diff --git a/COVID Neural Net.ipynb b/COVID Neural Net.ipynb
index ef502bb..768621e 100644
--- a/COVID Neural Net.ipynb	
+++ b/COVID Neural Net.ipynb	
@@ -2,8 +2,8 @@
  "cells": [
   {
    "cell_type": "code",
-   "execution_count": 1,
-   "id": "advance-demonstration",
+   "execution_count": 2,
+   "id": "center-collective",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -12,8 +12,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 2,
-   "id": "hollow-edinburgh",
+   "execution_count": 3,
+   "id": "established-temperature",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -28,8 +28,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 3,
-   "id": "innovative-preliminary",
+   "execution_count": 4,
+   "id": "entertaining-waste",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -39,8 +39,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 4,
-   "id": "hundred-elevation",
+   "execution_count": 5,
+   "id": "invalid-travel",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -49,7 +49,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "premium-symbol",
+   "id": "revolutionary-strategy",
    "metadata": {},
    "source": [
     "# Loading in Data"
@@ -57,8 +57,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 5,
-   "id": "sharp-swiss",
+   "execution_count": 6,
+   "id": "manufactured-soundtrack",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -67,8 +67,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
-   "id": "worthy-helmet",
+   "execution_count": 7,
+   "id": "fluid-accreditation",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -78,8 +78,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 7,
-   "id": "introductory-heart",
+   "execution_count": 8,
+   "id": "applicable-basics",
    "metadata": {},
    "outputs": [
     {
@@ -98,8 +98,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 8,
-   "id": "compliant-australia",
+   "execution_count": 9,
+   "id": "brazilian-skill",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -112,8 +112,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 9,
-   "id": "perfect-intelligence",
+   "execution_count": 10,
+   "id": "located-religion",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -123,17 +123,17 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 10,
-   "id": "collectible-minute",
+   "execution_count": 11,
+   "id": "alive-windsor",
    "metadata": {},
    "outputs": [
     {
      "data": {
       "text/plain": [
-       "<matplotlib.image.AxesImage at 0x1e80dc25190>"
+       "<matplotlib.image.AxesImage at 0x2831a5f12b0>"
       ]
      },
-     "execution_count": 10,
+     "execution_count": 11,
      "metadata": {},
      "output_type": "execute_result"
     },
@@ -156,17 +156,17 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 11,
-   "id": "sophisticated-passage",
+   "execution_count": 12,
+   "id": "mysterious-flood",
    "metadata": {},
    "outputs": [
     {
      "data": {
       "text/plain": [
-       "<matplotlib.image.AxesImage at 0x1e81ac064c0>"
+       "<matplotlib.image.AxesImage at 0x2831c1e45e0>"
       ]
      },
-     "execution_count": 11,
+     "execution_count": 12,
      "metadata": {},
      "output_type": "execute_result"
     },
@@ -189,8 +189,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 12,
-   "id": "extended-answer",
+   "execution_count": 13,
+   "id": "continued-nicaragua",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -199,8 +199,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 13,
-   "id": "eleven-recycling",
+   "execution_count": 14,
+   "id": "female-cement",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -212,8 +212,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 14,
-   "id": "broad-momentum",
+   "execution_count": 15,
+   "id": "sealed-oasis",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -223,8 +223,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 15,
-   "id": "earlier-procurement",
+   "execution_count": 16,
+   "id": "listed-arkansas",
    "metadata": {},
    "outputs": [
     {
@@ -978,7 +978,7 @@
        " 'Non-Covid']"
       ]
      },
-     "execution_count": 15,
+     "execution_count": 16,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -989,7 +989,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "relevant-trick",
+   "id": "interpreted-compact",
    "metadata": {},
    "source": [
     "# Loading in Data 2.0"
@@ -997,8 +997,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 16,
-   "id": "technological-dover",
+   "execution_count": 17,
+   "id": "express-lesson",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -1007,8 +1007,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 17,
-   "id": "proof-constraint",
+   "execution_count": 18,
+   "id": "alien-might",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -1039,7 +1039,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "persistent-looking",
+   "id": "annual-boundary",
    "metadata": {},
    "source": [
     "Fairly Balanced in the Training Set"
@@ -1047,8 +1047,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 18,
-   "id": "adjustable-robertson",
+   "execution_count": 19,
+   "id": "lasting-persian",
    "metadata": {},
    "outputs": [
     {
@@ -1065,8 +1065,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 19,
-   "id": "desperate-elimination",
+   "execution_count": 20,
+   "id": "assisted-lincoln",
    "metadata": {},
    "outputs": [
     {
@@ -1083,8 +1083,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 20,
-   "id": "forced-contractor",
+   "execution_count": 21,
+   "id": "animated-litigation",
    "metadata": {},
    "outputs": [
     {
@@ -1101,8 +1101,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 21,
-   "id": "bacterial-reform",
+   "execution_count": 22,
+   "id": "listed-crystal",
    "metadata": {
     "scrolled": true
    },
@@ -1122,14 +1122,14 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "portuguese-lighting",
+   "id": "stupid-cemetery",
    "metadata": {},
    "outputs": [],
    "source": []
   },
   {
    "cell_type": "markdown",
-   "id": "funded-separation",
+   "id": "competitive-criticism",
    "metadata": {},
    "source": [
     "# Data Preprocessing"
@@ -1137,67 +1137,49 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 22,
-   "id": "industrial-catholic",
+   "execution_count": 23,
+   "id": "understanding-english",
    "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Found 596 images belonging to 2 classes.\n",
-      "Found 73 images belonging to 2 classes.\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "from keras.preprocessing.image import ImageDataGenerator\n",
     "\n",
-    "# All images will be rescaled by 1./255\n",
-    "train_datagen = ImageDataGenerator(rescale=1./255)\n",
-    "test_datagen = ImageDataGenerator(rescale=1./255)\n",
+    "# # All images will be rescaled by 1./255\n",
+    "# train_datagen = ImageDataGenerator(rescale=1./255)\n",
+    "# test_datagen = ImageDataGenerator(rescale=1./255)\n",
     "\n",
-    "train_generator = train_datagen.flow_from_directory(\n",
-    "        # This is the target directory\n",
-    "        train_dir,\n",
-    "        # All images will be resized to 150x150\n",
-    "        target_size=(218, 178),\n",
-    "        batch_size=20,\n",
-    "        # Since we use binary_crossentropy loss, we need binary labels\n",
-    "        class_mode='binary')\n",
+    "# train_generator = train_datagen.flow_from_directory(\n",
+    "#         # This is the target directory\n",
+    "#         train_dir,\n",
+    "#         # All images will be resized to 150x150\n",
+    "#         target_size=(218, 178),\n",
+    "#         batch_size=20,\n",
+    "#         # Since we use binary_crossentropy loss, we need binary labels\n",
+    "#         class_mode='binary')\n",
     "\n",
-    "validation_generator = test_datagen.flow_from_directory(\n",
-    "        validation_dir,\n",
-    "        target_size=(218, 178),\n",
-    "        batch_size=20,\n",
-    "        class_mode='binary')"
+    "# validation_generator = test_datagen.flow_from_directory(\n",
+    "#         validation_dir,\n",
+    "#         target_size=(218, 178),\n",
+    "#         batch_size=20,\n",
+    "#         class_mode='binary')"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 23,
-   "id": "experimental-hardware",
+   "execution_count": 24,
+   "id": "preliminary-serbia",
    "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "data batch shape: (20, 218, 178, 3)\n",
-      "labels batch shape: (20,)\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
-    "for data_batch, labels_batch in train_generator:\n",
-    "    print('data batch shape:', data_batch.shape)\n",
-    "    print('labels batch shape:', labels_batch.shape)\n",
-    "    break"
+    "# for data_batch, labels_batch in train_generator:\n",
+    "#     print('data batch shape:', data_batch.shape)\n",
+    "#     print('labels batch shape:', labels_batch.shape)\n",
+    "#     break"
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "printable-burns",
+   "id": "recognized-thickness",
    "metadata": {},
    "source": [
     "# Building the Model"
@@ -1205,7 +1187,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "protected-inventory",
+   "id": "respiratory-source",
    "metadata": {},
    "source": [
     "We will be using a VGG-16 pre-trained model as the base for our image classification. I am keeping the base of the model, but am setting include_top to be False so that we can use our own densely connected classifier."
