inception-ft-multigpu.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "cb601888",
   "metadata": {
    "id": "l3FoTyRa9pLu",
    "papermill": {
     "duration": 0.01328,
     "end_time": "2022-11-15T19:33:30.510695",
     "exception": false,
     "start_time": "2022-11-15T19:33:30.497415",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# **LIBRARIES**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aaad533c",
   "metadata": {
    "papermill": {
     "duration": 0.007833,
     "end_time": "2022-11-15T19:33:30.528212",
     "exception": false,
     "start_time": "2022-11-15T19:33:30.520379",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "The first attempt tries to use Inception for TL with no other fancy things. \n",
    "MAYBE having a \"mix\" of receptive fields like inception does is good.  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "5114dd24",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-11-15T19:33:30.546157Z",
     "iopub.status.busy": "2022-11-15T19:33:30.545534Z",
     "iopub.status.idle": "2022-11-15T19:33:37.838871Z",
     "shell.execute_reply": "2022-11-15T19:33:37.836877Z"
    },
    "id": "f_sOaV1Y8NsL",
    "outputId": "62e1649f-24e0-4dce-daba-9580ccaac6ab",
    "papermill": {
     "duration": 7.305245,
     "end_time": "2022-11-15T19:33:37.841454",
     "exception": false,
     "start_time": "2022-11-15T19:33:30.536209",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.6.4\n"
     ]
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "from numpy import asarray\n",
    "import os\n",
    "import random\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import accuracy_score, f1_score, precision_score, recall_score\n",
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "from PIL import Image\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
    "from sklearn.utils import class_weight\n",
    "import datetime \n",
    "import pathlib\n",
    "\n",
    "tfk = tf.keras\n",
    "tfkl = tf.keras.layers\n",
    "print(tf.__version__)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "ce1779c1",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-11-15T19:33:37.854974Z",
     "iopub.status.busy": "2022-11-15T19:33:37.854486Z",
     "iopub.status.idle": "2022-11-15T19:33:37.860849Z",
     "shell.execute_reply": "2022-11-15T19:33:37.860003Z"
    },
    "id": "C7HYua8HUHIj",
    "papermill": {
     "duration": 0.015305,
     "end_time": "2022-11-15T19:33:37.862881",
     "exception": false,
     "start_time": "2022-11-15T19:33:37.847576",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "seed = 42127\n",
    "random.seed(seed)\n",
    "os.environ['PYTHONHASHSEED'] = str(seed)\n",
    "np.random.seed(seed)\n",
    "tf.random.set_seed(seed)\n",
    "tf.compat.v1.set_random_seed(seed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "e4892564",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-11-15T19:33:37.875814Z",
     "iopub.status.busy": "2022-11-15T19:33:37.875092Z",
     "iopub.status.idle": "2022-11-15T19:33:42.629314Z",
     "shell.execute_reply": "2022-11-15T19:33:42.628359Z"
    },
    "papermill": {
     "duration": 4.76347,
     "end_time": "2022-11-15T19:33:42.632015",
     "exception": false,
     "start_time": "2022-11-15T19:33:37.868545",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-11-15 19:33:38.002009: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2022-11-15 19:33:38.002981: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2022-11-15 19:33:38.165222: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2022-11-15 19:33:38.166182: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2022-11-15 19:33:38.167009: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2022-11-15 19:33:38.167808: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2022-11-15 19:33:38.171231: I tensorflow/core/platform/cpu_feature_guard.cc:142] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 AVX512F FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
      "2022-11-15 19:33:38.433840: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2022-11-15 19:33:38.434782: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2022-11-15 19:33:38.435506: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2022-11-15 19:33:38.436185: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2022-11-15 19:33:38.436845: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2022-11-15 19:33:38.437533: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2022-11-15 19:33:41.936217: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2022-11-15 19:33:41.937509: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2022-11-15 19:33:41.938640: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2022-11-15 19:33:41.939823: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2022-11-15 19:33:41.940982: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2022-11-15 19:33:41.941989: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1510] Created device /job:localhost/replica:0/task:0/device:GPU:0 with 13789 MB memory:  -> device: 0, name: Tesla T4, pci bus id: 0000:00:04.0, compute capability: 7.5\n",
      "2022-11-15 19:33:41.947216: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2022-11-15 19:33:41.948198: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1510] Created device /job:localhost/replica:0/task:0/device:GPU:1 with 13789 MB memory:  -> device: 1, name: Tesla T4, pci bus id: 0000:00:05.0, compute capability: 7.5\n"
     ]
    }
   ],
   "source": [
    "mirrored_strategy = tf.distribute.MirroredStrategy()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2179478a",
   "metadata": {
    "id": "BDM82PpE3VSg",
    "papermill": {
     "duration": 0.005651,
     "end_time": "2022-11-15T19:33:42.643959",
     "exception": false,
     "start_time": "2022-11-15T19:33:42.638308",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "### Suppress warnings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "f21bc354",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-11-15T19:33:42.658360Z",
     "iopub.status.busy": "2022-11-15T19:33:42.657763Z",
     "iopub.status.idle": "2022-11-15T19:33:42.664368Z",
     "shell.execute_reply": "2022-11-15T19:33:42.663664Z"
    },
    "id": "5fXtacjAqOIq",
