - val_loss: 0.6261 - val_accuracy: 0.8224\n","Epoch 227/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2011 - accuracy: 0.9518 - val_loss: 0.6394 - val_accuracy: 0.8224\n","Epoch 228/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2541 - accuracy: 0.9277 - val_loss: 0.6203 - val_accuracy: 0.8224\n","Epoch 229/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2750 - accuracy: 0.9116 - val_loss: 0.6314 - val_accuracy: 0.8318\n","Epoch 230/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2477 - accuracy: 0.9398 - val_loss: 0.6480 - val_accuracy: 0.8224\n","Epoch 231/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2539 - accuracy: 0.9357 - val_loss: 0.6203 - val_accuracy: 0.8224\n","Epoch 232/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2575 - accuracy: 0.9277 - val_loss: 0.6278 - val_accuracy: 0.8318\n","Epoch 233/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2160 - accuracy: 0.9357 - val_loss: 0.6481 - val_accuracy: 0.8224\n","Epoch 234/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2531 - accuracy: 0.9277 - val_loss: 0.6338 - val_accuracy: 0.8224\n","Epoch 235/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2527 - accuracy: 0.9116 - val_loss: 0.6435 - val_accuracy: 0.8131\n","Epoch 236/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2315 - accuracy: 0.9438 - val_loss: 0.6545 - val_accuracy: 0.8224\n","Epoch 237/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2408 - accuracy: 0.9398 - val_loss: 0.6255 - val_accuracy: 0.8318\n","Epoch 238/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2294 - accuracy: 0.9438 - val_loss: 0.6347 - val_accuracy: 0.8224\n","Epoch 239/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2395 - accuracy: 0.9398 - val_loss: 0.6217 - val_accuracy: 0.8318\n","Epoch 240/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2222 - accuracy: 0.9398 - val_loss: 0.6301 - val_accuracy: 0.8318\n","Epoch 241/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2129 - accuracy: 0.9317 - val_loss: 0.6445 - val_accuracy: 0.8131\n","Epoch 242/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2355 - accuracy: 0.9357 - val_loss: 0.6327 - val_accuracy: 0.8224\n","Epoch 243/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2333 - accuracy: 0.9357 - val_loss: 0.6441 - val_accuracy: 0.8224\n","Epoch 244/1000\n","8/8 [==============================] - 0s 9ms/step - loss: 0.2173 - accuracy: 0.9357 - val_loss: 0.6352 - val_accuracy: 0.8131\n","Epoch 245/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2539 - accuracy: 0.9116 - val_loss: 0.6350 - val_accuracy: 0.8318\n","Epoch 246/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2467 - accuracy: 0.9357 - val_loss: 0.6395 - val_accuracy: 0.8318\n","Epoch 247/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2285 - accuracy: 0.9398 - val_loss: 0.6355 - val_accuracy: 0.8131\n","Epoch 248/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2190 - accuracy: 0.9398 - val_loss: 0.6135 - val_accuracy: 0.8318\n","Epoch 249/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2248 - accuracy: 0.9438 - val_loss: 0.6275 - val_accuracy: 0.8318\n","Epoch 250/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2228 - accuracy: 0.9357 - val_loss: 0.6254 - val_accuracy: 0.8318\n","Epoch 251/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2298 - accuracy: 0.9398 - val_loss: 0.6499 - val_accuracy: 0.8131\n","Epoch 252/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2445 - accuracy: 0.9237 - val_loss: 0.6239 - val_accuracy: 0.8318\n","Epoch 253/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2199 - accuracy: 0.9438 - val_loss: 0.6109 - val_accuracy: 0.8224\n","Epoch 254/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2080 - accuracy: 0.9357 - val_loss: 0.6194 - val_accuracy: 0.8318\n","Epoch 255/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2289 - accuracy: 0.9478 - val_loss: 0.6070 - val_accuracy: 0.8318\n","Epoch 256/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2324 - accuracy: 0.9277 - val_loss: 0.6333 - val_accuracy: 0.8224\n","Epoch 257/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2116 - accuracy: 0.9438 - val_loss: 0.6275 - val_accuracy: 0.8411\n","Epoch 258/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2528 - accuracy: 0.9317 - val_loss: 0.6203 - val_accuracy: 0.8318\n","Epoch 259/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2017 - accuracy: 0.9438 - val_loss: 0.6310 - val_accuracy: 0.8224\n","Epoch 260/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2122 - accuracy: 0.9398 - val_loss: 0.6261 - val_accuracy: 0.8224\n","Epoch 261/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1881 - accuracy: 0.9438 - val_loss: 0.6361 - val_accuracy: 0.8224\n","Epoch 262/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2028 - accuracy: 0.9357 - val_loss: 0.6400 - val_accuracy: 0.8224\n","Epoch 263/1000\n","8/8 [==============================] - 0s 9ms/step - loss: 0.2134 - accuracy: 0.9478 - val_loss: 0.6217 - val_accuracy: 0.8224\n","Epoch 264/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.2064 - accuracy: 0.9518 - val_loss: 0.6264 - val_accuracy: 0.8224\n","Epoch 265/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.2186 - accuracy: 0.9398 - val_loss: 0.6339 - val_accuracy: 0.8224\n","Epoch 266/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1960 - accuracy: 0.9598 - val_loss: 0.6074 - val_accuracy: 0.8318\n","Epoch 267/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1955 - accuracy: 0.9357 - val_loss: 0.6269 - val_accuracy: 0.8318\n","Epoch 268/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1689 - accuracy: 0.9558 - val_loss: 0.6424 - val_accuracy: 0.8318\n","Epoch 269/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1990 - accuracy: 0.9478 - val_loss: 0.6193 - val_accuracy: 0.8318\n","Epoch 270/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1847 - accuracy: 0.9558 - val_loss: 0.6340 - val_accuracy: 0.8131\n","Epoch 271/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1928 - accuracy: 0.9478 - val_loss: 0.6573 - val_accuracy: 0.8224\n","Epoch 272/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1818 - accuracy: 0.9478 - val_loss: 0.6192 - val_accuracy: 0.8224\n","Epoch 273/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1800 - accuracy: 0.9478 - val_loss: 0.6208 - val_accuracy: 0.8224\n","Epoch 274/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1819 - accuracy: 0.9598 - val_loss: 0.6520 - val_accuracy: 0.8224\n","Epoch 275/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1828 - accuracy: 0.9558 - val_loss: 0.6257 - val_accuracy: 0.8318\n","Epoch 276/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1768 - accuracy: 0.9558 - val_loss: 0.6183 - val_accuracy: 0.8318\n","Epoch 277/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1828 - accuracy: 0.9639 - val_loss: 0.6259 - val_accuracy: 0.8318\n","Epoch 278/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1727 - accuracy: 0.9598 - val_loss: 0.6337 - val_accuracy: 0.8224\n","Epoch 279/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1773 - accuracy: 0.9558 - val_loss: 0.6363 - val_accuracy: 0.8318\n","Epoch 280/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1974 - accuracy: 0.9438 - val_loss: 0.6639 - val_accuracy: 0.8224\n","Epoch 281/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1868 - accuracy: 0.9558 - val_loss: 0.6220 - val_accuracy: 0.8318\n","Epoch 282/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1850 - accuracy: 