@@ -1213,8 +1195,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 24,
-   "id": "boxed-supplement",
+   "execution_count": 25,
+   "id": "regular-rescue",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -1227,8 +1209,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 25,
-   "id": "processed-confidence",
+   "execution_count": 26,
+   "id": "aging-advocacy",
    "metadata": {},
    "outputs": [
     {
@@ -1290,8 +1272,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 26,
-   "id": "charged-puzzle",
+   "execution_count": 27,
+   "id": "small-gazette",
    "metadata": {},
    "outputs": [
     {
@@ -1335,7 +1317,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "adjusted-circumstances",
+   "id": "mechanical-driving",
    "metadata": {},
    "source": [
     "flatten the features so they can be processed in a densely connected classifier."
@@ -1343,8 +1325,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 27,
-   "id": "radio-profile",
+   "execution_count": 28,
+   "id": "organic-seating",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -1355,8 +1337,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 28,
-   "id": "determined-distance",
+   "execution_count": 29,
+   "id": "frozen-dynamics",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -1372,8 +1354,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 29,
-   "id": "european-hands",
+   "execution_count": 30,
+   "id": "improved-snake",
    "metadata": {},
    "outputs": [
     {
@@ -1405,7 +1387,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "corrected-snake",
+   "id": "limited-refund",
    "metadata": {},
    "source": [
     "Before we compile and train our model, a very important thing to do is to freeze the convolutional base. \"Freezing\" a layer or set of \n",
@@ -1416,8 +1398,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 30,
-   "id": "neither-array",
+   "execution_count": 31,
+   "id": "elect-thanksgiving",
    "metadata": {},
    "outputs": [
     {
@@ -1435,8 +1417,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 31,
-   "id": "prerequisite-defeat",
+   "execution_count": 32,
+   "id": "chronic-species",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -1445,8 +1427,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 32,
-   "id": "metropolitan-reproduction",
+   "execution_count": 33,
+   "id": "enormous-marking",
    "metadata": {},
    "outputs": [
     {
@@ -1464,7 +1446,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "jewish-congress",
+   "id": "surgical-aquarium",
    "metadata": {},
    "source": [
     "Two weights in each dense layer, the bias and the weight vector"
@@ -1472,157 +1454,63 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 33,
-   "id": "sorted-competition",
+   "execution_count": 34,
+   "id": "legendary-analysis",
    "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Found 596 images belonging to 2 classes.\n",
-      "Found 73 images belonging to 2 classes.\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
-    "from keras import optimizers\n",
+    " from keras import optimizers\n",
     "\n",
-    "train_generator = train_datagen.flow_from_directory(\n",
-    "        # This is the target directory\n",
-    "        train_dir,\n",
-    "        # All images will be resized to 150x150\n",
-    "        target_size=(218, 178),\n",
-    "        batch_size=20,\n",
-    "        # Since we use binary_crossentropy loss, we need binary labels\n",
-    "        class_mode='binary')\n",
+    "# train_generator = train_datagen.flow_from_directory(\n",
+    "#         # This is the target directory\n",
+    "#         train_dir,\n",
+    "#         # All images will be resized to 150x150\n",
+    "#         target_size=(218, 178),\n",
+    "#         batch_size=20,\n",
+    "#         # Since we use binary_crossentropy loss, we need binary labels\n",
+    "#         class_mode='binary')\n",
     "\n",
-    "validation_generator = test_datagen.flow_from_directory(\n",
-    "        validation_dir,\n",
-    "        target_size=(218, 178),\n",
-    "        batch_size=20,\n",
-    "        class_mode='binary')\n",
+    "# validation_generator = test_datagen.flow_from_directory(\n",
+    "#         validation_dir,\n",
+    "#         target_size=(218, 178),\n",
+    "#         batch_size=20,\n",
+    "#         class_mode='binary')\n",
     "\n",
-    "model.compile(loss='binary_crossentropy',\n",
-    "              optimizer=optimizers.RMSprop(lr=2e-5),\n",
-    "              metrics=['acc'])\n"
+    "# model.compile(loss='binary_crossentropy',\n",
+    "#               optimizer=optimizers.RMSprop(lr=2e-5),\n",
+    "#               metrics=['acc'])\n"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 34,
-   "id": "after-arkansas",
+   "execution_count": 35,
+   "id": "prescription-sunglasses",
    "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "29 3\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
-    "print(596//batch_size, 73//batch_size)"
+    "# print(596//batch_size, 73//batch_size)"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 35,
-   "id": "meaningful-gothic",
+   "execution_count": 36,
+   "id": "smaller-opportunity",
    "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Epoch 1/35\n",
-      "29/29 [==============================] - 3s 87ms/step - loss: 0.6882 - acc: 0.5747 - val_loss: 0.5551 - val_acc: 0.7333\n",
-      "Epoch 2/35\n",
-      "29/29 [==============================] - 2s 58ms/step - loss: 0.5422 - acc: 0.7626 - val_loss: 0.4502 - val_acc: 0.7833\n",
-      "Epoch 3/35\n",
-      "29/29 [==============================] - 2s 58ms/step - loss: 0.4692 - acc: 0.7938 - val_loss: 0.4454 - val_acc: 0.7833\n",
-      "Epoch 4/35\n",
-      "29/29 [==============================] - 2s 56ms/step - loss: 0.4024 - acc: 0.8498 - val_loss: 0.4209 - val_acc: 0.8167\n",
-      "Epoch 5/35\n",
-      "29/29 [==============================] - 2s 58ms/step - loss: 0.3938 - acc: 0.8698 - val_loss: 0.4033 - val_acc: 0.8000\n",
-      "Epoch 6/35\n",
-      "29/29 [==============================] - 2s 58ms/step - loss: 0.3605 - acc: 0.8829 - val_loss: 0.4256 - val_acc: 0.7667\n",
-      "Epoch 7/35\n",
-      "29/29 [==============================] - 2s 59ms/step - loss: 0.3432 - acc: 0.8648 - val_loss: 0.3596 - val_acc: 0.8667\n",