    "papermill": {
     "duration": 0.016346,
     "end_time": "2022-11-15T19:33:42.666128",
     "exception": false,
     "start_time": "2022-11-15T19:33:42.649782",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "import warnings\n",
    "import logging\n",
    "\n",
    "os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'\n",
    "warnings.simplefilter(action='ignore', category=FutureWarning)\n",
    "warnings.simplefilter(action='ignore', category=Warning)\n",
    "tf.get_logger().setLevel('INFO')\n",
    "tf.autograph.set_verbosity(0)\n",
    "\n",
    "tf.get_logger().setLevel(logging.ERROR)\n",
    "tf.get_logger().setLevel('ERROR')\n",
    "tf.compat.v1.logging.set_verbosity(tf.compat.v1.logging.ERROR)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "799eb52f",
   "metadata": {
    "id": "MvjjNtBV_jBQ",
    "papermill": {
     "duration": 0.005514,
     "end_time": "2022-11-15T19:33:42.677177",
     "exception": false,
     "start_time": "2022-11-15T19:33:42.671663",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# **DATASET**\n",
    "\n",
    "\n",
    "RGB images of 96x96 pixels.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "dfa1d2ac",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-11-15T19:33:42.689744Z",
     "iopub.status.busy": "2022-11-15T19:33:42.689491Z",
     "iopub.status.idle": "2022-11-15T19:33:42.694050Z",
     "shell.execute_reply": "2022-11-15T19:33:42.693182Z"
    },
    "papermill": {
     "duration": 0.013042,
     "end_time": "2022-11-15T19:33:42.696073",
     "exception": false,
     "start_time": "2022-11-15T19:33:42.683031",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def genAugmLayer(seed): \n",
    "    return tf.keras.Sequential([\n",
    "     tfkl.RandomFlip(\"horizontal_and_vertical\"),\n",
    "     tfkl.RandomRotation(0.5),\n",
    "     tfkl.RandomZoom( height_factor=(-0.3, 0.3), width_factor=(-0.3, 0.3),fill_mode = 'reflect',seed=seed) ])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "798ae878",
   "metadata": {
    "id": "mgetvb8V3FDm",
    "papermill": {
     "duration": 0.005551,
     "end_time": "2022-11-15T19:33:42.707595",
     "exception": false,
     "start_time": "2022-11-15T19:33:42.702044",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "***SELECTIVE AUGMENTATION DONE TWO TIMES TO FEED THE NN MORE SAMPLES FROM CLASSES 1 AND 6 (~ 400 PER CLASS)***\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "c8057635",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-11-15T19:33:42.720109Z",
     "iopub.status.busy": "2022-11-15T19:33:42.719839Z",
     "iopub.status.idle": "2022-11-15T19:33:43.595760Z",
     "shell.execute_reply": "2022-11-15T19:33:43.594806Z"
    },
    "papermill": {
     "duration": 0.884963,
     "end_time": "2022-11-15T19:33:43.598351",
     "exception": false,
     "start_time": "2022-11-15T19:33:42.713388",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 414 files belonging to 8 classes.\n",
      "Using 332 files for training.\n"
     ]
    }
   ],
   "source": [
    "imbalClasses = tf.keras.utils.image_dataset_from_directory( \n",
    "    '../input/correctedbal',\n",
    "    labels='inferred',\n",
    "    label_mode='categorical',\n",
    "    class_names=None,\n",
    "    color_mode='rgb',\n",
    "    batch_size=32,\n",
    "    image_size=(96,96),\n",
    "    subset = 'training',\n",
    "    validation_split=0.2,\n",
    "    seed=seed,\n",
    ")\n",
    "\n",
    "data_augmentation1 = genAugmLayer(17) \n",
    "dataAugmImbal = imbalClasses.map(lambda x , y : (data_augmentation1(x), y)) \n",
    "data_augmentation2 = genAugmLayer(271) \n",
    "dataAugmImbal = dataAugmImbal.concatenate(imbalClasses.map(lambda x , y : (data_augmentation2(x), y))) #double samples of classes 1 and 6 "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d6b9ebd7",
   "metadata": {
    "papermill": {
     "duration": 0.005692,
     "end_time": "2022-11-15T19:33:43.610069",
     "exception": false,
     "start_time": "2022-11-15T19:33:43.604377",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "***UNIVERSAL AUGMENTATION***"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "deb6a61d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-11-15T19:33:43.623451Z",
     "iopub.status.busy": "2022-11-15T19:33:43.622579Z",
     "iopub.status.idle": "2022-11-15T19:33:50.522711Z",
     "shell.execute_reply": "2022-11-15T19:33:50.520925Z"
    },
    "papermill": {
     "duration": 6.90916,
     "end_time": "2022-11-15T19:33:50.524935",
     "exception": false,
     "start_time": "2022-11-15T19:33:43.615775",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 3542 files belonging to 8 classes.\n",
      "Using 2834 files for training.\n",
      "Found 3542 files belonging to 8 classes.\n",
      "Using 708 files for validation.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-11-15 19:33:44.219950: I tensorflow/compiler/mlir/mlir_graph_optimization_pass.cc:185] None of the MLIR Optimization Passes are enabled (registered 2)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{0: 2.393581081081081, 1: 0.8296252927400468, 2: 0.8923173803526449, 3: 0.843452380952381, 4: 0.8162442396313364, 5: 2.012784090909091, 6: 0.8032879818594104, 7: 0.9060102301790282}\n"
     ]
    }
   ],
   "source": [
    "datapath = '../input/annhw1/training_data_final'\n",
    "training_set = tf.keras.utils.image_dataset_from_directory(\n",
    "    datapath,\n",
    "    labels='inferred',\n",
    "    label_mode='categorical',\n",
    "    class_names=None,\n",
    "    color_mode='rgb',\n",
    "    batch_size=32, \n",
    "    image_size=(96,96),\n",
    "    shuffle=True,\n",
    "    seed=seed,\n",
    "    validation_split=0.2,\n",
    "    subset='training'\n",
    ")\n",
    "\n",
    "validation_set = tf.keras.utils.image_dataset_from_directory(\n",
    "    datapath,\n",
    "    labels='inferred',\n",
    "    label_mode='categorical',\n",
    "    class_names=None,\n",
    "    color_mode='rgb',\n",
    "    batch_size=32, \n",
    "    image_size=(96,96),\n",
    "    shuffle=True,\n",
    "    seed=seed,\n",
    "    validation_split=0.2,\n",
    "    subset='validation'\n",
    ")\n",
    "\n",
    "\n",
