heuzef
/
jan24_cds_mushrooms
1
0
Fork 0
Code Issues 0 Pull Requests 0 Releases 0 Wiki Activity
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
3 Commits
1 Branch
332 MiB
 
 
 
Branch: master
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from 'master'
${ noResults }
jan24_cds_mushrooms/notebooks/transfer_learning_ResNet50/ResNet50_L2.ipynb

404 lines
25 KiB
Raw Permalink Blame History 

  1. {

  2.  "cells": [

  3.   {

  4.    "cell_type": "code",

  5.    "execution_count": 1,

  6.    "metadata": {},

  7.    "outputs": [

  8.     {

  9.      "name": "stdout",

  10.      "output_type": "stream",

  11.      "text": [

  12.       "Found 2199 images belonging to 2 classes.\n",

  13.       "Found 549 images belonging to 2 classes.\n",

  14.       "Training samples: 2199\n",

  15.       "Validation samples: 549\n"

  16.      ]

  17.     }

  18.    ],

  19.    "source": [

  20.     "import tensorflow as tf\n",

  21.     "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",

  22.     "\n",

  23.     "dataset_dir = 'C:/Users/vsavelev/GITHUB/DS_projet/jan24_cds_mushrooms/data'\n",

  24.     "\n",

  25.     "# Create ImageDataGenerator with validation split\n",

  26.     "datagen = ImageDataGenerator(rescale=1.0/255, validation_split=0.2)\n",

  27.     "\n",

  28.     "train_generator = datagen.flow_from_directory(\n",

  29.     "    dataset_dir,\n",

  30.     "    target_size=(224, 224),\n",

  31.     "    batch_size=32,\n",

  32.     "    class_mode='categorical',\n",

  33.     "    subset='training'  # Set as training data\n",

  34.     ")\n",

  35.     "\n",

  36.     "validation_generator = datagen.flow_from_directory(\n",

  37.     "    dataset_dir,\n",

  38.     "    target_size=(224, 224),\n",

  39.     "    batch_size=32,\n",

  40.     "    class_mode='categorical',\n",

  41.     "    subset='validation'  # Set as validation data\n",

  42.     ")\n",

  43.     "\n",

  44.     "print(f'Training samples: {train_generator.samples}')\n",

  45.     "print(f'Validation samples: {validation_generator.samples}')"

  46.    ]

  47.   },

  48.   {

  49.    "cell_type": "code",

  50.    "execution_count": 2,

  51.    "metadata": {},

  52.    "outputs": [

  53.     {

  54.      "data": {

  55.       "text/html": [

  56.        "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"sequential\"</span>\n",

  57.        "</pre>\n"

  58.       ],

  59.       "text/plain": [

  60.        "\u001b[1mModel: \"sequential\"\u001b[0m\n"

  61.       ]

  62.      },

  63.      "metadata": {},

  64.      "output_type": "display_data"

  65.     },

  66.     {

  67.      "data": {

  68.       "text/html": [

  69.        "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",

  70.        "┃<span style=\"font-weight: bold\"> Layer (type)                    </span>┃<span style=\"font-weight: bold\"> Output Shape           </span>┃<span style=\"font-weight: bold\">       Param # </span>┃\n",

  71.        "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",

  72.        "│ resnet50 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Functional</span>)           │ ?                      │    <span style=\"color: #00af00; text-decoration-color: #00af00\">23,587,712</span> │\n",

  73.        "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",

  74.        "│ global_average_pooling2d        │ ?                      │   <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (unbuilt) │\n",

  75.        "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">GlobalAveragePooling2D</span>)        │                        │               │\n",

  76.        "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",

  77.        "│ dense (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                   │ ?                      │   <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (unbuilt) │\n",

  78.        "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",

  79.        "│ dropout (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dropout</span>)               │ ?                      │   <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (unbuilt) │\n",

  80.        "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",

  81.        "│ dense_1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                 │ ?                      │   <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (unbuilt) │\n",

  82.        "└─────────────────────────────────┴────────────────────────┴───────────────┘\n",

  83.        "</pre>\n"

  84.       ],

  85.       "text/plain": [

  86.        "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",

  87.        "┃\u001b[1m \u001b[0m\u001b[1mLayer (type)                   \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape          \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m      Param #\u001b[0m\u001b[1m \u001b[0m┃\n",

  88.        "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",

  89.        "│ resnet50 (\u001b[38;5;33mFunctional\u001b[0m)           │ ?                      │    \u001b[38;5;34m23,587,712\u001b[0m │\n",