-      "Epoch 8/35\n",
-      "29/29 [==============================] - 2s 58ms/step - loss: 0.3261 - acc: 0.8838 - val_loss: 0.3304 - val_acc: 0.8500\n",
-      "Epoch 9/35\n",
-      "29/29 [==============================] - 2s 58ms/step - loss: 0.2657 - acc: 0.9221 - val_loss: 0.3691 - val_acc: 0.8333\n",
-      "Epoch 10/35\n",
-      "29/29 [==============================] - 2s 57ms/step - loss: 0.2475 - acc: 0.9347 - val_loss: 0.3405 - val_acc: 0.8500\n",
-      "Epoch 11/35\n",
-      "29/29 [==============================] - 2s 58ms/step - loss: 0.2289 - acc: 0.9241 - val_loss: 0.3436 - val_acc: 0.8500\n",
-      "Epoch 12/35\n",
-      "29/29 [==============================] - 2s 57ms/step - loss: 0.2160 - acc: 0.9419 - val_loss: 0.3527 - val_acc: 0.8333\n",
-      "Epoch 13/35\n",
-      "29/29 [==============================] - 2s 57ms/step - loss: 0.1889 - acc: 0.9524 - val_loss: 0.2974 - val_acc: 0.9000\n",
-      "Epoch 14/35\n",
-      "29/29 [==============================] - 2s 57ms/step - loss: 0.1774 - acc: 0.9664 - val_loss: 0.3310 - val_acc: 0.8500\n",
-      "Epoch 15/35\n",
-      "29/29 [==============================] - 2s 58ms/step - loss: 0.1863 - acc: 0.9649 - val_loss: 0.2959 - val_acc: 0.9167\n",
-      "Epoch 16/35\n",
-      "29/29 [==============================] - 2s 58ms/step - loss: 0.1669 - acc: 0.9717 - val_loss: 0.2850 - val_acc: 0.9000\n",
-      "Epoch 17/35\n",
-      "29/29 [==============================] - 2s 57ms/step - loss: 0.1625 - acc: 0.9688 - val_loss: 0.3316 - val_acc: 0.8333\n",
-      "Epoch 18/35\n",
-      "29/29 [==============================] - 2s 57ms/step - loss: 0.1520 - acc: 0.9830 - val_loss: 0.3502 - val_acc: 0.8500\n",
-      "Epoch 19/35\n",
-      "29/29 [==============================] - 2s 56ms/step - loss: 0.1374 - acc: 0.9851 - val_loss: 0.2882 - val_acc: 0.8833\n",
-      "Epoch 20/35\n",
-      "29/29 [==============================] - 2s 58ms/step - loss: 0.1346 - acc: 0.9763 - val_loss: 0.2851 - val_acc: 0.8667\n",
-      "Epoch 21/35\n",
-      "29/29 [==============================] - 2s 57ms/step - loss: 0.1216 - acc: 0.9816 - val_loss: 0.3446 - val_acc: 0.8500\n",
-      "Epoch 22/35\n",
-      "29/29 [==============================] - 2s 57ms/step - loss: 0.1077 - acc: 0.9891 - val_loss: 0.2455 - val_acc: 0.8667\n",
-      "Epoch 23/35\n",
-      "29/29 [==============================] - 2s 57ms/step - loss: 0.0943 - acc: 0.9947 - val_loss: 0.2579 - val_acc: 0.9000\n",
-      "Epoch 24/35\n",
-      "29/29 [==============================] - 2s 56ms/step - loss: 0.0939 - acc: 0.9986 - val_loss: 0.2669 - val_acc: 0.8833\n",
-      "Epoch 25/35\n",
-      "29/29 [==============================] - 2s 57ms/step - loss: 0.0812 - acc: 0.9907 - val_loss: 0.3305 - val_acc: 0.8667\n",
-      "Epoch 26/35\n",
-      "29/29 [==============================] - 2s 57ms/step - loss: 0.0765 - acc: 0.9998 - val_loss: 0.2578 - val_acc: 0.8833\n",
-      "Epoch 27/35\n",
-      "29/29 [==============================] - 2s 57ms/step - loss: 0.0765 - acc: 1.0000 - val_loss: 0.2883 - val_acc: 0.8667\n",
-      "Epoch 28/35\n",
-      "29/29 [==============================] - 2s 58ms/step - loss: 0.0670 - acc: 0.9990 - val_loss: 0.2817 - val_acc: 0.8833\n",
-      "Epoch 29/35\n",
-      "29/29 [==============================] - 2s 57ms/step - loss: 0.0618 - acc: 0.9980 - val_loss: 0.2338 - val_acc: 0.9167\n",
-      "Epoch 30/35\n",
-      "29/29 [==============================] - 2s 57ms/step - loss: 0.0627 - acc: 0.9958 - val_loss: 0.2536 - val_acc: 0.8833\n",
-      "Epoch 31/35\n",
-      "29/29 [==============================] - 2s 57ms/step - loss: 0.0581 - acc: 1.0000 - val_loss: 0.2244 - val_acc: 0.9000\n",
-      "Epoch 32/35\n",
-      "29/29 [==============================] - 2s 56ms/step - loss: 0.0541 - acc: 1.0000 - val_loss: 0.2464 - val_acc: 0.8833\n",
-      "Epoch 33/35\n",
-      "29/29 [==============================] - 2s 57ms/step - loss: 0.0490 - acc: 0.9981 - val_loss: 0.3140 - val_acc: 0.8500\n",
-      "Epoch 34/35\n",
-      "29/29 [==============================] - 2s 59ms/step - loss: 0.0405 - acc: 1.0000 - val_loss: 0.3190 - val_acc: 0.8333\n",
-      "Epoch 35/35\n",
-      "29/29 [==============================] - 2s 57ms/step - loss: 0.0393 - acc: 1.0000 - val_loss: 0.2224 - val_acc: 0.9000\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
-    "history = model.fit(\n",
-    "      train_generator,\n",
-    "      steps_per_epoch=29,\n",
-    "      epochs=35,\n",
-    "      validation_data=validation_generator,\n",
-    "      validation_steps=3,\n",
-    "      verbose=1)"
+    "# history = model.fit(\n",
+    "#       train_generator,\n",
+    "#       steps_per_epoch=29,\n",
+    "#       epochs=35,\n",
+    "#       validation_data=validation_generator,\n",
+    "#       validation_steps=3,\n",
+    "#       verbose=1)"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 36,
-   "id": "vanilla-radio",
+   "execution_count": 37,
+   "id": "rocky-compensation",
    "metadata": {},
    "outputs": [
     {
@@ -1634,7 +1522,7 @@
       "memory_limit: 268435456\n",
       "locality {\n",
       "}\n",
-      "incarnation: 16838192518558930312\n",
+      "incarnation: 10009789183137472670\n",
       ", name: \"/device:GPU:0\"\n",
       "device_type: \"GPU\"\n",
       "memory_limit: 6910041152\n",
@@ -1643,7 +1531,7 @@
       "  links {\n",
       "  }\n",
       "}\n",
-      "incarnation: 5839872070457136224\n",
+      "incarnation: 4216352920101208748\n",
       "physical_device_desc: \"device: 0, name: GeForce RTX 3070, pci bus id: 0000:09:00.0, compute capability: 8.6\"\n",
       "]\n"
      ]
@@ -1656,8 +1544,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 37,
-   "id": "monthly-judge",
+   "execution_count": 38,
+   "id": "another-framework",
    "metadata": {},
    "outputs": [
     {
@@ -1676,72 +1564,55 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 38,
-   "id": "sunrise-football",
+   "execution_count": 39,
+   "id": "reserved-arthritis",
    "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "image/png": "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\n",
-      "text/plain": [
-       "<Figure size 432x288 with 1 Axes>"
-      ]
-     },
-     "metadata": {