0.9639 - val_loss: 0.6273 - val_accuracy: 0.8505\n","Epoch 283/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1903 - accuracy: 0.9598 - val_loss: 0.6465 - val_accuracy: 0.8224\n","Epoch 284/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1782 - accuracy: 0.9558 - val_loss: 0.6377 - val_accuracy: 0.8318\n","Epoch 285/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1676 - accuracy: 0.9598 - val_loss: 0.6173 - val_accuracy: 0.8318\n","Epoch 286/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1854 - accuracy: 0.9598 - val_loss: 0.6389 - val_accuracy: 0.8224\n","Epoch 287/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1691 - accuracy: 0.9679 - val_loss: 0.6269 - val_accuracy: 0.8224\n","Epoch 288/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1576 - accuracy: 0.9558 - val_loss: 0.6202 - val_accuracy: 0.8318\n","Epoch 289/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1827 - accuracy: 0.9518 - val_loss: 0.6620 - val_accuracy: 0.8318\n","Epoch 290/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1699 - accuracy: 0.9478 - val_loss: 0.6282 - val_accuracy: 0.8224\n","Epoch 291/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1577 - accuracy: 0.9719 - val_loss: 0.6324 - val_accuracy: 0.8224\n","Epoch 292/1000\n","8/8 [==============================] - 0s 9ms/step - loss: 0.1829 - accuracy: 0.9357 - val_loss: 0.6278 - val_accuracy: 0.8318\n","Epoch 293/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1647 - accuracy: 0.9639 - val_loss: 0.6159 - val_accuracy: 0.8411\n","Epoch 294/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1813 - accuracy: 0.9398 - val_loss: 0.6618 - val_accuracy: 0.8224\n","Epoch 295/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1686 - accuracy: 0.9639 - val_loss: 0.6391 - val_accuracy: 0.8318\n","Epoch 296/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1602 - accuracy: 0.9558 - val_loss: 0.6535 - val_accuracy: 0.8224\n","Epoch 297/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1450 - accuracy: 0.9639 - val_loss: 0.6429 - val_accuracy: 0.8224\n","Epoch 298/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1466 - accuracy: 0.9598 - val_loss: 0.6159 - val_accuracy: 0.8224\n","Epoch 299/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1635 - accuracy: 0.9518 - val_loss: 0.6418 - val_accuracy: 0.8318\n","Epoch 300/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1886 - accuracy: 0.9558 - val_loss: 0.6578 - val_accuracy: 0.8224\n","Epoch 301/1000\n","8/8 [==============================] - 0s 9ms/step - loss: 0.1831 - accuracy: 0.9639 - val_loss: 0.6550 - val_accuracy: 0.8318\n","Epoch 302/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1658 - accuracy: 0.9598 - val_loss: 0.6564 - val_accuracy: 0.8224\n","Epoch 303/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1627 - accuracy: 0.9598 - val_loss: 0.6586 - val_accuracy: 0.8224\n","Epoch 304/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1636 - accuracy: 0.9558 - val_loss: 0.6370 - val_accuracy: 0.8224\n","Epoch 305/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1403 - accuracy: 0.9639 - val_loss: 0.6624 - val_accuracy: 0.8224\n","Epoch 306/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1609 - accuracy: 0.9478 - val_loss: 0.6475 - val_accuracy: 0.8318\n","Epoch 307/1000\n","8/8 [==============================] - 0s 9ms/step - loss: 0.1759 - accuracy: 0.9558 - val_loss: 0.6231 - val_accuracy: 0.8411\n","Epoch 308/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1430 - accuracy: 0.9759 - val_loss: 0.6406 - val_accuracy: 0.8411\n","Epoch 309/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1342 - accuracy: 0.9719 - val_loss: 0.6520 - val_accuracy: 0.8224\n","Epoch 310/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1941 - accuracy: 0.9478 - val_loss: 0.6459 - val_accuracy: 0.8131\n","Epoch 311/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1403 - accuracy: 0.9719 - val_loss: 0.6357 - val_accuracy: 0.8224\n","Epoch 312/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1508 - accuracy: 0.9679 - val_loss: 0.6337 - val_accuracy: 0.8318\n","Epoch 313/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1466 - accuracy: 0.9518 - val_loss: 0.6535 - val_accuracy: 0.8224\n","Epoch 314/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1446 - accuracy: 0.9598 - val_loss: 0.6566 - val_accuracy: 0.8318\n","Epoch 315/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1626 - accuracy: 0.9558 - val_loss: 0.6219 - val_accuracy: 0.8224\n","Epoch 316/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1691 - accuracy: 0.9478 - val_loss: 0.6529 - val_accuracy: 0.8131\n","Epoch 317/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1603 - accuracy: 0.9518 - val_loss: 0.6402 - val_accuracy: 0.8224\n","Epoch 318/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1557 - accuracy: 0.9598 - val_loss: 0.6335 - val_accuracy: 0.8131\n","Epoch 319/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1488 - accuracy: 0.9679 - val_loss: 0.6529 - val_accuracy: 0.8224\n","Epoch 320/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1313 - accuracy: 0.9799 - val_loss: 0.6520 - val_accuracy: 0.8224\n","Epoch 321/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1387 - accuracy: 0.9518 - val_loss: 0.6390 - val_accuracy: 0.8224\n","Epoch 322/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1274 - accuracy: 0.9799 - val_loss: 0.6576 - val_accuracy: 0.8131\n","Epoch 323/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1513 - accuracy: 0.9639 - val_loss: 0.6392 - val_accuracy: 0.8411\n","Epoch 324/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1471 - accuracy: 0.9598 - val_loss: 0.6690 - val_accuracy: 0.8224\n","Epoch 325/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1309 - accuracy: 0.9679 - val_loss: 0.6545 - val_accuracy: 0.8411\n","Epoch 326/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1385 - accuracy: 0.9518 - val_loss: 0.6750 - val_accuracy: 0.8131\n","Epoch 327/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1586 - accuracy: 0.9598 - val_loss: 0.6619 - val_accuracy: 0.8224\n","Epoch 328/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1362 - accuracy: 0.9719 - val_loss: 0.6554 - val_accuracy: 0.8318\n","Epoch 329/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1456 - accuracy: 0.9639 - val_loss: 0.6389 - val_accuracy: 0.8318\n","Epoch 330/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1124 - accuracy: 0.9759 - val_loss: 0.6524 - val_accuracy: 0.8318\n","Epoch 331/1000\n","8/8 [==============================] - 0s 9ms/step - loss: 0.1345 - accuracy: 0.9639 - val_loss: 0.6362 - val_accuracy: 0.8318\n","Epoch 332/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1483 - accuracy: 0.9598 - val_loss: 0.6502 - val_accuracy: 0.8224\n","Epoch 333/1000\n","8/8 [==============================] - 0s 9ms/step - loss: 0.1196 - accuracy: 0.9679 - val_loss: 0.6329 - val_accuracy: 0.8224\n","Epoch 334/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1167 - accuracy: 0.9719 - val_loss: 0.6355 - val_accuracy: 0.8224\n","Epoch 335/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1373 - accuracy: 0.9639 - val_loss: 0.6648 - val_accuracy: 0.8224\n","Epoch 336/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1316 - accuracy: 0.9759 - val_loss: 0.6453 - val_accuracy: 0.8318\n","Epoch 337/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1315 - accuracy: 0.9558 - val_loss: 0.6502 - val_accuracy: 0.8318\n","Epoch 