  90.        "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",

  91.        "│ global_average_pooling2d        │ ?                      │   \u001b[38;5;34m0\u001b[0m (unbuilt) │\n",

  92.        "│ (\u001b[38;5;33mGlobalAveragePooling2D\u001b[0m)        │                        │               │\n",

  93.        "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",

  94.        "│ dense (\u001b[38;5;33mDense\u001b[0m)                   │ ?                      │   \u001b[38;5;34m0\u001b[0m (unbuilt) │\n",

  95.        "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",

  96.        "│ dropout (\u001b[38;5;33mDropout\u001b[0m)               │ ?                      │   \u001b[38;5;34m0\u001b[0m (unbuilt) │\n",

  97.        "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",

  98.        "│ dense_1 (\u001b[38;5;33mDense\u001b[0m)                 │ ?                      │   \u001b[38;5;34m0\u001b[0m (unbuilt) │\n",

  99.        "└─────────────────────────────────┴────────────────────────┴───────────────┘\n"

  100.       ]

  101.      },

  102.      "metadata": {},

  103.      "output_type": "display_data"

  104.     },

  105.     {

  106.      "data": {

  107.       "text/html": [

  108.        "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">23,587,712</span> (89.98 MB)\n",

  109.        "</pre>\n"

  110.       ],

  111.       "text/plain": [

  112.        "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m23,587,712\u001b[0m (89.98 MB)\n"

  113.       ]

  114.      },

  115.      "metadata": {},

  116.      "output_type": "display_data"

  117.     },

  118.     {

  119.      "data": {

  120.       "text/html": [

  121.        "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (0.00 B)\n",

  122.        "</pre>\n"

  123.       ],

  124.       "text/plain": [

  125.        "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"

  126.       ]

  127.      },

  128.      "metadata": {},

  129.      "output_type": "display_data"

  130.     },

  131.     {

  132.      "data": {

  133.       "text/html": [

  134.        "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">23,587,712</span> (89.98 MB)\n",

  135.        "</pre>\n"

  136.       ],

  137.       "text/plain": [

  138.        "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m23,587,712\u001b[0m (89.98 MB)\n"

  139.       ]

  140.      },

  141.      "metadata": {},

  142.      "output_type": "display_data"

  143.     }

  144.    ],

  145.    "source": [

  146.     "from tensorflow.keras.applications import ResNet50\n",

  147.     "from tensorflow.keras import layers, models\n",

  148.     "\n",

  149.     "# Load and Configure the Pre-trained ResNet50 Model\n",

  150.     "base_model = ResNet50(weights='imagenet', include_top=False, input_shape=(224, 224, 3))\n",

  151.     "\"\"\"\n",

  152.     "weights='imagenet': Loads the pre-trained weights from the ImageNet dataset.\n",

  153.     "include_top=False: Excludes the top fully-connected layers of the ResNet50 model, enabling you to add your own custom layers.\n",

  154.     "input_shape=(224, 224, 3): Specifies the input shape of the images (224x224 pixels, with 3 color channels - RGB).\n",

  155.     "\"\"\"\n",

  156.     "\n",

  157.     "# Freeze the base model (to freeze the pre-trained layers)\n",

  158.     "base_model.trainable = False\n",

  159.     "\n",

  160.     "# Add custom layers on top of the base model\n",

  161.     "model = models.Sequential([   #allows to stack layers linearly\n",

  162.     "    base_model,\n",

  163.     "    layers.GlobalAveragePooling2D(),\n",

  164.     "    layers.Dense(1024, activation='relu', kernel_regularizer=tf.keras.regularizers.l2(0.01)),  # L2 regularization\n",

  165.     "    layers.Dropout(0.5),\n",

  166.     "    layers.Dense(train_generator.num_classes, activation='softmax')\n",

  167.     "])\n",

  168.     "\n",

  169.     "\"\"\"\n",

  170.     "GlobalAveragePooling2D(): Reduces each feature map to a single number by taking the average, \n",

  171.     "which helps to reduce the size of the model and prevent overfitting.\n",

  172.     "Dense(1024, activation='relu'): Adds a fully connected layer with 1024 units and ReLU activation function.\n",

  173.     "Dropout(0.5): Adds a dropout layer with a 50% dropout rate to prevent overfitting by randomly setting half of the input units \n",

  174.     "to 0 at each update during training.\n",

  175.     "Dense(train_generator.num_classes, activation='softmax'): \n",

  176.     "Adds the final output layer with units equal to the number of classes in your dataset, using the softmax activation function for multi-class classification.\n",