    "#-------------WEIGHTS ----------------------------------\n",
    "labels = np.concatenate([y for x, y in training_set], axis=0) \n",
    "class_weights = class_weight.compute_class_weight(class_weight='balanced' ,classes = [i for i in range(8)] , y = np.argmax(labels, axis=-1) ) \n",
    "weights = dict(enumerate(class_weights)) \n",
    "print(weights)\n",
    "#-------------------------------------------------------\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "815be4b3",
   "metadata": {
    "papermill": {
     "duration": 0.006031,
     "end_time": "2022-11-15T19:33:50.537910",
     "exception": false,
     "start_time": "2022-11-15T19:33:50.531879",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "***RESCALING***"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "037b53a9",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-11-15T19:33:50.551031Z",
     "iopub.status.busy": "2022-11-15T19:33:50.550725Z",
     "iopub.status.idle": "2022-11-15T19:33:50.901320Z",
     "shell.execute_reply": "2022-11-15T19:33:50.900399Z"
    },
    "papermill": {
     "duration": 0.360031,
     "end_time": "2022-11-15T19:33:50.903947",
     "exception": false,
     "start_time": "2022-11-15T19:33:50.543916",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "data_augmentation = genAugmLayer(89)\n",
    "training_set = training_set.concatenate(training_set.map(lambda x , y : (data_augmentation(x), y)) ) #total augm\n",
    "training_set = training_set.concatenate(dataAugmImbal)\n",
    "#training_set = training_set.concatenate(imbalClasses) #OVERSAMPLING classes 1 and 6 "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "4c9dc301",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-11-15T19:33:50.917972Z",
     "iopub.status.busy": "2022-11-15T19:33:50.917664Z",
     "iopub.status.idle": "2022-11-15T19:33:50.925634Z",
     "shell.execute_reply": "2022-11-15T19:33:50.924778Z"
    },
    "papermill": {
     "duration": 0.017537,
     "end_time": "2022-11-15T19:33:50.927615",
     "exception": false,
     "start_time": "2022-11-15T19:33:50.910078",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "AUTOTUNE = tf.data.AUTOTUNE\n",
    "training_set = training_set.cache().shuffle(1000).prefetch(buffer_size=AUTOTUNE)\n",
    "validation_set = validation_set.cache().prefetch(buffer_size=AUTOTUNE)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d9fc171b",
   "metadata": {
    "id": "kprNnVJ7I9HU",
    "papermill": {
     "duration": 0.006097,
     "end_time": "2022-11-15T19:33:50.940060",
     "exception": false,
     "start_time": "2022-11-15T19:33:50.933963",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# **Transfer Learning**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "cca6434c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-11-15T19:33:50.953629Z",
     "iopub.status.busy": "2022-11-15T19:33:50.953367Z",
     "iopub.status.idle": "2022-11-15T19:33:50.957888Z",
     "shell.execute_reply": "2022-11-15T19:33:50.956867Z"
    },
    "id": "7YcxBMJhl4EM",
    "papermill": {
     "duration": 0.01406,
     "end_time": "2022-11-15T19:33:50.960201",
     "exception": false,
     "start_time": "2022-11-15T19:33:50.946141",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "initial_time = datetime.datetime.now()\n",
    "input_shape = (96, 96, 3)\n",
    "epochs = 90 #to be defined "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "66be2f36",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-11-15T19:33:50.974363Z",
     "iopub.status.busy": "2022-11-15T19:33:50.974087Z",
     "iopub.status.idle": "2022-11-15T19:33:51.019627Z",
     "shell.execute_reply": "2022-11-15T19:33:51.018830Z"
    },
    "papermill": {
     "duration": 0.055245,
     "end_time": "2022-11-15T19:33:51.021650",
     "exception": false,
     "start_time": "2022-11-15T19:33:50.966405",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "training_set = training_set.map(lambda x,y: (tf.keras.applications.inception_v3.preprocess_input(x), y))\n",
    "validation_set = validation_set.map(lambda x,y: (tf.keras.applications.inception_v3.preprocess_input(x),y))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "66447b55",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-11-15T19:33:51.035524Z",
     "iopub.status.busy": "2022-11-15T19:33:51.035255Z",
     "iopub.status.idle": "2022-11-15T19:33:57.542740Z",
     "shell.execute_reply": "2022-11-15T19:33:57.541752Z"
    },
    "papermill": {
     "duration": 6.517458,
     "end_time": "2022-11-15T19:33:57.545320",
     "exception": false,
     "start_time": "2022-11-15T19:33:51.027862",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/inception_v3/inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
      "87916544/87910968 [==============================] - 1s 0us/step\n",
      "87924736/87910968 [==============================] - 1s 0us/step\n"
     ]
    }
   ],
   "source": [
    "#ARCHITECTURE\n",
    "with mirrored_strategy.scope():\n",
    "    supernet = tf.keras.applications.inception_v3.InceptionV3( include_top=False, weights='imagenet')\n",
    "    supernet.trainable = True\n",
    "    input_layer = tfkl.Input(input_shape)\n",
    "    x = tfkl.Resizing(height = 299, width = 299, interpolation = 'bicubic' )(input_layer)\n",
    "    x = supernet(x)\n",
    "    x = tfkl.GlobalAveragePooling2D()(x) \n",
    "    x = tfkl.Dropout(rate=0.3)(x)\n",
    "    x = tfkl.Dense(units=128, name='Dense3', kernel_initializer=tfk.initializers.HeUniform(seed),  activation='relu')(x)\n",
    "    x = tfkl.Dropout(rate=0.3)(x)\n",
    "    output_layer = tfkl.Dense(units=8, activation='softmax', kernel_initializer=tfk.initializers.GlorotUniform(seed), name='output_layer')(x)\n",
    "        \n",
    "    model = tfk.Model(inputs=input_layer, outputs=output_layer, name='model')\n",
    "    model.compile(loss=tfk.losses.CategoricalCrossentropy(), optimizer=tf.keras.optimizers.Adam(), metrics='accuracy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "4e1da02f",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-11-15T19:33:57.561836Z",
     "iopub.status.busy": "2022-11-15T19:33:57.561535Z",
     "iopub.status.idle": "2022-11-15T19:33:57.567497Z",
     "shell.execute_reply": "2022-11-15T19:33:57.566555Z"
    },