-      "needs_background": "light"
-     },
-     "output_type": "display_data"
-    },
-    {
-     "data": {
-      "image/png": "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\n",
-      "text/plain": [
-       "<Figure size 432x288 with 1 Axes>"
-      ]
-     },
-     "metadata": {
-      "needs_background": "light"
-     },
-     "output_type": "display_data"
-    }
-   ],
+   "outputs": [],
    "source": [
-    "import matplotlib.pyplot as plt\n",
+    "# import matplotlib.pyplot as plt\n",
     "\n",
-    "acc = history.history['acc']\n",
-    "val_acc = history.history['val_acc']\n",
-    "loss = history.history['loss']\n",
-    "val_loss = history.history['val_loss']\n",
+    "# acc = history.history['acc']\n",
+    "# val_acc = history.history['val_acc']\n",
+    "# loss = history.history['loss']\n",
+    "# val_loss = history.history['val_loss']\n",
     "\n",
-    "epochs = range(len(acc))\n",
+    "# epochs = range(len(acc))\n",
     "\n",
-    "plt.plot(epochs, acc, 'bo', label='Training acc')\n",
-    "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
-    "plt.title('Training and validation accuracy')\n",
-    "plt.legend()\n",
+    "# plt.plot(epochs, acc, 'bo', label='Training acc')\n",
+    "# plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
+    "# plt.title('Training and validation accuracy')\n",
+    "# plt.legend()\n",
     "\n",
-    "plt.figure()\n",
+    "# plt.figure()\n",
     "\n",
-    "plt.plot(epochs, loss, 'bo', label='Training loss')\n",
-    "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
-    "plt.title('Training and validation loss')\n",
-    "plt.legend()\n",
+    "# plt.plot(epochs, loss, 'bo', label='Training loss')\n",
+    "# plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
+    "# plt.title('Training and validation loss')\n",
+    "# plt.legend()\n",
     "\n",
-    "plt.show()\n"
+    "# plt.show()\n"
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "white-rider",
+   "id": "emotional-front",
    "metadata": {},
    "source": [
     "Lots of overfitting even from the start, data augmentation can help fix this. But it requires GPU usage in order to be feasible."
    ]
   },
+  {
+   "cell_type": "markdown",
+   "id": "psychological-nepal",
+   "metadata": {},
+   "source": [
+    "# Data Preprocessing with Augmentation"
+   ]
+  },
   {
    "cell_type": "code",
-   "execution_count": 39,
-   "id": "composed-doubt",
+   "execution_count": 47,
+   "id": "engaging-grocery",
    "metadata": {},
    "outputs": [
     {
@@ -1791,8 +1662,53 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 41,
-   "id": "eastern-melissa",
+   "execution_count": 48,
+   "id": "exceptional-parade",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "data batch shape: (20, 218, 178, 3)\n",
+      "labels batch shape: (20,)\n"
+     ]
+    }
+   ],
+   "source": [
+    "for data_batch, labels_batch in train_generator:\n",
+    "    print('data batch shape:', data_batch.shape)\n",
+    "    print('labels batch shape:', labels_batch.shape)\n",
+    "    break"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 49,
+   "id": "western-stranger",
+   "metadata": {
+    "scrolled": true
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "<tensorflow.python.keras.preprocessing.image.DirectoryIterator at 0x28359d638b0>"
+      ]
+     },
+     "execution_count": 49,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "train_generator"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 50,
+   "id": "baking-armor",
    "metadata": {},
    "outputs": [
     {
@@ -1800,69 +1716,73 @@
      "output_type": "stream",
      "text": [
       "Epoch 1/30\n",
-      "29/29 [==============================] - 5s 179ms/step - loss: 0.5866 - acc: 0.7069 - val_loss: 0.3745 - val_acc: 0.8333\n",
+      "29/29 [==============================] - 6s 182ms/step - loss: 0.7200 - acc: 0.5112 - val_loss: 0.6353 - val_acc: 0.6667\n",
       "Epoch 2/30\n",
-      "29/29 [==============================] - 5s 173ms/step - loss: 0.5038 - acc: 0.7535 - val_loss: 0.3365 - val_acc: 0.8500\n",
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.6241 - acc: 0.6435 - val_loss: 0.5594 - val_acc: 0.6667\n",
       "Epoch 3/30\n",
-      "29/29 [==============================] - 5s 172ms/step - loss: 0.5150 - acc: 0.7292 - val_loss: 0.3872 - val_acc: 0.8167\n",
+      "29/29 [==============================] - 5s 171ms/step - loss: 0.6239 - acc: 0.6879 - val_loss: 0.5561 - val_acc: 0.6833\n",
       "Epoch 4/30\n",
-      "29/29 [==============================] - 5s 172ms/step - loss: 0.4991 - acc: 0.7604 - val_loss: 0.3126 - val_acc: 0.8167\n",
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.5846 - acc: 0.6973 - val_loss: 0.6462 - val_acc: 0.6500\n",
       "Epoch 5/30\n",
-      "29/29 [==============================] - 5s 171ms/step - loss: 0.5020 - acc: 0.7500 - val_loss: 0.3045 - val_acc: 0.9000\n",
+      "29/29 [==============================] - 5s 170ms/step - loss: 0.5706 - acc: 0.6893 - val_loss: 0.5446 - val_acc: 0.7000\n",
       "Epoch 6/30\n",
-      "29/29 [==============================] - 5s 170ms/step - loss: 0.4615 - acc: 0.7830 - val_loss: 0.3367 - val_acc: 0.8500\n",
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.5354 - acc: 0.7523 - val_loss: 0.5141 - val_acc: 0.7500\n",
       "Epoch 7/30\n",
-      "29/29 [==============================] - 5s 172ms/step - loss: 0.4737 - acc: 0.7812 - val_loss: 0.3246 - val_acc: 0.8833\n",
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.5456 - acc: 0.7270 - val_loss: 0.5143 - val_acc: 0.7833\n",
       "Epoch 8/30\n",
-      "29/29 [==============================] - 5s 171ms/step - loss: 0.4718 - acc: 0.7691 - val_loss: 0.4177 - val_acc: 0.8333\n",
+      "29/29 [==============================] - 5s 170ms/step - loss: 0.5685 - acc: 0.7323 - val_loss: 0.4981 - val_acc: 0.7500\n",
       "Epoch 9/30\n",
-      "29/29 [==============================] - 5s 172ms/step - loss: 0.4742 - acc: 0.7604 - val_loss: 0.3013 - val_acc: 0.8667\n",
+      "29/29 [==============================] - 5s 173ms/step - loss: 0.5514 - acc: 0.7144 - val_loss: 0.5160 - val_acc: 0.7167\n",
       "Epoch 10/30\n",
-      "29/29 [==============================] - 5s 170ms/step - loss: 0.4541 - acc: 0.7726 - val_loss: 0.3413 - val_acc: 0.8500\n",