  177.     "\"\"\"\n",

  178.     "model.compile(optimizer=tf.keras.optimizers.Adam(),\n",

  179.     "              loss='categorical_crossentropy',\n",

  180.     "              metrics=['accuracy'])\n",

  181.     "\n",

  182.     "\"\"\"\n",

  183.     "optimizer=tf.keras.optimizers.Adam(): Uses the Adam optimizer, which is an adaptive learning rate optimization algorithm.\n",

  184.     "loss='categorical_crossentropy': Uses categorical cross-entropy as the loss function, suitable for multi-class classification.\n",

  185.     "metrics=['accuracy']: Tracks accuracy as the metric to evaluate the model's performance during training and testing.\n",

  186.     "\"\"\"\n",

  187.     "model.summary()\n",

  188.     "\n"

  189.    ]

  190.   },

  191.   {

  192.    "cell_type": "code",

  193.    "execution_count": 3,

  194.    "metadata": {},

  195.    "outputs": [

  196.     {

  197.      "name": "stdout",

  198.      "output_type": "stream",

  199.      "text": [

  200.       "Epoch 1/10\n"

  201.      ]

  202.     },

  203.     {

  204.      "name": "stderr",

  205.      "output_type": "stream",

  206.      "text": [

  207.       "c:\\Users\\vsavelev\\AppData\\Local\\anaconda3\\Lib\\site-packages\\keras\\src\\trainers\\data_adapters\\py_dataset_adapter.py:121: UserWarning: Your `PyDataset` class should call `super().__init__(**kwargs)` in its constructor. `**kwargs` can include `workers`, `use_multiprocessing`, `max_queue_size`. Do not pass these arguments to `fit()`, as they will be ignored.\n",

  208.       "  self._warn_if_super_not_called()\n"

  209.      ]

  210.     },

  211.     {

  212.      "name": "stdout",

  213.      "output_type": "stream",

  214.      "text": [

  215.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m143s\u001b[0m 2s/step - accuracy: 0.8540 - loss: 6.8043 - val_accuracy: 0.8860 - val_loss: 0.6636\n",

  216.       "Epoch 2/10\n",

  217.       "\u001b[1m 1/68\u001b[0m \u001b[37m━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[1m2:09\u001b[0m 2s/step - accuracy: 0.8438 - loss: 0.8550"

  218.      ]

  219.     },

  220.     {

  221.      "name": "stderr",

  222.      "output_type": "stream",

  223.      "text": [

  224.       "c:\\Users\\vsavelev\\AppData\\Local\\anaconda3\\Lib\\contextlib.py:158: UserWarning: Your input ran out of data; interrupting training. Make sure that your dataset or generator can generate at least `steps_per_epoch * epochs` batches. You may need to use the `.repeat()` function when building your dataset.\n",

  225.       "  self.gen.throw(typ, value, traceback)\n"

  226.      ]

  227.     },

  228.     {

  229.      "name": "stdout",

  230.      "output_type": "stream",

  231.      "text": [

  232.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 11ms/step - accuracy: 0.8438 - loss: 0.8550 - val_accuracy: 1.0000 - val_loss: 0.4526\n",

  233.       "Epoch 3/10\n",

  234.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m146s\u001b[0m 2s/step - accuracy: 0.8873 - loss: 0.6507 - val_accuracy: 0.8879 - val_loss: 0.4938\n",

  235.       "Epoch 4/10\n",

  236.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 6ms/step - accuracy: 0.8750 - loss: 0.5112 - val_accuracy: 0.8000 - val_loss: 0.6924\n",

  237.       "Epoch 5/10\n",

  238.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m149s\u001b[0m 2s/step - accuracy: 0.8857 - loss: 0.5040 - val_accuracy: 0.8879 - val_loss: 0.4471\n",

  239.       "Epoch 6/10\n",

  240.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 6ms/step - accuracy: 0.8750 - loss: 0.4322 - val_accuracy: 0.8000 - val_loss: 0.6413\n",

  241.       "Epoch 7/10\n",

  242.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m166s\u001b[0m 2s/step - accuracy: 0.8926 - loss: 0.4378 - val_accuracy: 0.8860 - val_loss: 0.4396\n",

  243.       "Epoch 8/10\n",

  244.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 7ms/step - accuracy: 0.9375 - loss: 0.3477 - val_accuracy: 1.0000 - val_loss: 0.2697\n",

  245.       "Epoch 9/10\n",

  246.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m173s\u001b[0m 3s/step - accuracy: 0.8890 - loss: 0.4132 - val_accuracy: 0.8879 - val_loss: 0.4444\n",

  247.       "Epoch 10/10\n",

  248.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 7ms/step - accuracy: 0.9688 - loss: 0.3492 - val_accuracy: 0.8000 - val_loss: 0.5367\n"

  249.      ]

  250.     }

  251.    ],

  252.    "source": [

  253.     "history = model.fit(\n",

  254.     "    train_generator,\n",

  255.     "    steps_per_epoch=train_generator.samples // train_generator.batch_size,\n",

  256.     "    validation_data=validation_generator,\n",

  257.     "    validation_steps=validation_generator.samples // validation_generator.batch_size,\n",

  258.     "    epochs=10\n",

  259.     ")\n",

  260.     "\n",

  261.     "#This specifies the number of complete passes through the training dataset. Here, the model will train for 10 epochs."