    "id": "6qbixQiprQl_",
    "papermill": {
     "duration": 0.01693,
     "end_time": "2022-11-15T19:33:57.569636",
     "exception": false,
     "start_time": "2022-11-15T19:33:57.552706",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "from datetime import datetime\n",
    "def create_folders_and_callbacks(model_name):\n",
    "    '''\n",
    "    exps_dir = os.path.join('experiments')\n",
    "    if not os.path.exists(exps_dir):\n",
    "        os.makedirs(exps_dir)\n",
    "    now = datetime.now().strftime('%b%d_%H-%M-%S')\n",
    "    exp_dir = os.path.join(exps_dir, model_name + '_' + str(now))\n",
    "    if not os.path.exists(exp_dir):\n",
    "        os.makedirs(exp_dir)\n",
    "    '''\n",
    "    callbacks = []\n",
    "    \n",
    "    '''\n",
    "    # Model checkpoint\n",
    "      # ----------------\n",
    "    ckpt_dir = os.path.join(exp_dir, 'ckpts')\n",
    "    if not os.path.exists(ckpt_dir):\n",
    "        os.makedirs(ckpt_dir)\n",
    "          \n",
    "    ckpt_callback = tf.keras.callbacks.ModelCheckpoint(filepath=os.path.join(ckpt_dir, 'cp.ckpt'), save_weights_only=True, save_best_only=True) #play here \n",
    "      \n",
    "    callbacks.append(ckpt_callback)\n",
    "    '''\n",
    "    # Early Stopping\n",
    "    # --------------\n",
    "    es_callback = tf.keras.callbacks.EarlyStopping(monitor='val_accuracy', patience=20, restore_best_weights=True) #tuning point\n",
    "    callbacks.append(es_callback)\n",
    "    return callbacks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "a1d25a38",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-11-15T19:33:57.584781Z",
     "iopub.status.busy": "2022-11-15T19:33:57.584518Z",
     "iopub.status.idle": "2022-11-15T19:33:57.604765Z",
     "shell.execute_reply": "2022-11-15T19:33:57.603693Z"
    },
    "id": "NwsjIp_crOKq",
    "outputId": "32ffd07a-e461-464f-c4fa-a6bcbc307cbf",
    "papermill": {
     "duration": 0.030304,
     "end_time": "2022-11-15T19:33:57.607135",
     "exception": false,
     "start_time": "2022-11-15T19:33:57.576831",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_2 (InputLayer)         [(None, 96, 96, 3)]       0         \n",
      "_________________________________________________________________\n",
      "resizing (Resizing)          (None, 299, 299, 3)       0         \n",
      "_________________________________________________________________\n",
      "inception_v3 (Functional)    (None, None, None, 2048)  21802784  \n",
      "_________________________________________________________________\n",
      "global_average_pooling2d (Gl (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "Dense3 (Dense)               (None, 128)               262272    \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "output_layer (Dense)         (None, 8)                 1032      \n",
      "=================================================================\n",
      "Total params: 22,066,088\n",
      "Trainable params: 22,031,656\n",
      "Non-trainable params: 34,432\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "301c004d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-11-15T19:33:57.622507Z",
     "iopub.status.busy": "2022-11-15T19:33:57.622228Z",
     "iopub.status.idle": "2022-11-15T19:33:57.626028Z",
     "shell.execute_reply": "2022-11-15T19:33:57.625061Z"
    },
    "papermill": {
     "duration": 0.013742,
     "end_time": "2022-11-15T19:33:57.628166",
     "exception": false,
     "start_time": "2022-11-15T19:33:57.614424",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "#WEIGHTS RESTORING WHEN CRASHES \n",
    "#model.restore(\"./experiments/myModel_TESTING_Nov14_10-17-04\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "03793699",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-11-15T19:33:57.643614Z",
     "iopub.status.busy": "2022-11-15T19:33:57.643359Z",
     "iopub.status.idle": "2022-11-15T19:33:57.672537Z",
     "shell.execute_reply": "2022-11-15T19:33:57.671413Z"
    },
    "papermill": {
     "duration": 0.039085,
     "end_time": "2022-11-15T19:33:57.674526",
     "exception": false,
     "start_time": "2022-11-15T19:33:57.635441",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_2 (InputLayer)         [(None, 96, 96, 3)]       0         \n",
      "_________________________________________________________________\n",
      "resizing (Resizing)          (None, 299, 299, 3)       0         \n",
      "_________________________________________________________________\n",
      "inception_v3 (Functional)    (None, None, None, 2048)  21802784  \n",
      "_________________________________________________________________\n",
      "global_average_pooling2d (Gl (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "Dense3 (Dense)               (None, 128)               262272    \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "output_layer (Dense)         (None, 8)                 1032      \n",
      "=================================================================\n",
      "Total params: 22,066,088\n",
      "Trainable params: 830,264\n",
      "Non-trainable params: 21,235,824\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "\n",
    "for i, layer in enumerate(model.get_layer('inception_v3').layers[:-1]):\n",
    "    layer.trainable=False\n",
    "\n",
    "for i, layer in enumerate(model.get_layer('inception_v3').layers[:15]):\n",
    "    layer.trainable=True\n",
    "    \n",
    "for i, layer in enumerate(model.get_layer('inception_v3').layers[-15:-1]):\n",
    "    layer.trainable = True\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "787b9817",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-11-15T19:33:57.690408Z",
     "iopub.status.busy": "2022-11-15T19:33:57.689242Z",
     "iopub.status.idle": "2022-11-15T22:03:41.732294Z",
     "shell.execute_reply": "2022-11-15T22:03:41.731210Z"
    },
    "id": "pXAHAuoSz-_S",
    "outputId": "05f554a5-f149-41ae-c9ac-68221b6f2345",
    "papermill": {
     "duration": 8984.496277,
     "end_time": "2022-11-15T22:03:42.177501",
     "exception": false,