+      "29/29 [==============================] - 5s 177ms/step - loss: 0.5189 - acc: 0.7434 - val_loss: 0.4775 - val_acc: 0.7500\n",
       "Epoch 11/30\n",
-      "29/29 [==============================] - 5s 171ms/step - loss: 0.4457 - acc: 0.7934 - val_loss: 0.3308 - val_acc: 0.8667\n",
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.5023 - acc: 0.7701 - val_loss: 0.4596 - val_acc: 0.7667\n",
       "Epoch 12/30\n",
-      "29/29 [==============================] - 5s 172ms/step - loss: 0.4183 - acc: 0.7914 - val_loss: 0.4123 - val_acc: 0.8167\n",
+      "29/29 [==============================] - 5s 173ms/step - loss: 0.4970 - acc: 0.7871 - val_loss: 0.4886 - val_acc: 0.7500\n",
       "Epoch 13/30\n",
-      "29/29 [==============================] - 5s 171ms/step - loss: 0.4291 - acc: 0.7882 - val_loss: 0.3907 - val_acc: 0.8333\n",
+      "29/29 [==============================] - 5s 174ms/step - loss: 0.4825 - acc: 0.7920 - val_loss: 0.4593 - val_acc: 0.8000\n",
       "Epoch 14/30\n",
-      "29/29 [==============================] - 5s 172ms/step - loss: 0.4276 - acc: 0.7969 - val_loss: 0.3833 - val_acc: 0.8333\n",
+      "29/29 [==============================] - 5s 171ms/step - loss: 0.4801 - acc: 0.7709 - val_loss: 0.4556 - val_acc: 0.7833\n",
       "Epoch 15/30\n",
-      "29/29 [==============================] - 5s 171ms/step - loss: 0.4648 - acc: 0.7830 - val_loss: 0.2974 - val_acc: 0.8333\n",
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.5159 - acc: 0.7390 - val_loss: 0.4586 - val_acc: 0.8000\n",
       "Epoch 16/30\n",
-      "29/29 [==============================] - 5s 172ms/step - loss: 0.4490 - acc: 0.7899 - val_loss: 0.3490 - val_acc: 0.8500\n",
+      "29/29 [==============================] - 5s 173ms/step - loss: 0.5095 - acc: 0.7580 - val_loss: 0.4778 - val_acc: 0.7500\n",
       "Epoch 17/30\n",
-      "29/29 [==============================] - 5s 172ms/step - loss: 0.4138 - acc: 0.8160 - val_loss: 0.2703 - val_acc: 0.9167\n",
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.5000 - acc: 0.7435 - val_loss: 0.4974 - val_acc: 0.7333\n",
       "Epoch 18/30\n",
-      "29/29 [==============================] - 5s 171ms/step - loss: 0.4213 - acc: 0.7969 - val_loss: 0.3588 - val_acc: 0.8333\n",
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.4824 - acc: 0.7718 - val_loss: 0.4858 - val_acc: 0.7833\n",
       "Epoch 19/30\n",
-      "29/29 [==============================] - 5s 171ms/step - loss: 0.4459 - acc: 0.7899 - val_loss: 0.3524 - val_acc: 0.8000\n",
+      "29/29 [==============================] - 5s 173ms/step - loss: 0.4989 - acc: 0.7631 - val_loss: 0.4600 - val_acc: 0.7833\n",
       "Epoch 20/30\n",
-      "29/29 [==============================] - 5s 173ms/step - loss: 0.4354 - acc: 0.7897 - val_loss: 0.3583 - val_acc: 0.8500\n",
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.4900 - acc: 0.7419 - val_loss: 0.4774 - val_acc: 0.7333\n",
       "Epoch 21/30\n",
-      "29/29 [==============================] - 5s 172ms/step - loss: 0.4161 - acc: 0.7969 - val_loss: 0.3354 - val_acc: 0.8500\n",
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.4416 - acc: 0.7792 - val_loss: 0.4740 - val_acc: 0.8000\n",
       "Epoch 22/30\n",
-      "29/29 [==============================] - 5s 175ms/step - loss: 0.4113 - acc: 0.8108 - val_loss: 0.3735 - val_acc: 0.8333\n",
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.4249 - acc: 0.8063 - val_loss: 0.4523 - val_acc: 0.8000\n",
       "Epoch 23/30\n",
-      "29/29 [==============================] - 5s 176ms/step - loss: 0.4326 - acc: 0.8090 - val_loss: 0.3121 - val_acc: 0.8333\n",
+      "29/29 [==============================] - 5s 173ms/step - loss: 0.4380 - acc: 0.7973 - val_loss: 0.4799 - val_acc: 0.7500\n",
       "Epoch 24/30\n",
-      "29/29 [==============================] - 5s 183ms/step - loss: 0.3955 - acc: 0.8207 - val_loss: 0.3023 - val_acc: 0.8833\n",
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.4183 - acc: 0.8223 - val_loss: 0.4384 - val_acc: 0.8167\n",
       "Epoch 25/30\n",
-      "29/29 [==============================] - 5s 174ms/step - loss: 0.4033 - acc: 0.8056 - val_loss: 0.3970 - val_acc: 0.8167\n",
+      "29/29 [==============================] - 5s 173ms/step - loss: 0.4468 - acc: 0.7928 - val_loss: 0.5173 - val_acc: 0.7500\n",
       "Epoch 26/30\n",
-      "29/29 [==============================] - 5s 172ms/step - loss: 0.4049 - acc: 0.8177 - val_loss: 0.3546 - val_acc: 0.8333\n",
+      "29/29 [==============================] - 5s 171ms/step - loss: 0.4629 - acc: 0.7985 - val_loss: 0.4874 - val_acc: 0.7333\n",
       "Epoch 27/30\n",
-      "29/29 [==============================] - 5s 172ms/step - loss: 0.3811 - acc: 0.8438 - val_loss: 0.3010 - val_acc: 0.8500\n",
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.4193 - acc: 0.8165 - val_loss: 0.4212 - val_acc: 0.8500\n",
       "Epoch 28/30\n",
-      "29/29 [==============================] - 5s 171ms/step - loss: 0.3983 - acc: 0.8177 - val_loss: 0.3496 - val_acc: 0.8333\n",
+      "29/29 [==============================] - 5s 171ms/step - loss: 0.4375 - acc: 0.7688 - val_loss: 0.4807 - val_acc: 0.7000\n",
       "Epoch 29/30\n",
-      "29/29 [==============================] - 5s 172ms/step - loss: 0.4177 - acc: 0.8090 - val_loss: 0.3825 - val_acc: 0.8333\n",
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.4335 - acc: 0.7780 - val_loss: 0.4413 - val_acc: 0.8000\n",
       "Epoch 30/30\n",
-      "29/29 [==============================] - 5s 171ms/step - loss: 0.3906 - acc: 0.8073 - val_loss: 0.3535 - val_acc: 0.8333\n"
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.4693 - acc: 0.7795 - val_loss: 0.4325 - val_acc: 0.8167\n"
      ]
     }
    ],
    "source": [
+    "import random\n",
+    "\n",
+    "random.seed(123)\n",
+    "\n",
     "history = model.fit(\n",
     "      train_generator,\n",
     "      steps_per_epoch=29,\n",
@@ -1874,23 +1794,23 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 42,
-   "id": "moderate-bottle",
+   "execution_count": 51,
+   "id": "shared-jacob",
    "metadata": {},
    "outputs": [],
    "source": [
-    "model.save('covid_data_aug.h5')"
+    "model.save('covid_model1.h5')"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 43,
-   "id": "addressed-stretch",
+   "execution_count": 52,
