aptos-model-statistics.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "_cell_guid": "79c7e3d0-c299-4dcb-8224-4455121ee9b0",
    "_uuid": "d629ff2d2480ee46fbb7e2d37f6b5fab8052498a"
   },
   "outputs": [],
   "source": [
    "import os\n",
    "import zipfile\n",
    "import numpy as np\n",
    "import sys\n",
    "import pandas as pd\n",
    "import tensorflow as tf\n",
    "from keras_preprocessing.image import load_img\n",
    "from keras_preprocessing.image import img_to_array\n",
    "from kaggle_datasets import KaggleDatasets\n",
    "from numpy import save\n",
    "from numpy import asarray\n",
    "from os import listdir\n",
    "import matplotlib as mpl\n",
    "from numpy import load\n",
    "from tensorflow.keras.optimizers import RMSprop\n",
    "import matplotlib.pyplot as plt\n",
    "import random\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "validation_df=pd.read_csv('/kaggle/input/bestmodelvalidationused/validation.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "gs://kds-db6211cd4531736bf26a8f5404ff6a6e65a1dfc68ccc6145be1375fd\n"
     ]
    }
   ],
   "source": [
    "path_for_gcs=KaggleDatasets().get_gcs_path('aptos2019-blindness-detection')\n",
    "print(path_for_gcs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "validation_files_path=[path_for_gcs+'/train_images/'+ fname for fname in validation_df['id_code']]\n",
    "validation_labels=list(validation_df['diagnosis'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Not Running on TPU\n",
      "REPLICAS  1\n"
     ]
    }
   ],
   "source": [
    "#TPU setup\n",
    "try:\n",
    "    tpu=tf.distribute.cluster_resolver.TPUClusterResolver()\n",
    "    print(\"Running on TPU\")\n",
    "except ValueError:\n",
    "    tpu=None\n",
    "if tpu:\n",
    "    tf.config.experimental_connect_to_cluster(tpu)\n",
    "    tf.tpu.experimental.initialize_tpu_system(tpu)\n",
    "    tpu_strategy=tf.distribute.experimental.TPUStrategy(tpu)\n",
    "else:\n",
    "    print(\"Not Running on TPU\")\n",
    "    tpu_strategy=tf.distribute.get_strategy()\n",
    "print(\"REPLICAS \",tpu_strategy.num_replicas_in_sync)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "#predefined variable\n",
    "IMG_WIDTH=512\n",
    "IMG_HEIGHT=512\n",
    "BATCH_SIZE=16*tpu_strategy.num_replicas_in_sync#take thumb rule\n",
    "AUTOTUNE = tf.data.experimental.AUTOTUNE "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def parse_function_for_validate(filename,label):\n",
    "    image_string=tf.io.read_file(filename)\n",
    "    image_decoded=tf.image.decode_png(image_string,channels=3)\n",
    "    image_resized=tf.image.resize(image_decoded,[IMG_WIDTH,IMG_HEIGHT])\n",
    "    image_normalized=image_resized/255.0\n",
    "    label=tf.dtypes.cast(label,tf.int32)\n",
    "    label=tf.one_hot(label,5)\n",
    "    return image_normalized,label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_dataset(filenames, labels, is_training=True):\n",
    "    dataset = tf.data.Dataset.from_tensor_slices((filenames, labels))\n",
    "    if is_training:\n",
    "        dataset = dataset.map(parse_function_for_train, num_parallel_calls=AUTOTUNE)\n",
    "    else:\n",
    "        dataset = dataset.map(parse_function_for_validate, num_parallel_calls=AUTOTUNE)\n",
    "    dataset = dataset.prefetch(buffer_size=AUTOTUNE)\n",
    "    dataset = dataset.batch(BATCH_SIZE)\n",
    "    return dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "validation_dataset=create_dataset(validation_files_path,validation_labels,is_training=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "!pip install -q efficientnet\n",
    "import efficientnet.tfkeras as efn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "metrices=[tf.keras.metrics.CategoricalAccuracy(name='acc')]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "efficientnet-b7 (Model)      (None, 16, 16, 2560)      64097680  \n",
      "_________________________________________________________________\n",
      "global_average_pooling2d (Gl (None, 2560)              0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 5)                 12805     \n",
      "=================================================================\n",
      "Total params: 64,110,485\n",
      "Trainable params: 63,799,765\n",
      "Non-trainable params: 310,720\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "with tpu_strategy.scope():\n",
    "    enet = efn.EfficientNetB7(\n",
    "        input_shape=(IMG_WIDTH, IMG_HEIGHT, 3),\n",
    "        weights=None,#'imagenet if training for first time'\n",
    "        include_top=False\n",
    "    )\n",
    "    enet.trainable = True \n",
    "    model = tf.keras.Sequential([\n",
    "        enet,\n",
    "        tf.keras.layers.GlobalAveragePooling2D(),\n",
    "        tf.keras.layers.Dense(5, activation='softmax')\n",
    "    ])\n",
    "    model.compile(\n",
    "        optimizer=tf.keras.optimizers.Adam(1e-5),\n",
    "        loss = 'categorical_crossentropy',\n",
    "        metrics=metrices\n",
    "    )\n",