     "start_time": "2022-11-15T19:33:57.681224",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-11-15 19:33:57.783018: W tensorflow/core/grappler/optimizers/data/auto_shard.cc:695] AUTO sharding policy will apply DATA sharding policy as it failed to apply FILE sharding policy because of the following reason: Found an unshardable source dataset: name: \"TensorSliceDataset/_1\"\n",
      "op: \"TensorSliceDataset\"\n",
      "input: \"Placeholder/_0\"\n",
      "attr {\n",
      "  key: \"Toutput_types\"\n",
      "  value {\n",
      "    list {\n",
      "      type: DT_STRING\n",
      "    }\n",
      "  }\n",
      "}\n",
      "attr {\n",
      "  key: \"output_shapes\"\n",
      "  value {\n",
      "    list {\n",
      "      shape {\n",
      "      }\n",
      "    }\n",
      "  }\n",
      "}\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/90\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-11-15 19:34:56.152733: I tensorflow/core/kernels/data/shuffle_dataset_op.cc:175] Filling up shuffle buffer (this may take a while): 197 of 1000\n",
      "2022-11-15 19:34:56.293514: I tensorflow/core/kernels/data/shuffle_dataset_op.cc:228] Shuffle buffer filled.\n",
      "2022-11-15 19:34:57.719358: I tensorflow/stream_executor/cuda/cuda_dnn.cc:369] Loaded cuDNN version 8005\n",
      "2022-11-15 19:34:58.150625: I tensorflow/stream_executor/cuda/cuda_dnn.cc:369] Loaded cuDNN version 8005\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "200/200 [==============================] - ETA: 0s - loss: 1.5953 - accuracy: 0.4090"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-11-15 19:36:50.706215: W tensorflow/core/grappler/optimizers/data/auto_shard.cc:695] AUTO sharding policy will apply DATA sharding policy as it failed to apply FILE sharding policy because of the following reason: Found an unshardable source dataset: name: \"TensorSliceDataset/_1\"\n",
      "op: \"TensorSliceDataset\"\n",
      "input: \"Placeholder/_0\"\n",
      "attr {\n",
      "  key: \"Toutput_types\"\n",
      "  value {\n",
      "    list {\n",
      "      type: DT_STRING\n",
      "    }\n",
      "  }\n",
      "}\n",
      "attr {\n",
      "  key: \"output_shapes\"\n",
      "  value {\n",
      "    list {\n",
      "      shape {\n",
      "      }\n",
      "    }\n",
      "  }\n",
      "}\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "200/200 [==============================] - 182s 552ms/step - loss: 1.5953 - accuracy: 0.4090 - val_loss: 4.3521 - val_accuracy: 0.2415\n",
      "Epoch 2/90\n",
      "200/200 [==============================] - 99s 492ms/step - loss: 1.3395 - accuracy: 0.5190 - val_loss: 4.1505 - val_accuracy: 0.3517\n",
      "Epoch 3/90\n",
      "200/200 [==============================] - 99s 493ms/step - loss: 1.1323 - accuracy: 0.6008 - val_loss: 2.0894 - val_accuracy: 0.4831\n",
      "Epoch 4/90\n",
      "200/200 [==============================] - 98s 489ms/step - loss: 0.9529 - accuracy: 0.6691 - val_loss: 5.9032 - val_accuracy: 0.1723\n",
      "Epoch 5/90\n",
      "200/200 [==============================] - 98s 488ms/step - loss: 0.8225 - accuracy: 0.7187 - val_loss: 3.6163 - val_accuracy: 0.3616\n",
      "Epoch 6/90\n",
      "200/200 [==============================] - 97s 487ms/step - loss: 0.7084 - accuracy: 0.7541 - val_loss: 2.4174 - val_accuracy: 0.3489\n",
      "Epoch 7/90\n",
      "200/200 [==============================] - 99s 493ms/step - loss: 0.6095 - accuracy: 0.7858 - val_loss: 2.1719 - val_accuracy: 0.5254\n",
      "Epoch 8/90\n",
      "200/200 [==============================] - 98s 490ms/step - loss: 0.5100 - accuracy: 0.8249 - val_loss: 1.4441 - val_accuracy: 0.5862\n",
      "Epoch 9/90\n",
      "200/200 [==============================] - 97s 487ms/step - loss: 0.4451 - accuracy: 0.8468 - val_loss: 2.2650 - val_accuracy: 0.4944\n",
      "Epoch 10/90\n",
      "200/200 [==============================] - 98s 490ms/step - loss: 0.3935 - accuracy: 0.8669 - val_loss: 1.1774 - val_accuracy: 0.6766\n",
      "Epoch 11/90\n",
      "200/200 [==============================] - 98s 491ms/step - loss: 0.2814 - accuracy: 0.9095 - val_loss: 2.3018 - val_accuracy: 0.6398\n",
      "Epoch 12/90\n",
      "200/200 [==============================] - 98s 488ms/step - loss: 0.2878 - accuracy: 0.8989 - val_loss: 1.5279 - val_accuracy: 0.5706\n",
      "Epoch 13/90\n",
      "200/200 [==============================] - 98s 488ms/step - loss: 0.2647 - accuracy: 0.9086 - val_loss: 1.4807 - val_accuracy: 0.7006\n",
      "Epoch 14/90\n",
      "200/200 [==============================] - 98s 488ms/step - loss: 0.1941 - accuracy: 0.9365 - val_loss: 2.0944 - val_accuracy: 0.5621\n",
      "Epoch 15/90\n",
      "200/200 [==============================] - 97s 487ms/step - loss: 0.2016 - accuracy: 0.9334 - val_loss: 1.5612 - val_accuracy: 0.6398\n",
      "Epoch 16/90\n",
      "200/200 [==============================] - 97s 486ms/step - loss: 0.1983 - accuracy: 0.9367 - val_loss: 4.1282 - val_accuracy: 0.4619\n",
      "Epoch 17/90\n",
      "200/200 [==============================] - 97s 485ms/step - loss: 0.1645 - accuracy: 0.9480 - val_loss: 1.3855 - val_accuracy: 0.6907\n",
      "Epoch 18/90\n",
      "200/200 [==============================] - 97s 486ms/step - loss: 0.1208 - accuracy: 0.9604 - val_loss: 1.7784 - val_accuracy: 0.6455\n",
      "Epoch 19/90\n",
      "200/200 [==============================] - 100s 498ms/step - loss: 0.1176 - accuracy: 0.9611 - val_loss: 1.1471 - val_accuracy: 0.7486\n",
      "Epoch 20/90\n",
      "200/200 [==============================] - 99s 494ms/step - loss: 0.1742 - accuracy: 0.9452 - val_loss: 3.6301 - val_accuracy: 0.4280\n",
      "Epoch 21/90\n",
      "200/200 [==============================] - 98s 491ms/step - loss: 0.1074 - accuracy: 0.9654 - val_loss: 2.7621 - val_accuracy: 0.5452\n",
      "Epoch 22/90\n",
      "200/200 [==============================] - 99s 494ms/step - loss: 0.1297 - accuracy: 0.9593 - val_loss: 3.0132 - val_accuracy: 0.4972\n",
      "Epoch 23/90\n",
      "200/200 [==============================] - 99s 495ms/step - loss: 0.1152 - accuracy: 0.9632 - val_loss: 1.4869 - val_accuracy: 0.7189\n",
      "Epoch 24/90\n",
      "200/200 [==============================] - 98s 491ms/step - loss: 0.1070 - accuracy: 0.9657 - val_loss: 2.3717 - val_accuracy: 0.6271\n",