+   "id": "nutritional-baghdad",
    "metadata": {},
    "outputs": [
     {
      "data": {
-      "image/png": "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\n",
+      "image/png": "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\n",
       "text/plain": [
        "<Figure size 432x288 with 1 Axes>"
       ]
@@ -1902,7 +1822,7 @@
     },
     {
      "data": {
-      "image/png": "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\n",
+      "image/png": "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\n",
       "text/plain": [
        "<Figure size 432x288 with 1 Axes>"
       ]
@@ -1938,7 +1858,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "positive-silly",
+   "id": "distinguished-nature",
    "metadata": {},
    "source": [
     "## Fine-tuning\n",
@@ -1952,10 +1872,400 @@
     "![fine-tuning VGG16](https://s3.amazonaws.com/book.keras.io/img/ch5/vgg16_fine_tuning.png)"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": 53,
+   "id": "skilled-pickup",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Model: \"vgg16\"\n",
+      "_________________________________________________________________\n",
+      "Layer (type)                 Output Shape              Param #   \n",
+      "=================================================================\n",
+      "input_1 (InputLayer)         [(None, 218, 178, 3)]     0         \n",
+      "_________________________________________________________________\n",
+      "block1_conv1 (Conv2D)        (None, 218, 178, 64)      1792      \n",
+      "_________________________________________________________________\n",
+      "block1_conv2 (Conv2D)        (None, 218, 178, 64)      36928     \n",
+      "_________________________________________________________________\n",
+      "block1_pool (MaxPooling2D)   (None, 109, 89, 64)       0         \n",
+      "_________________________________________________________________\n",
+      "block2_conv1 (Conv2D)        (None, 109, 89, 128)      73856     \n",
+      "_________________________________________________________________\n",
+      "block2_conv2 (Conv2D)        (None, 109, 89, 128)      147584    \n",
+      "_________________________________________________________________\n",
+      "block2_pool (MaxPooling2D)   (None, 54, 44, 128)       0         \n",
+      "_________________________________________________________________\n",
+      "block3_conv1 (Conv2D)        (None, 54, 44, 256)       295168    \n",
+      "_________________________________________________________________\n",
+      "block3_conv2 (Conv2D)        (None, 54, 44, 256)       590080    \n",
+      "_________________________________________________________________\n",
+      "block3_conv3 (Conv2D)        (None, 54, 44, 256)       590080    \n",
+      "_________________________________________________________________\n",
+      "block3_pool (MaxPooling2D)   (None, 27, 22, 256)       0         \n",
+      "_________________________________________________________________\n",
+      "block4_conv1 (Conv2D)        (None, 27, 22, 512)       1180160   \n",
+      "_________________________________________________________________\n",
+      "block4_conv2 (Conv2D)        (None, 27, 22, 512)       2359808   \n",
+      "_________________________________________________________________\n",
+      "block4_conv3 (Conv2D)        (None, 27, 22, 512)       2359808   \n",
+      "_________________________________________________________________\n",
+      "block4_pool (MaxPooling2D)   (None, 13, 11, 512)       0         \n",
+      "_________________________________________________________________\n",
+      "block5_conv1 (Conv2D)        (None, 13, 11, 512)       2359808   \n",
+      "_________________________________________________________________\n",
+      "block5_conv2 (Conv2D)        (None, 13, 11, 512)       2359808   \n",
+      "_________________________________________________________________\n",
+      "block5_conv3 (Conv2D)        (None, 13, 11, 512)       2359808   \n",
+      "_________________________________________________________________\n",
+      "block5_pool (MaxPooling2D)   (None, 6, 5, 512)         0         \n",
+      "=================================================================\n",
+      "Total params: 14,714,688\n",
+      "Trainable params: 0\n",
+      "Non-trainable params: 14,714,688\n",
+      "_________________________________________________________________\n"
+     ]
+    }
+   ],
+   "source": [
+    "conv_base.summary()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "excellent-windsor",
+   "metadata": {},
+   "source": [
+    "\n",
+    "We will fine-tune the last 3 convolutional layers, which means that all layers up until `block4_pool` should be frozen, and the layers \n",
+    "`block5_conv1`, `block5_conv2` and `block5_conv3` should be trainable.\n",
+    "\n",
+    "Why not fine-tune more layers? Why not fine-tune the entire convolutional base? We could. However, we need to consider that:\n",
+    "\n",
+    "* Earlier layers in the convolutional base encode more generic, reusable features, while layers higher up encode more specialized features. It is \n",
+    "more useful to fine-tune the more specialized features, as these are the ones that need to be repurposed on our new problem. There would \n",
+    "be fast-decreasing returns in fine-tuning lower layers.\n",
+    "* The more parameters we are training, the more we are at risk of overfitting. The convolutional base has 15M parameters, so it would be \n",
+    "risky to attempt to train it on our small dataset.\n",
+    "\n",
+    "Thus, in our situation, it is a good strategy to only fine-tune the top 2 to 3 layers in the convolutional base.\n",
+    "\n",
+    "Let's set this up, starting from where we left off in the previous example:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 54,
+   "id": "romantic-blackberry",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "conv_base.trainable = True\n",
+    "\n",
+    "set_trainable = False\n",
+    "for layer in conv_base.layers:\n",
+    "    if layer.name == 'block5_conv1':\n",
+    "        set_trainable = True\n",
+    "    if set_trainable:\n",
+    "        layer.trainable = True\n",
+    "    else:\n",
+    "        layer.trainable = False"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 55,
