tensorflow__resnet.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Aplicação da ResNet 50"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
    "from tensorflow.keras.applications import ResNet50\n",
    "from tensorflow.keras.layers import Dense, GlobalAveragePooling2D\n",
    "from tensorflow.keras.models import Model\n",
    "from tensorflow.keras.optimizers import Adam"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Diretórios de treino e teste\n",
    "train_dir = 'PokemonSplit/train'\n",
    "test_dir = 'PokemonSplit/test'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 5391 images belonging to 150 classes.\n",
      "Found 814 images belonging to 150 classes.\n",
      "Epoch 1/10\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Users\\madus\\AppData\\Local\\Programs\\Python\\Python312\\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",
      "  self._warn_if_super_not_called()\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m169/169\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m415s\u001b[0m 2s/step - accuracy: 0.0079 - loss: 5.1506 - val_accuracy: 0.0221 - val_loss: 4.9445\n",
      "Epoch 2/10\n",
      "\u001b[1m169/169\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m371s\u001b[0m 2s/step - accuracy: 0.0244 - loss: 4.9393 - val_accuracy: 0.0332 - val_loss: 4.8458\n",
      "Epoch 3/10\n",
      "\u001b[1m169/169\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m359s\u001b[0m 2s/step - accuracy: 0.0273 - loss: 4.8415 - val_accuracy: 0.0405 - val_loss: 4.7783\n",
      "Epoch 4/10\n",
      "\u001b[1m169/169\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m353s\u001b[0m 2s/step - accuracy: 0.0287 - loss: 4.7616 - val_accuracy: 0.0418 - val_loss: 4.7338\n",
      "Epoch 5/10\n",
      "\u001b[1m169/169\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m350s\u001b[0m 2s/step - accuracy: 0.0419 - loss: 4.6750 - val_accuracy: 0.0369 - val_loss: 4.6691\n",
      "Epoch 6/10\n",
      "\u001b[1m169/169\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m341s\u001b[0m 2s/step - accuracy: 0.0486 - loss: 4.6306 - val_accuracy: 0.0577 - val_loss: 4.6419\n",
      "Epoch 7/10\n",
      "\u001b[1m169/169\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m384s\u001b[0m 2s/step - accuracy: 0.0580 - loss: 4.5243 - val_accuracy: 0.0627 - val_loss: 4.5835\n",
      "Epoch 8/10\n",
      "\u001b[1m169/169\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m372s\u001b[0m 2s/step - accuracy: 0.0650 - loss: 4.4802 - val_accuracy: 0.0602 - val_loss: 4.5294\n",
      "Epoch 9/10\n",
      "\u001b[1m169/169\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m363s\u001b[0m 2s/step - accuracy: 0.0767 - loss: 4.4228 - val_accuracy: 0.0713 - val_loss: 4.4838\n",
      "Epoch 10/10\n",
      "\u001b[1m169/169\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m368s\u001b[0m 2s/step - accuracy: 0.0804 - loss: 4.3853 - val_accuracy: 0.0713 - val_loss: 4.4757\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.src.callbacks.history.History at 0x238ae37bd40>"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Hiperparâmetros\n",
    "img_height, img_width = 224, 224\n",
    "batch_size = 32\n",
    "epochs = 10\n",
    "\n",
    "# Criação dos geradores de dados\n",
    "train_datagen = ImageDataGenerator(\n",
    "    rescale=1./255,\n",
    "    shear_range=0.2,\n",
    "    zoom_range=0.2,\n",
    "    horizontal_flip=True)\n",
    "\n",
    "test_datagen = ImageDataGenerator(rescale=1./255)\n",
    "\n",
    "train_generator = train_datagen.flow_from_directory(\n",
    "    train_dir,\n",
    "    target_size=(img_height, img_width),\n",
    "    batch_size=batch_size,\n",
    "    class_mode='categorical')\n",
    "\n",
    "test_generator = test_datagen.flow_from_directory(\n",
    "    test_dir,\n",
    "    target_size=(img_height, img_width),\n",
    "    batch_size=batch_size,\n",
    "    class_mode='categorical')\n",
    "\n",
    "# Carregamento do modelo ResNet-50 pré-treinado no ImageNet\n",
    "base_model = ResNet50(weights='imagenet', include_top=False, input_shape=(img_height, img_width, 3))\n",