      "Epoch 25/90\n",
      "200/200 [==============================] - 99s 492ms/step - loss: 0.0823 - accuracy: 0.9728 - val_loss: 1.3381 - val_accuracy: 0.6638\n",
      "Epoch 26/90\n",
      "200/200 [==============================] - 99s 492ms/step - loss: 0.1349 - accuracy: 0.9578 - val_loss: 1.6648 - val_accuracy: 0.6554\n",
      "Epoch 27/90\n",
      "200/200 [==============================] - 98s 491ms/step - loss: 0.1030 - accuracy: 0.9686 - val_loss: 5.7785 - val_accuracy: 0.4661\n",
      "Epoch 28/90\n",
      "200/200 [==============================] - 98s 491ms/step - loss: 0.0686 - accuracy: 0.9793 - val_loss: 2.5009 - val_accuracy: 0.5650\n",
      "Epoch 29/90\n",
      "200/200 [==============================] - 98s 491ms/step - loss: 0.0878 - accuracy: 0.9730 - val_loss: 1.4240 - val_accuracy: 0.7232\n",
      "Epoch 30/90\n",
      "200/200 [==============================] - 99s 495ms/step - loss: 0.0887 - accuracy: 0.9725 - val_loss: 2.5752 - val_accuracy: 0.5763\n",
      "Epoch 31/90\n",
      "200/200 [==============================] - 98s 491ms/step - loss: 0.0615 - accuracy: 0.9801 - val_loss: 1.8706 - val_accuracy: 0.7062\n",
      "Epoch 32/90\n",
      "200/200 [==============================] - 99s 493ms/step - loss: 0.0630 - accuracy: 0.9806 - val_loss: 1.6128 - val_accuracy: 0.7359\n",
      "Epoch 33/90\n",
      "200/200 [==============================] - 99s 496ms/step - loss: 0.1195 - accuracy: 0.9626 - val_loss: 1.8123 - val_accuracy: 0.6596\n",
      "Epoch 34/90\n",
      "200/200 [==============================] - 98s 493ms/step - loss: 0.0729 - accuracy: 0.9788 - val_loss: 1.7514 - val_accuracy: 0.6653\n",
      "Epoch 35/90\n",
      "200/200 [==============================] - 99s 493ms/step - loss: 0.0664 - accuracy: 0.9780 - val_loss: 2.1919 - val_accuracy: 0.6554\n",
      "Epoch 36/90\n",
      "200/200 [==============================] - 99s 495ms/step - loss: 0.0738 - accuracy: 0.9771 - val_loss: 1.2095 - val_accuracy: 0.7910\n",
      "Epoch 37/90\n",
      "200/200 [==============================] - 99s 494ms/step - loss: 0.1148 - accuracy: 0.9621 - val_loss: 1.4129 - val_accuracy: 0.7316\n",
      "Epoch 38/90\n",
      "200/200 [==============================] - 98s 490ms/step - loss: 0.0454 - accuracy: 0.9853 - val_loss: 2.5906 - val_accuracy: 0.6907\n",
      "Epoch 39/90\n",
      "200/200 [==============================] - 99s 494ms/step - loss: 0.1499 - accuracy: 0.9550 - val_loss: 1.5471 - val_accuracy: 0.6864\n",
      "Epoch 40/90\n",
      "200/200 [==============================] - 99s 493ms/step - loss: 0.0510 - accuracy: 0.9855 - val_loss: 1.3548 - val_accuracy: 0.7514\n",
      "Epoch 41/90\n",
      "200/200 [==============================] - 98s 492ms/step - loss: 0.0451 - accuracy: 0.9859 - val_loss: 2.1206 - val_accuracy: 0.7020\n",
      "Epoch 42/90\n",
      "200/200 [==============================] - 99s 495ms/step - loss: 0.0215 - accuracy: 0.9938 - val_loss: 1.3696 - val_accuracy: 0.7740\n",
      "Epoch 43/90\n",
      "200/200 [==============================] - 99s 495ms/step - loss: 0.0425 - accuracy: 0.9877 - val_loss: 1.4346 - val_accuracy: 0.7232\n",
      "Epoch 44/90\n",
      "200/200 [==============================] - 98s 491ms/step - loss: 0.1406 - accuracy: 0.9623 - val_loss: 2.2034 - val_accuracy: 0.5254\n",
      "Epoch 45/90\n",
      "200/200 [==============================] - 98s 492ms/step - loss: 0.0976 - accuracy: 0.9684 - val_loss: 3.5143 - val_accuracy: 0.5480\n",
      "Epoch 46/90\n",
      "200/200 [==============================] - 98s 492ms/step - loss: 0.0486 - accuracy: 0.9844 - val_loss: 2.1000 - val_accuracy: 0.6624\n",
      "Epoch 47/90\n",
      "200/200 [==============================] - 99s 494ms/step - loss: 0.0380 - accuracy: 0.9886 - val_loss: 2.2163 - val_accuracy: 0.6624\n",
      "Epoch 48/90\n",
      "200/200 [==============================] - 99s 493ms/step - loss: 0.0420 - accuracy: 0.9886 - val_loss: 1.7194 - val_accuracy: 0.6723\n",
      "Epoch 49/90\n",
      "200/200 [==============================] - 98s 491ms/step - loss: 0.0491 - accuracy: 0.9861 - val_loss: 1.8914 - val_accuracy: 0.6695\n",
      "Epoch 50/90\n",
      "200/200 [==============================] - 98s 492ms/step - loss: 0.0394 - accuracy: 0.9875 - val_loss: 2.4771 - val_accuracy: 0.6836\n",
      "Epoch 51/90\n",
      "200/200 [==============================] - 99s 495ms/step - loss: 0.0391 - accuracy: 0.9875 - val_loss: 1.6255 - val_accuracy: 0.7542\n",
      "Epoch 52/90\n",
      "200/200 [==============================] - 99s 493ms/step - loss: 0.0922 - accuracy: 0.9705 - val_loss: 8.2242 - val_accuracy: 0.3545\n",
      "Epoch 53/90\n",
      "200/200 [==============================] - 98s 491ms/step - loss: 0.0829 - accuracy: 0.9760 - val_loss: 1.7547 - val_accuracy: 0.7189\n",
      "Epoch 54/90\n",
      "200/200 [==============================] - 98s 491ms/step - loss: 0.0279 - accuracy: 0.9912 - val_loss: 1.2422 - val_accuracy: 0.8206\n",
      "Epoch 55/90\n",
      "200/200 [==============================] - 99s 495ms/step - loss: 0.0266 - accuracy: 0.9912 - val_loss: 2.7870 - val_accuracy: 0.6017\n",
      "Epoch 56/90\n",
      "200/200 [==============================] - 99s 495ms/step - loss: 0.0320 - accuracy: 0.9918 - val_loss: 1.8396 - val_accuracy: 0.7331\n",
      "Epoch 57/90\n",
      "200/200 [==============================] - 99s 493ms/step - loss: 0.0507 - accuracy: 0.9848 - val_loss: 2.0835 - val_accuracy: 0.6893\n",
      "Epoch 58/90\n",
      "200/200 [==============================] - 98s 492ms/step - loss: 0.1153 - accuracy: 0.9653 - val_loss: 2.8547 - val_accuracy: 0.5466\n",
      "Epoch 59/90\n",
      "200/200 [==============================] - 98s 491ms/step - loss: 0.0413 - accuracy: 0.9874 - val_loss: 1.7338 - val_accuracy: 0.7161\n",
      "Epoch 60/90\n",
      "200/200 [==============================] - 98s 492ms/step - loss: 0.0450 - accuracy: 0.9863 - val_loss: 2.3029 - val_accuracy: 0.7006\n",
      "Epoch 61/90\n",
      "200/200 [==============================] - 98s 491ms/step - loss: 0.0558 - accuracy: 0.9828 - val_loss: 2.3287 - val_accuracy: 0.6879\n",
      "Epoch 62/90\n",