+   "id": "thermal-porcelain",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "C:\\Users\\jaych\\anaconda3\\envs\\SDP-GPU\\lib\\site-packages\\tensorflow\\python\\keras\\engine\\training.py:1844: UserWarning: `Model.fit_generator` is deprecated and will be removed in a future version. Please use `Model.fit`, which supports generators.\n",
+      "  warnings.warn('`Model.fit_generator` is deprecated and '\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Epoch 1/50\n",
+      "29/29 [==============================] - 11s 218ms/step - loss: 0.4142 - acc: 0.7980 - val_loss: 0.4281 - val_acc: 0.8000\n",
+      "Epoch 2/50\n",
+      "29/29 [==============================] - 5s 174ms/step - loss: 0.4081 - acc: 0.8046 - val_loss: 0.4504 - val_acc: 0.8000\n",
+      "Epoch 3/50\n",
+      "29/29 [==============================] - 5s 174ms/step - loss: 0.3876 - acc: 0.8379 - val_loss: 0.5194 - val_acc: 0.7167\n",
+      "Epoch 4/50\n",
+      "29/29 [==============================] - 5s 178ms/step - loss: 0.4223 - acc: 0.8063 - val_loss: 0.4536 - val_acc: 0.7833\n",
+      "Epoch 5/50\n",
+      "29/29 [==============================] - 5s 176ms/step - loss: 0.3860 - acc: 0.8193 - val_loss: 0.4335 - val_acc: 0.7500\n",
+      "Epoch 6/50\n",
+      "29/29 [==============================] - 5s 174ms/step - loss: 0.3687 - acc: 0.8472 - val_loss: 0.4138 - val_acc: 0.8000\n",
+      "Epoch 7/50\n",
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.3575 - acc: 0.8418 - val_loss: 0.3892 - val_acc: 0.8500\n",
+      "Epoch 8/50\n",
+      "29/29 [==============================] - 5s 177ms/step - loss: 0.3418 - acc: 0.8633 - val_loss: 0.5656 - val_acc: 0.7667\n",
+      "Epoch 9/50\n",
+      "29/29 [==============================] - 5s 174ms/step - loss: 0.3603 - acc: 0.8489 - val_loss: 0.4100 - val_acc: 0.8333\n",
+      "Epoch 10/50\n",
+      "29/29 [==============================] - 5s 173ms/step - loss: 0.3357 - acc: 0.8319 - val_loss: 0.4074 - val_acc: 0.8000\n",
+      "Epoch 11/50\n",
+      "29/29 [==============================] - 5s 179ms/step - loss: 0.2893 - acc: 0.8637 - val_loss: 0.4802 - val_acc: 0.7500\n",
+      "Epoch 12/50\n",
+      "29/29 [==============================] - 5s 176ms/step - loss: 0.3006 - acc: 0.8605 - val_loss: 0.4265 - val_acc: 0.8000\n",
+      "Epoch 13/50\n",
+      "29/29 [==============================] - 5s 175ms/step - loss: 0.3555 - acc: 0.8261 - val_loss: 0.3927 - val_acc: 0.8167\n",
+      "Epoch 14/50\n",
+      "29/29 [==============================] - 5s 174ms/step - loss: 0.3609 - acc: 0.8393 - val_loss: 0.4127 - val_acc: 0.8000\n",
+      "Epoch 15/50\n",
+      "29/29 [==============================] - 5s 175ms/step - loss: 0.3061 - acc: 0.8687 - val_loss: 0.3916 - val_acc: 0.8667\n",
+      "Epoch 16/50\n",
+      "29/29 [==============================] - 5s 176ms/step - loss: 0.2545 - acc: 0.8935 - val_loss: 0.4475 - val_acc: 0.7667\n",
+      "Epoch 17/50\n",
+      "29/29 [==============================] - 5s 173ms/step - loss: 0.2599 - acc: 0.8939 - val_loss: 0.4097 - val_acc: 0.8000\n",
+      "Epoch 18/50\n",
+      "29/29 [==============================] - 5s 176ms/step - loss: 0.3367 - acc: 0.8597 - val_loss: 0.4205 - val_acc: 0.8167\n",
+      "Epoch 19/50\n",
+      "29/29 [==============================] - 5s 174ms/step - loss: 0.2781 - acc: 0.8748 - val_loss: 0.4452 - val_acc: 0.8500\n",
+      "Epoch 20/50\n",
+      "29/29 [==============================] - 5s 171ms/step - loss: 0.2929 - acc: 0.8455 - val_loss: 0.4453 - val_acc: 0.8000\n",
+      "Epoch 21/50\n",
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.2763 - acc: 0.8794 - val_loss: 0.4106 - val_acc: 0.8333\n",
+      "Epoch 22/50\n",
+      "29/29 [==============================] - 5s 171ms/step - loss: 0.2996 - acc: 0.8699 - val_loss: 0.3621 - val_acc: 0.8167\n",
+      "Epoch 23/50\n",
+      "29/29 [==============================] - 5s 171ms/step - loss: 0.2572 - acc: 0.8813 - val_loss: 0.6007 - val_acc: 0.7333\n",
+      "Epoch 24/50\n",
+      "29/29 [==============================] - 5s 171ms/step - loss: 0.2595 - acc: 0.8779 - val_loss: 0.4118 - val_acc: 0.7833\n",
+      "Epoch 25/50\n",
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.2492 - acc: 0.8883 - val_loss: 0.3396 - val_acc: 0.8167\n",
+      "Epoch 26/50\n",
+      "29/29 [==============================] - 5s 171ms/step - loss: 0.2444 - acc: 0.8887 - val_loss: 0.4979 - val_acc: 0.8000\n",
+      "Epoch 27/50\n",
+      "29/29 [==============================] - 5s 171ms/step - loss: 0.2619 - acc: 0.8787 - val_loss: 0.3875 - val_acc: 0.8333\n",
+      "Epoch 28/50\n",
+      "29/29 [==============================] - 5s 171ms/step - loss: 0.2518 - acc: 0.8770 - val_loss: 0.4676 - val_acc: 0.8000\n",
+      "Epoch 29/50\n",
+      "29/29 [==============================] - 5s 175ms/step - loss: 0.2266 - acc: 0.8929 - val_loss: 0.3966 - val_acc: 0.8167\n",
+      "Epoch 30/50\n",
+      "29/29 [==============================] - 5s 175ms/step - loss: 0.2043 - acc: 0.9102 - val_loss: 0.5099 - val_acc: 0.8167\n",
+      "Epoch 31/50\n",
+      "29/29 [==============================] - 5s 173ms/step - loss: 0.2443 - acc: 0.8717 - val_loss: 0.4444 - val_acc: 0.8333\n",
+      "Epoch 32/50\n",
+      "29/29 [==============================] - 5s 173ms/step - loss: 0.2290 - acc: 0.9006 - val_loss: 0.3245 - val_acc: 0.8333\n",
+      "Epoch 33/50\n",
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.1786 - acc: 0.9424 - val_loss: 0.3532 - val_acc: 0.8500\n",
+      "Epoch 34/50\n",
+      "29/29 [==============================] - 5s 177ms/step - loss: 0.2153 - acc: 0.9151 - val_loss: 0.4316 - val_acc: 0.8167\n",
+      "Epoch 35/50\n",
+      "29/29 [==============================] - 5s 171ms/step - loss: 0.1965 - acc: 0.9361 - val_loss: 0.4256 - val_acc: 0.8500\n",
+      "Epoch 36/50\n",
+      "29/29 [==============================] - 5s 171ms/step - loss: 0.2662 - acc: 0.8737 - val_loss: 0.3961 - val_acc: 0.8167\n",
+      "Epoch 37/50\n",
+      "29/29 [==============================] - 5s 174ms/step - loss: 0.1835 - acc: 0.9193 - val_loss: 0.4914 - val_acc: 0.8333\n",
+      "Epoch 38/50\n",
+      "29/29 [==============================] - 5s 171ms/step - loss: 0.1771 - acc: 0.9316 - val_loss: 0.6900 - val_acc: 0.8000\n",