338/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1265 - accuracy: 0.9679 - val_loss: 0.6693 - val_accuracy: 0.8318\n","Epoch 339/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1231 - accuracy: 0.9679 - val_loss: 0.6561 - val_accuracy: 0.8318\n","Epoch 340/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1094 - accuracy: 0.9759 - val_loss: 0.6783 - val_accuracy: 0.8224\n","Epoch 341/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1329 - accuracy: 0.9719 - val_loss: 0.6809 - val_accuracy: 0.8224\n","Epoch 342/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1376 - accuracy: 0.9598 - val_loss: 0.6554 - val_accuracy: 0.8224\n","Epoch 343/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1390 - accuracy: 0.9639 - val_loss: 0.6594 - val_accuracy: 0.8224\n","Epoch 344/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1414 - accuracy: 0.9518 - val_loss: 0.6552 - val_accuracy: 0.8318\n","Epoch 345/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1117 - accuracy: 0.9719 - val_loss: 0.6599 - val_accuracy: 0.8318\n","Epoch 346/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1113 - accuracy: 0.9759 - val_loss: 0.6693 - val_accuracy: 0.8224\n","Epoch 347/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1294 - accuracy: 0.9598 - val_loss: 0.6559 - val_accuracy: 0.8224\n","Epoch 348/1000\n","8/8 [==============================] - 0s 9ms/step - loss: 0.1267 - accuracy: 0.9759 - val_loss: 0.6296 - val_accuracy: 0.8318\n","Epoch 349/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1210 - accuracy: 0.9639 - val_loss: 0.6532 - val_accuracy: 0.8318\n","Epoch 350/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1377 - accuracy: 0.9558 - val_loss: 0.6517 - val_accuracy: 0.8224\n","Epoch 351/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1166 - accuracy: 0.9759 - val_loss: 0.6590 - val_accuracy: 0.8224\n","Epoch 352/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1122 - accuracy: 0.9759 - val_loss: 0.6748 - val_accuracy: 0.8318\n","Epoch 353/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1025 - accuracy: 0.9759 - val_loss: 0.6718 - val_accuracy: 0.8318\n","Epoch 354/1000\n","8/8 [==============================] - 0s 8ms/step - loss: 0.1051 - accuracy: 0.9719 - val_loss: 0.6465 - val_accuracy: 0.8318\n","Epoch 355/1000\n","8/8 [==============================] - 0s 7ms/step - loss: 0.1123 - accuracy: 0.9639 - val_loss: 0.6723 - val_accuracy: 0.8318\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"G2pCN5RsVgtM","colab_type":"code","outputId":"62a889c6-9a82-4763-a7e5-af95c90265fa","executionInfo":{"status":"ok","timestamp":1592049360615,"user_tz":-480,"elapsed":77587,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}},"colab":{"base_uri":"https://localhost:8080/","height":302}},"source":["# Training Accuracy 與 Validation Accuracy 的變化\n","print(\"The training accuracy: {}%\".format(round(history.history[\"accuracy\"][-1]*100, 2)))\n","print(\"The validation accuracy: {}%\".format(round(history.history[\"val_accuracy\"][-1]*100, 2)))\n","\n","acc = history.history[\"accuracy\"]\n","val_acc = history.history[\"val_accuracy\"]\n","\n","plt.plot(acc, \"b-\", label=\"acc\")\n","plt.plot(val_acc, \"r-\", label=\"val_acc\")\n","plt.legend()\n","plt.show()"],"execution_count":29,"outputs":[{"output_type":"stream","text":["The training accuracy: 96.39%\n","The validation accuracy: 83.18%\n"],"name":"stdout"},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"QVvGctwPOWZm","colab_type":"text"},"source":["Q1: PCA or AutoEncoder? Which is better for data representation?\n","\n","A1: \n","\n","AutoEncoder 對於影像資料的呈現與降維的能力比 PCA 更好，因為\n","\n","1) AutoEncoder 的概念類似影像的壓縮，將原始影像對應到某個 latent space，因此在 encoded 後的影像，仍能保有原本影像的部分模樣，而若是使用 PCA 做降維，則就只是單純的降維，看不到原始影像的樣態。