    "    model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.load_weights('/kaggle/input/projectfinal2/efficientnetb7epochs36weightswithoversamplingverticalimgaugpart1.hdf5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "history1to10=np.load('/kaggle/input/histories/history1to10.npy',allow_pickle=True)\n",
    "history11to20=np.load('/kaggle/input/histories/history10to20.npy',allow_pickle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([{'loss': [0.7429718223970327], 'acc': [0.7285448], 'val_loss': [2.3325783014297485], 'val_acc': [0.5652174]},\n",
       "       {'loss': [0.2829801651524074], 'acc': [0.90298504], 'val_loss': [2.1125738620758057], 'val_acc': [0.47826087]},\n",
       "       {'loss': [0.16766206564298317], 'acc': [0.94962686], 'val_loss': [1.4791942834854126], 'val_acc': [0.6956522]},\n",
       "       {'loss': [0.10276847057489318], 'acc': [0.9785448], 'val_loss': [1.0975428223609924], 'val_acc': [0.6956522]},\n",
       "       {'loss': [0.08571873897158389], 'acc': [0.9692164], 'val_loss': [1.3228583931922913], 'val_acc': [0.86956525]},\n",
       "       {'loss': [0.08110816198498455], 'acc': [0.9757463], 'val_loss': [1.127720832824707], 'val_acc': [0.86956525]},\n",
       "       {'loss': [0.0722345094015794], 'acc': [0.98600745], 'val_loss': [1.5155670046806335], 'val_acc': [0.7826087]},\n",
       "       {'loss': [0.0604201923611004], 'acc': [0.98134327], 'val_loss': [1.1462824940681458], 'val_acc': [0.826087]},\n",
       "       {'loss': [0.06378606799989939], 'acc': [0.98134327], 'val_loss': [1.3474608659744263], 'val_acc': [0.7826087]},\n",
       "       {'loss': [0.05742654251629737], 'acc': [0.9850746], 'val_loss': [1.7078197002410889], 'val_acc': [0.73913044]}],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "history1to10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([{'loss': [0.6259168341978273], 'acc': [0.7835821], 'val_loss': [0.9490711987018585], 'val_acc': [0.73913044]},\n",
       "       {'loss': [0.2883383865231898], 'acc': [0.9048507], 'val_loss': [0.9239094257354736], 'val_acc': [0.826087]},\n",
       "       {'loss': [0.20106189064125515], 'acc': [0.9393656], 'val_loss': [1.2790570855140686], 'val_acc': [0.7826087]},\n",
       "       {'loss': [0.1525687783845325], 'acc': [0.95522386], 'val_loss': [0.9420528411865234], 'val_acc': [0.7826087]},\n",
       "       {'loss': [0.12183386184719962], 'acc': [0.95988804], 'val_loss': [1.1153591871261597], 'val_acc': [0.826087]},\n",
       "       {'loss': [0.10094482340474627], 'acc': [0.9720149], 'val_loss': [0.9824252724647522], 'val_acc': [0.7826087]},\n",
       "       {'loss': [0.08775933881971373], 'acc': [0.9766791], 'val_loss': [1.1176316142082214], 'val_acc': [0.7826087]},\n",
       "       {'loss': [0.08796753801071822], 'acc': [0.9766791], 'val_loss': [1.5668790936470032], 'val_acc': [0.826087]},\n",
       "       {'loss': [0.08804098399939822], 'acc': [0.9776119], 'val_loss': [1.176470309495926], 'val_acc': [0.7826087]},\n",
       "       {'loss': [0.07805884648948463], 'acc': [0.9766791], 'val_loss': [0.9303404092788696], 'val_acc': [0.86956525]}],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "history11to20"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot(history1,history2):\n",
    "    acc=[]\n",
    "    val_acc=[]\n",
    "    loss=[]\n",
    "    val_loss=[]\n",
    "    for hist in history1:\n",
    "        acc.append(hist['acc'])\n",
    "        loss.append(hist['loss'])\n",
    "        val_acc.append(hist['val_acc'])\n",
    "        val_loss.append(hist['val_loss'])\n",
    "    for hist in history2:\n",
    "        acc.append(hist['acc'])\n",
    "        loss.append(hist['loss'])\n",
    "        val_acc.append(hist['val_acc'])\n",
    "        val_loss.append(hist['val_loss'])\n",
    "    epochs=range(len(acc))\n",
    "    plt.plot(epochs,acc)\n",
    "    plt.plot(epochs,val_acc)\n",
    "    plt.title('Training accuracy and Validation accuracy')\n",
    "    plt.figure()\n",
    "    plt.plot(epochs,loss)\n",
    "    plt.plot(epochs,val_loss)\n",
    "    plt.title('Training and validation loss')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "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": [
    "plot(history1to10,history11to20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import classification_report"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions=model.predict(validation_dataset)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions=list(np.argmax(predictions,1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "import seaborn as sn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plotconfusionmatrix(y_actual,y_pred):\n",
    "    data = {'y_Actual':    y_actual,'y_Predicted': y_pred}\n",
    "    df = pd.DataFrame(data, columns=['y_Actual','y_Predicted'])\n",
    "    confusion_matrix = pd.crosstab(df['y_Actual'], df['y_Predicted'], rownames=['Actual'], colnames=['Predicted'])\n",
    "    sn.heatmap(confusion_matrix, annot=True)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plotconfusionmatrix(validation_labels,predictions)"
   ]
  }
 ],
 "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}