+      "Epoch 39/50\n",
+      "29/29 [==============================] - 5s 170ms/step - loss: 0.1932 - acc: 0.9201 - val_loss: 0.4495 - val_acc: 0.7833\n",
+      "Epoch 40/50\n",
+      "29/29 [==============================] - 5s 170ms/step - loss: 0.2163 - acc: 0.9058 - val_loss: 0.5080 - val_acc: 0.8667\n",
+      "Epoch 41/50\n",
+      "29/29 [==============================] - 5s 171ms/step - loss: 0.2098 - acc: 0.9262 - val_loss: 0.3225 - val_acc: 0.8500\n",
+      "Epoch 42/50\n",
+      "29/29 [==============================] - 5s 173ms/step - loss: 0.1621 - acc: 0.9353 - val_loss: 0.4470 - val_acc: 0.8833\n",
+      "Epoch 43/50\n",
+      "29/29 [==============================] - 5s 170ms/step - loss: 0.1876 - acc: 0.9185 - val_loss: 0.2944 - val_acc: 0.8333\n",
+      "Epoch 44/50\n",
+      "29/29 [==============================] - 5s 171ms/step - loss: 0.1272 - acc: 0.9601 - val_loss: 0.3339 - val_acc: 0.8333\n",
+      "Epoch 45/50\n",
+      "29/29 [==============================] - 5s 173ms/step - loss: 0.1609 - acc: 0.9323 - val_loss: 0.3098 - val_acc: 0.8500\n",
+      "Epoch 46/50\n",
+      "29/29 [==============================] - 5s 170ms/step - loss: 0.1500 - acc: 0.9472 - val_loss: 0.4263 - val_acc: 0.8333\n",
+      "Epoch 47/50\n",
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.1740 - acc: 0.9311 - val_loss: 0.3486 - val_acc: 0.8333\n",
+      "Epoch 48/50\n",
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.1576 - acc: 0.9395 - val_loss: 0.5968 - val_acc: 0.8333\n",
+      "Epoch 49/50\n",
+      "29/29 [==============================] - 5s 170ms/step - loss: 0.2000 - acc: 0.9099 - val_loss: 0.4477 - val_acc: 0.8167\n",
+      "Epoch 50/50\n",
+      "29/29 [==============================] - 5s 172ms/step - loss: 0.1385 - acc: 0.9443 - val_loss: 0.3172 - val_acc: 0.8667\n"
+     ]
+    }
+   ],
+   "source": [
+    "model.compile(loss='binary_crossentropy',\n",
+    "              optimizer=optimizers.RMSprop(lr=1e-5),\n",
+    "              metrics=['acc'])\n",
+    "\n",
+    "history = model.fit_generator(\n",
+    "      train_generator,\n",
+    "      steps_per_epoch=29,\n",
+    "      epochs=50,\n",
+    "      validation_data=validation_generator,\n",
+    "      validation_steps=3,\n",
+    "    verbose = 1)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 56,
+   "id": "liquid-muslim",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "model.save('covid_fine_tuned.h5')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 57,
+   "id": "julian-impossible",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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\n",
+      "text/plain": [
+       "<Figure size 432x288 with 1 Axes>"
+      ]
+     },
+     "metadata": {
+      "needs_background": "light"
+     },
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "image/png": "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\n",
+      "text/plain": [
+       "<Figure size 432x288 with 1 Axes>"
+      ]
+     },
+     "metadata": {
+      "needs_background": "light"
+     },
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "acc = history.history['acc']\n",
+    "val_acc = history.history['val_acc']\n",
+    "loss = history.history['loss']\n",
+    "val_loss = history.history['val_loss']\n",
+    "\n",
+    "epochs = range(len(acc))\n",
+    "\n",
+    "plt.plot(epochs, acc, 'bo', label='Training acc')\n",
+    "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
+    "plt.title('Training and validation accuracy')\n",
+    "plt.legend()\n",
+    "\n",
+    "plt.figure()\n",
+    "\n",
+    "plt.plot(epochs, loss, 'bo', label='Training loss')\n",
+    "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
+    "plt.title('Training and validation loss')\n",
+    "plt.legend()\n",
+    "\n",
+    "plt.show()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 58,
+   "id": "still-plain",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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\n",
+      "text/plain": [
+       "<Figure size 432x288 with 1 Axes>"
+      ]
+     },
+     "metadata": {
+      "needs_background": "light"
+     },
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "image/png": "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\n",
+      "text/plain": [
+       "<Figure size 432x288 with 1 Axes>"
+      ]
+     },
+     "metadata": {
+      "needs_background": "light"
+     },
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "def smooth_curve(points, factor=0.8):\n",
+    "  smoothed_points = []\n",
+    "  for point in points:\n",
+    "    if smoothed_points:\n",
+    "      previous = smoothed_points[-1]\n",
+    "      smoothed_points.append(previous * factor + point * (1 - factor))\n",
+    "    else:\n",
+    "      smoothed_points.append(point)\n",
+    "  return smoothed_points\n",
+    "\n",
+    "plt.plot(epochs,\n",
+    "         smooth_curve(acc), 'bo', label='Smoothed training acc')\n",
+    "plt.plot(epochs,\n",
+    "         smooth_curve(val_acc), 'b', label='Smoothed validation acc')\n",
+    "plt.title('Training and validation accuracy')\n",
+    "plt.legend()\n",
+    "\n",
+    "plt.figure()\n",
+    "\n",
+    "plt.plot(epochs,\n",
+    "         smooth_curve(loss), 'bo', label='Smoothed training loss')\n",
+    "plt.plot(epochs,\n",
+    "         smooth_curve(val_loss), 'b', label='Smoothed validation loss')\n",
+    "plt.title('Training and validation loss')\n",
+    "plt.legend()\n",
+    "\n",
+    "plt.show()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 60,
+   "id": "central-american",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Found 77 images belonging to 2 classes.\n",
+      "test acc: 0.7922077775001526\n"
+     ]
+    }
+   ],
+   "source": [
+    "test_generator = test_datagen.flow_from_directory(\n",
+    "        test_dir,\n",
+    "        target_size=(218, 178),\n",
+    "        batch_size=20,\n",
+    "        class_mode='binary')\n",
+    "\n",
+    "test_loss, test_acc = model.evaluate_generator(test_generator, steps=4)\n",
+    "print('test acc:', test_acc)"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "played-emerald",
+   "id": "greenhouse-louisville",
    "metadata": {},
    "outputs": [],
    "source": []