  262.    ]

  263.   },

  264.   {

  265.    "cell_type": "code",

  266.    "execution_count": 12,

  267.    "metadata": {},

  268.    "outputs": [

  269.     {

  270.      "name": "stdout",

  271.      "output_type": "stream",

  272.      "text": [

  273.       "Model compiled successfully.\n",

  274.       "Callbacks created successfully.\n",

  275.       "Epoch 1/20\n",

  276.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m267s\u001b[0m 4s/step - accuracy: 0.8686 - loss: 0.6860 - val_accuracy: 0.8897 - val_loss: 0.5719\n",

  277.       "Epoch 2/20\n",

  278.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 9ms/step - accuracy: 0.8750 - loss: 0.6368 - val_accuracy: 0.6000 - val_loss: 0.7365\n",

  279.       "Epoch 3/20\n",

  280.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m286s\u001b[0m 4s/step - accuracy: 0.8856 - loss: 0.6046 - val_accuracy: 0.8897 - val_loss: 0.6145\n",

  281.       "Epoch 4/20\n",

  282.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 9ms/step - accuracy: 0.8438 - loss: 0.5484 - val_accuracy: 0.6000 - val_loss: 0.7114\n",

  283.       "Epoch 5/20\n",

  284.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m323s\u001b[0m 5s/step - accuracy: 0.8840 - loss: 0.5082 - val_accuracy: 0.8860 - val_loss: 0.5360\n",

  285.       "Epoch 6/20\n",

  286.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 7ms/step - accuracy: 0.8750 - loss: 0.4655 - val_accuracy: 1.0000 - val_loss: 0.4314\n",

  287.       "Epoch 7/20\n",

  288.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m299s\u001b[0m 4s/step - accuracy: 0.8770 - loss: 0.4487 - val_accuracy: 0.8879 - val_loss: 0.4546\n",

  289.       "Epoch 8/20\n",

  290.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 6ms/step - accuracy: 0.9375 - loss: 0.3386 - val_accuracy: 0.8000 - val_loss: 0.5507\n",

  291.       "Epoch 9/20\n",

  292.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m276s\u001b[0m 4s/step - accuracy: 0.8946 - loss: 0.3906 - val_accuracy: 0.8879 - val_loss: 0.4199\n",

  293.       "Epoch 10/20\n",

  294.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 7ms/step - accuracy: 0.8438 - loss: 0.4413 - val_accuracy: 0.8000 - val_loss: 0.5716\n",

  295.       "Epoch 11/20\n",

  296.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m293s\u001b[0m 4s/step - accuracy: 0.8840 - loss: 0.3867 - val_accuracy: 0.8860 - val_loss: 0.4153\n",

  297.       "Epoch 12/20\n",

  298.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 7ms/step - accuracy: 0.9565 - loss: 0.2604 - val_accuracy: 1.0000 - val_loss: 0.2164\n",

  299.       "Epoch 13/20\n",

  300.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m290s\u001b[0m 4s/step - accuracy: 0.8951 - loss: 0.3485 - val_accuracy: 0.8915 - val_loss: 0.3989\n",

  301.       "Epoch 14/20\n",

  302.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 7ms/step - accuracy: 0.9375 - loss: 0.2921 - val_accuracy: 0.4000 - val_loss: 1.3309\n",

  303.       "Epoch 15/20\n",

  304.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m273s\u001b[0m 4s/step - accuracy: 0.8950 - loss: 0.3413 - val_accuracy: 0.8860 - val_loss: 0.4095\n",

  305.       "Epoch 16/20\n",

  306.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 7ms/step - accuracy: 0.9062 - loss: 0.3136 - val_accuracy: 1.0000 - val_loss: 0.1990\n",

  307.       "Epoch 17/20\n",

  308.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m270s\u001b[0m 4s/step - accuracy: 0.8919 - loss: 0.3295 - val_accuracy: 0.8860 - val_loss: 0.4054\n",