      "200/200 [==============================] - 98s 490ms/step - loss: 0.0388 - accuracy: 0.9886 - val_loss: 4.0512 - val_accuracy: 0.5621\n",
      "Epoch 63/90\n",
      "200/200 [==============================] - 99s 495ms/step - loss: 0.0251 - accuracy: 0.9913 - val_loss: 1.3445 - val_accuracy: 0.7585\n",
      "Epoch 64/90\n",
      "200/200 [==============================] - 99s 494ms/step - loss: 0.0089 - accuracy: 0.9984 - val_loss: 4.7664 - val_accuracy: 0.6088\n",
      "Epoch 65/90\n",
      "200/200 [==============================] - 99s 494ms/step - loss: 0.0931 - accuracy: 0.9698 - val_loss: 1.7156 - val_accuracy: 0.6850\n",
      "Epoch 66/90\n",
      "200/200 [==============================] - 99s 494ms/step - loss: 0.0654 - accuracy: 0.9806 - val_loss: 1.4097 - val_accuracy: 0.7726\n",
      "Epoch 67/90\n",
      "200/200 [==============================] - 99s 493ms/step - loss: 0.0045 - accuracy: 0.9992 - val_loss: 1.1343 - val_accuracy: 0.8291\n",
      "Epoch 68/90\n",
      "200/200 [==============================] - 99s 493ms/step - loss: 0.0369 - accuracy: 0.9896 - val_loss: 2.2210 - val_accuracy: 0.7161\n",
      "Epoch 69/90\n",
      "200/200 [==============================] - 99s 495ms/step - loss: 0.0447 - accuracy: 0.9875 - val_loss: 1.5998 - val_accuracy: 0.7444\n",
      "Epoch 70/90\n",
      "200/200 [==============================] - 99s 495ms/step - loss: 0.0406 - accuracy: 0.9891 - val_loss: 1.3930 - val_accuracy: 0.7698\n",
      "Epoch 71/90\n",
      "200/200 [==============================] - 98s 491ms/step - loss: 0.0550 - accuracy: 0.9828 - val_loss: 1.6596 - val_accuracy: 0.7274\n",
      "Epoch 72/90\n",
      "200/200 [==============================] - 99s 496ms/step - loss: 0.0251 - accuracy: 0.9919 - val_loss: 1.5620 - val_accuracy: 0.7571\n",
      "Epoch 73/90\n",
      "200/200 [==============================] - 98s 492ms/step - loss: 0.0150 - accuracy: 0.9951 - val_loss: 2.6941 - val_accuracy: 0.6540\n",
      "Epoch 74/90\n",
      "200/200 [==============================] - 99s 493ms/step - loss: 0.0704 - accuracy: 0.9799 - val_loss: 2.4403 - val_accuracy: 0.6568\n",
      "Epoch 75/90\n",
      "200/200 [==============================] - 99s 493ms/step - loss: 0.0413 - accuracy: 0.9866 - val_loss: 1.6378 - val_accuracy: 0.7274\n",
      "Epoch 76/90\n",
      "200/200 [==============================] - 98s 491ms/step - loss: 0.0407 - accuracy: 0.9888 - val_loss: 5.9630 - val_accuracy: 0.4590\n",
      "Epoch 77/90\n",
      "200/200 [==============================] - 98s 491ms/step - loss: 0.0354 - accuracy: 0.9888 - val_loss: 1.9166 - val_accuracy: 0.6723\n",
      "Epoch 78/90\n",
      "200/200 [==============================] - 99s 495ms/step - loss: 0.0252 - accuracy: 0.9924 - val_loss: 1.8299 - val_accuracy: 0.7105\n",
      "Epoch 79/90\n",
      "200/200 [==============================] - 99s 495ms/step - loss: 0.0289 - accuracy: 0.9915 - val_loss: 2.1137 - val_accuracy: 0.6935\n",
      "Epoch 80/90\n",
      "200/200 [==============================] - 99s 494ms/step - loss: 0.0280 - accuracy: 0.9912 - val_loss: 1.7855 - val_accuracy: 0.7613\n",
      "Epoch 81/90\n",
      "200/200 [==============================] - 99s 494ms/step - loss: 0.0266 - accuracy: 0.9924 - val_loss: 2.0431 - val_accuracy: 0.7316\n",
      "Epoch 82/90\n",
      "200/200 [==============================] - 99s 495ms/step - loss: 0.1202 - accuracy: 0.9673 - val_loss: 1.4575 - val_accuracy: 0.7726\n",
      "Epoch 83/90\n",
      "200/200 [==============================] - 99s 494ms/step - loss: 0.0111 - accuracy: 0.9961 - val_loss: 1.4713 - val_accuracy: 0.7712\n",
      "Epoch 84/90\n",
      "200/200 [==============================] - 99s 496ms/step - loss: 0.0041 - accuracy: 0.9989 - val_loss: 1.5575 - val_accuracy: 0.7712\n",
      "Epoch 85/90\n",
      "200/200 [==============================] - 99s 494ms/step - loss: 0.0560 - accuracy: 0.9844 - val_loss: 1.6511 - val_accuracy: 0.7288\n",
      "Epoch 86/90\n",
      "200/200 [==============================] - 99s 494ms/step - loss: 0.0411 - accuracy: 0.9880 - val_loss: 2.4652 - val_accuracy: 0.6384\n",
      "Epoch 87/90\n",
      "200/200 [==============================] - 99s 493ms/step - loss: 0.0198 - accuracy: 0.9946 - val_loss: 1.4539 - val_accuracy: 0.7684\n"
     ]
    }
   ],
   "source": [
    "callbacks = create_folders_and_callbacks(model_name='myModel_TESTING')\n",
    "\n",
    "history = model.fit(\n",
    "    x = training_set,\n",
    "    epochs = epochs,\n",
    "    validation_data = validation_set, \n",
    "    callbacks = callbacks,\n",
    "    batch_size = 256\n",
    "    #,class_weight = weights \n",
    ").history"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "70c26b79",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-11-15T22:03:44.239692Z",
     "iopub.status.busy": "2022-11-15T22:03:44.239341Z",
     "iopub.status.idle": "2022-11-15T22:03:44.244939Z",
     "shell.execute_reply": "2022-11-15T22:03:44.243762Z"
    },
    "id": "7YDDlgkUWMpO",
    "papermill": {
     "duration": 1.041885,
     "end_time": "2022-11-15T22:03:44.247801",
     "exception": false,
     "start_time": "2022-11-15T22:03:43.205916",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training took 2:29:53.286156\n"
     ]
    }
   ],
   "source": [
    "final_time = datetime.now()\n",
    "print(\"Training took %s\" % (final_time - initial_time))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "439150ee",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-11-15T22:03:46.557895Z",
     "iopub.status.busy": "2022-11-15T22:03:46.557489Z",
     "iopub.status.idle": "2022-11-15T22:03:47.149529Z",
     "shell.execute_reply": "2022-11-15T22:03:47.148546Z"
    },
    "id": "3fziDMO1NTG2",
    "papermill": {
     "duration": 1.92515,
     "end_time": "2022-11-15T22:03:47.151780",
     "exception": false,
     "start_time": "2022-11-15T22:03:45.226630",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15,5))\n",
    "plt.plot(history['loss'], label='Training', alpha=.3, color='#ff7f0e', linestyle='--')\n",
    "plt.plot(history['val_loss'], label='Validation', alpha=.8, color='#ff7f0e')\n",
    "plt.legend(loc='upper left')\n",
    "plt.title('Categorical Crossentropy')\n",