\n","\n","2) PCA 只能做線性的降維，然而 AutoEncoder 透過 NN 可以做到非線性的降維。因此在此題影像處理上，影像的 pixel 之間很有可能有非線性的關聯，因此透過 AutoEncoder 來做降維可以保留更多重要的特徵。"]},{"cell_type":"markdown","metadata":{"id":"BFaFurKhRsxM","colab_type":"text"},"source":["Q2: What is your design of NN? Your classification accuracy rate?\n","\n","A2: \n","\n","先訓練一個 AutoEncoder 來做資料的降維，而後將原始影像當作 input 進入此 AutoEncoder 的 Encoder，得到降維後的影像，再將此降維後的影像當作 input 進入另一個 CNN 分類模型進行訓練，最後得到訓練好的 CNN 模型。\n","\n","最終 CNN model 在 training data 上的準確率為 96.39%，在 validation data 上的準確率為 83.18%"]},{"cell_type":"markdown","metadata":{"id":"DU10wjpIY1va","colab_type":"text"},"source":["Q3: Implement certain ML algorithms on the same dataset. ML is better or NN is better?"]},{"cell_type":"code","metadata":{"id":"3nSy_UB2b2JW","colab_type":"code","colab":{}},"source":["dt_encoded_images = encoded_images.reshape(encoded_images.shape[0], 256)"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"I54YKJrTb-tB","colab_type":"code","colab":{}},"source":["from sklearn.model_selection import train_test_split\n","# Train, test splits\n","X_train, X_test, y_train, y_test = train_test_split(dt_encoded_images, train_labels, test_size=.3, stratify=train_labels, random_state=42)"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"wjscImqmaJbR","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":119},"outputId":"defa41bf-7ada-42b1-d8a0-6bdf4b0ba782","executionInfo":{"status":"ok","timestamp":1592049360617,"user_tz":-480,"elapsed":77496,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}}},"source":["from sklearn.tree import DecisionTreeClassifier\n","\n","dt = DecisionTreeClassifier(max_depth=10, random_state=42)\n","dt.fit(X_train, y_train)"],"execution_count":32,"outputs":[{"output_type":"execute_result","data":{"text/plain":["DecisionTreeClassifier(ccp_alpha=0.0, class_weight=None, criterion='gini',\n","                       max_depth=10, max_features=None, max_leaf_nodes=None,\n","                       min_impurity_decrease=0.0, min_impurity_split=None,\n","                       min_samples_leaf=1, min_samples_split=2,\n","                       min_weight_fraction_leaf=0.0, presort='deprecated',\n","                       random_state=42, splitter='best')"]},"metadata":{"tags":[]},"execution_count":32}]},{"cell_type":"code","metadata":{"id":"9YAdeYrFbW9o","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":34},"outputId":"492f51ec-a7e8-489f-ee61-20a3b18327ae","executionInfo":{"status":"ok","timestamp":1592049360618,"user_tz":-480,"elapsed":77477,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}}},"source":["train_acc = round(dt.score(X_train, y_train)*100, 2)\n","print(\"Training Accuracy: {} %\".format(train_acc))"],"execution_count":33,"outputs":[{"output_type":"stream","text":["Training Accuracy: 89.96 %\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"9-kkvVWrbk-e","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":34},"outputId":"33d30b70-3ba2-40f8-976e-b94f5a3ec451","executionInfo":{"status":"ok","timestamp":1592049360618,"user_tz":-480,"elapsed":77458,"user":{"displayName":"郭宇雋","photoUrl":"","userId":"01461113603099005024"}}},"source":["from sklearn.metrics import accuracy_score, roc_auc_score\n","# 預測 Test\n","y_pred = dt.predict(X_test)\n","# 計算 Accuracy\n","test_acc = round(accuracy_score(y_test, y_pred)*100, 2)\n","print('Accuracy: {}%'.format(test_acc))"],"execution_count":34,"outputs":[{"output_type":"stream","text":["Accuracy: 71.03%\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"vfS0aPtBc9Ju","colab_type":"text"},"source":["A3: \n","\n","使用傳統 ML 方法 (Decision Tree)，在 training data 上的準確率為 89.96%，在 validation data 上的準確率為 71.03%。\n","\n","可以發現 NN 在 training 與 validation 上的準確率都比傳統 ML 方法 (Decision Tree) 來的好一些。"]}]}