  309.       "Epoch 18/20\n",

  310.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 8ms/step - accuracy: 0.8750 - loss: 0.3683 - val_accuracy: 1.0000 - val_loss: 0.1749\n",

  311.       "Epoch 19/20\n",

  312.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m279s\u001b[0m 4s/step - accuracy: 0.8906 - loss: 0.3170 - val_accuracy: 0.8860 - val_loss: 0.4067\n",

  313.       "Epoch 20/20\n",

  314.       "\u001b[1m68/68\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 8ms/step - accuracy: 0.9375 - loss: 0.2421 - val_accuracy: 1.0000 - val_loss: 0.2140\n"

  315.      ]

  316.     }

  317.    ],

  318.    "source": [

  319.     "from tensorflow.keras.callbacks import EarlyStopping, ModelCheckpoint\n",

  320.     "\n",

  321.     "\n",

  322.     "# Unfreeze some layers\n",

  323.     "base_model.trainable = True\n",

  324.     "fine_tune_at = 100  # fine-tune from this layer onwards\n",

  325.     "\n",

  326.     "for layer in base_model.layers[:fine_tune_at]:\n",

  327.     "    layer.trainable = False\n",

  328.     "\n",

  329.     "#  # Unfreeze more layers gradually\n",

  330.     "# for layer in base_model.layers[:-10]:  # Unfreeze all layers except the last 10 layers\n",

  331.     "#     layer.trainable = False\n",

  332.     "\n",

  333.     "model.compile(optimizer=tf.keras.optimizers.Adam(learning_rate=1e-6), #learning_rate=1e-5\n",

  334.     "              loss='categorical_crossentropy', #The categorical cross-entropy loss function is used because this is a multi-class classification problem\n",

  335.     "              metrics=['accuracy'])\n",

  336.     "\n",

  337.     "print(\"Model compiled successfully.\")\n",

  338.     "\n",

  339.     "\"\"\"\n",

  340.     "The Adam optimizer is used with a very small learning rate (1e-5). Fine-tuning typically \n",

  341.     "uses a smaller learning rate to prevent large updates to the weights, which could potentially destroy the learned features in the pre-trained model.\n",

  342.     "\"\"\"\n",

  343.     "\n",

  344.     "early_stopping = EarlyStopping(monitor='val_loss', patience=5, restore_best_weights=True)\n",

  345.     "model_checkpoint = ModelCheckpoint('best_model.keras', save_best_only=True, monitor='val_loss')\n",

  346.     "\n",

  347.     "print(\"Callbacks created successfully.\")\n",

  348.     "\n",

  349.     "\n",

  350.     "history_fine = model.fit(\n",

  351.     "    train_generator,\n",

  352.     "    steps_per_epoch=train_generator.samples // train_generator.batch_size,\n",

  353.     "    validation_data=validation_generator,\n",

  354.     "    validation_steps=validation_generator.samples // validation_generator.batch_size,\n",

  355.     "    epochs=20,\n",

  356.     "    #callbacks=[early_stopping, model_checkpoint]\n",

  357.     ")\n"

  358.    ]

  359.   },

  360.   {

  361.    "cell_type": "code",

  362.    "execution_count": 13,

  363.    "metadata": {},

  364.    "outputs": [

  365.     {

  366.      "name": "stdout",

  367.      "output_type": "stream",

  368.      "text": [

  369.       "\u001b[1m18/18\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m22s\u001b[0m 1s/step - accuracy: 0.8704 - loss: 0.4312\n",

  370.       "Validation loss: 0.4050602614879608\n",

  371.       "Validation accuracy: 0.8870673775672913\n"

  372.      ]

  373.     }

  374.    ],

  375.    "source": [

  376.     "loss, accuracy = model.evaluate(validation_generator)\n",

  377.     "print(f'Validation loss: {loss}')\n",

  378.     "print(f'Validation accuracy: {accuracy}')\n"

  379.    ]

  380.   }

  381.  ],

  382.  "metadata": {

  383.   "kernelspec": {

  384.    "display_name": "base",

  385.    "language": "python",

  386.    "name": "python3"

  387.   },

  388.   "language_info": {

  389.    "codemirror_mode": {

  390.     "name": "ipython",

  391.     "version": 3

  392.    },

  393.    "file_extension": ".py",

  394.    "mimetype": "text/x-python",

  395.    "name": "python",

  396.    "nbconvert_exporter": "python",

  397.    "pygments_lexer": "ipython3",

  398.    "version": "3.11.7"

  399.   }

  400.  },

  401.  "nbformat": 4,

  402.  "nbformat_minor": 2

  403. }