    "plt.grid(alpha=.3)\n",
    "\n",
    "plt.figure(figsize=(15,5))\n",
    "plt.plot(history['accuracy'], label='Training', alpha=.8, color='#ff7f0e', linestyle='--')\n",
    "plt.plot(history['val_accuracy'], label='Validation', alpha=.8, color='#ff7f0e')\n",
    "plt.legend(loc='upper left')\n",
    "plt.title('Accuracy')\n",
    "plt.grid(alpha=.3)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "bde42187",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-11-15T22:03:49.172843Z",
     "iopub.status.busy": "2022-11-15T22:03:49.172474Z",
     "iopub.status.idle": "2022-11-15T22:04:36.975088Z",
     "shell.execute_reply": "2022-11-15T22:04:36.974028Z"
    },
    "id": "IwhA2qPP0Nvg",
    "papermill": {
     "duration": 48.85976,
     "end_time": "2022-11-15T22:04:36.977935",
     "exception": false,
     "start_time": "2022-11-15T22:03:48.118175",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-11-15 22:04:06.130714: W tensorflow/python/util/util.cc:348] Sets are not currently considered sequences, but this may change in the future, so consider avoiding using them.\n"
     ]
    }
   ],
   "source": [
    "model.save(\"models/myModel\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "b64d1f41",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-11-15T22:04:39.084303Z",
     "iopub.status.busy": "2022-11-15T22:04:39.083901Z",
     "iopub.status.idle": "2022-11-15T22:04:49.470133Z",
     "shell.execute_reply": "2022-11-15T22:04:49.469234Z"
    },
    "papermill": {
     "duration": 11.515662,
     "end_time": "2022-11-15T22:04:49.473392",
     "exception": false,
     "start_time": "2022-11-15T22:04:37.957730",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-11-15 22:04:39.119711: W tensorflow/core/grappler/optimizers/data/auto_shard.cc:695] AUTO sharding policy will apply DATA sharding policy as it failed to apply FILE sharding policy because of the following reason: Found an unshardable source dataset: name: \"TensorSliceDataset/_1\"\n",
      "op: \"TensorSliceDataset\"\n",
      "input: \"Placeholder/_0\"\n",
      "attr {\n",
      "  key: \"Toutput_types\"\n",
      "  value {\n",
      "    list {\n",
      "      type: DT_STRING\n",
      "    }\n",
      "  }\n",
      "}\n",
      "attr {\n",
      "  key: \"output_shapes\"\n",
      "  value {\n",
      "    list {\n",
      "      shape {\n",
      "      }\n",
      "    }\n",
      "  }\n",
      "}\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "23/23 [==============================] - 2s 93ms/step - loss: 1.1363 - accuracy: 0.8277\n",
      "{'loss': 1.1363028287887573, 'accuracy': 0.8276836276054382}\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-11-15 22:04:41.573196: W tensorflow/core/grappler/optimizers/data/auto_shard.cc:695] AUTO sharding policy will apply DATA sharding policy as it failed to apply FILE sharding policy because of the following reason: Found an unshardable source dataset: name: \"TensorSliceDataset/_1\"\n",
      "op: \"TensorSliceDataset\"\n",
      "input: \"Placeholder/_0\"\n",
      "attr {\n",
      "  key: \"Toutput_types\"\n",
      "  value {\n",
      "    list {\n",
      "      type: DT_STRING\n",
      "    }\n",
      "  }\n",
      "}\n",
      "attr {\n",
      "  key: \"output_shapes\"\n",
      "  value {\n",
      "    list {\n",
      "      shape {\n",
      "      }\n",
      "    }\n",
      "  }\n",
      "}\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.8277\n",
      "Precision: 0.8181\n",
      "Recall: 0.8016\n",
      "F1: 0.8055\n",
      "F1 (per class):\n",
      "[0.49275362 0.83870968 0.87068966 0.83937824 0.80365297 0.89411765\n",
      " 0.93684211 0.76777251]\n",
      "WORST CLASS: Species: 1\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 540x540 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#model = tf.keras.models.load_model(\"models/myModel_AUGMENTEDDATA_WEIGHTEDCLASSES\")\n",
    "metrics = model.evaluate(validation_set, return_dict=True)\n",
    "print(metrics)\n",
    "\n",
    "DATASET_EVAL = validation_set\n",
    "labels = np.concatenate([y for x, y in DATASET_EVAL], axis=0) \n",
    "predictions = model.predict(validation_set) \n",
    "\n",
    "assert len(labels) == len(predictions)\n",
    "\n",
    "# Compute the confusion matrix\n",
    "cm = confusion_matrix(np.argmax(labels, axis=-1), np.argmax(predictions, axis=-1))\n",
    "\n",
    "# Compute the classification metrics\n",
    "accuracy = accuracy_score(np.argmax(labels, axis=-1), np.argmax(predictions, axis=-1))\n",
    "precision = precision_score(np.argmax(labels, axis=-1), np.argmax(predictions, axis=-1), average='macro')\n",
    "recall = recall_score(np.argmax(labels, axis=-1), np.argmax(predictions, axis=-1), average='macro')\n",
    "f1 = f1_score(np.argmax(labels, axis=-1), np.argmax(predictions, axis=-1), average='macro')\n",
    "f1_classes = f1_score(np.argmax(labels, axis=-1), np.argmax(predictions, axis=-1), average=None)\n",
    "\n",
    "print('Accuracy:',accuracy.round(4))\n",
    "print('Precision:',precision.round(4))\n",
    "print('Recall:',recall.round(4))\n",
    "print('F1:',f1.round(4))\n",
    "print('F1 (per class):')\n",
    "print(f1_classes)\n",
    "\n",
    "print('WORST CLASS: Species:', np.argmin(f1_classes)+1)\n",
    "\n",
    "# Plot the confusion matrix\n",
    "ig, ax = plt.subplots(figsize=(7.5, 7.5))\n",
    "ax.matshow(cm, cmap=plt.cm.Blues, alpha=0.3)\n",
    "for i in range(cm.shape[0]):\n",
    "    for j in range(cm.shape[1]):\n",
    "        ax.text(x=j, y=i,s=cm[i, j], va='center', ha='center', size='xx-large')\n",
    "\n",
    "        \n",
    "plt.xlabel('Predictions', fontsize=18)\n",
    "plt.ylabel('Actuals', fontsize=18)\n",
    "plt.title('Confusion Matrix', fontsize=18)\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.12"
  },
  "papermill": {
   "default_parameters": {},
   "duration": 9092.445605,
   "end_time": "2022-11-15T22:04:54.511877",
   "environment_variables": {},
   "exception": null,
   "input_path": "__notebook__.ipynb",
   "output_path": "__notebook__.ipynb",
   "parameters": {},
   "start_time": "2022-11-15T19:33:22.066272",
   "version": "2.3.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}