    "\n",
    "# Adição de camadas customizadas\n",
    "x = base_model.output\n",
    "x = GlobalAveragePooling2D()(x)\n",
    "x = Dense(1024, activation='relu')(x)\n",
    "predictions = Dense(150, activation='softmax')(x)\n",
    "\n",
    "# Criação do modelo final\n",
    "model = Model(inputs=base_model.input, outputs=predictions)\n",
    "\n",
    "# Congelamento das camadas base do ResNet-50\n",
    "for layer in base_model.layers:\n",
    "    layer.trainable = False\n",
    "\n",
    "# Compilação do modelo\n",
    "model.compile(optimizer=Adam(), loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "\n",
    "# Treinamento do modelo\n",
    "model.fit(\n",
    "    train_generator,\n",
    "    epochs=epochs,\n",
    "    validation_data=test_generator)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m26/26\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m47s\u001b[0m 2s/step - accuracy: 0.0767 - loss: 4.4877\n",
      "Test accuracy: 7.13%\n"
     ]
    }
   ],
   "source": [
    "# Avaliação do modelo\n",
    "loss, accuracy = model.evaluate(test_generator)\n",
    "print(f'Test accuracy: {accuracy * 100:.2f}%')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Após treinar o modelo\n",
    "model.save('pokemon_classifier.h5')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Teste com 1 imagem"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow.keras.preprocessing import image\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.models import load_model\n",
    "\n",
    "# Carregar o modelo salvo\n",
    "model = load_model('pokemon_classifier.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Caminho da imagem de exemplo (pikachu)\n",
    "img_path = 'PokemonSplit/test/Pikachu/00000059.jpg'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'model' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "Cell \u001b[1;32mIn[3], line 13\u001b[0m\n\u001b[0;32m     10\u001b[0m img_array \u001b[38;5;241m=\u001b[39m load_and_preprocess_image(img_path)\n\u001b[0;32m     12\u001b[0m \u001b[38;5;66;03m# Fazer a previsão com o modelo treinado\u001b[39;00m\n\u001b[1;32m---> 13\u001b[0m predictions \u001b[38;5;241m=\u001b[39m \u001b[43mmodel\u001b[49m\u001b[38;5;241m.\u001b[39mpredict(img_array)\n\u001b[0;32m     14\u001b[0m predicted_class \u001b[38;5;241m=\u001b[39m np\u001b[38;5;241m.\u001b[39margmax(predictions, axis\u001b[38;5;241m=\u001b[39m\u001b[38;5;241m1\u001b[39m)\n\u001b[0;32m     16\u001b[0m \u001b[38;5;66;03m# Mapeamento de índices de classes para nomes de classes\u001b[39;00m\n",
      "\u001b[1;31mNameError\u001b[0m: name 'model' is not defined"
     ]
    }
   ],
   "source": [
    "# Função para carregar e preparar uma imagem para a previsão\n",
    "def load_and_preprocess_image(img_path, target_size=(224, 224)):\n",
    "    img = image.load_img(img_path, target_size=target_size)\n",
    "    img_array = image.img_to_array(img)\n",
    "    img_array = np.expand_dims(img_array, axis=0)\n",
    "    img_array /= 255.0  # Normalização\n",
    "    return img_array\n",
    "\n",
    "# Carregar e preprocessar a imagem\n",
    "img_array = load_and_preprocess_image(img_path)\n",
    "\n",
    "# Fazer a previsão com o modelo treinado\n",
    "predictions = model.predict(img_array)\n",
    "predicted_class = np.argmax(predictions, axis=1)\n",
    "\n",
    "# Mapeamento de índices de classes para nomes de classes\n",
    "class_indices = train_generator.class_indices\n",
    "class_names = {v: k for k, v in class_indices.items()}\n",
    "\n",
    "# Mostrar a imagem e a previsão\n",
    "plt.imshow(image.load_img(img_path))\n",
    "plt.title(f'Predicted: {class_names[predicted_class[0]]}')\n",
    "plt.axis('off')\n",
    "plt.show()\n",
    "\n",
    "# Avaliação do modelo no conjunto de teste\n",
    "loss, accuracy = model.evaluate(test_generator)\n",
    "print(f'Test accuracy: {accuracy * 100:.2f}%')\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.11.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}