leaf_classification.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### The notebook was run on colab and the dataset was imported directly from google drive"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Link to dataset: https://drive.google.com/file/d/1SKzdTgInW7VUQuuFkjhCKjfOesRIcRbo/view"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "0QvPbnCa1GAd",
    "outputId": "f95a26fe-cef6-407b-adb4-d033cb3a2d5f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
     ]
    }
   ],
   "source": [
    "from google.colab import drive\n",
    "drive.mount('/content/drive')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 136
    },
    "colab_type": "code",
    "id": "91jK-_ta1VhW",
    "outputId": "45b9884e-00e4-41a3-892b-d5af64a86275"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "all_images_train\tleafnetmodeltest.json  one_hot_labels_test.csv\n",
      "leafnetmodel2.h5\tleafnetmodelvgg.h5     one_hot_labels_train.csv\n",
      "leafnetmodel2.json\tleafnetmodelvgg.json   test_images_arrays.npz\n",
      "leafnetmodel.h5\t\tmodel2.h5\t       train\n",
      "leafnetmodel.json\tmodel2.json\t       train_images_arrays.npz\n",
      "leafnetmodeltestaug.h5\tmodel.h5\n",
      "leafnetmodeltest.h5\tmodel.json\n"
     ]
    }
   ],
   "source": [
    "!ls \"/content/drive/My Drive/leaf_classification/dataset/\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "48yAvRFu1gEp"
   },
   "outputs": [],
   "source": [
    "path=\"/content/drive/My Drive/leaf_classification/dataset/\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Import necessary libraries and keras functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "pFzng0Ca1kD4",
    "outputId": "c88cc89d-e730-4fcc-a985-8e4a87f0daf2"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import PIL\n",
    "import keras\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, Flatten\n",
    "from keras.layers import Conv2D, MaxPooling2D\n",
    "from keras.utils import to_categorical\n",
    "from keras.models import Sequential, Model, load_model\n",
    "from keras.layers import Input, Conv1D, Conv2D, MaxPooling1D, MaxPooling2D, Dense, Dropout, Activation, Flatten\n",
    "from keras.layers.normalization import BatchNormalization\n",
    "from keras.optimizers import SGD\n",
    "from keras.callbacks import ModelCheckpoint,ReduceLROnPlateau\n",
    "from keras import regularizers\n",
    "from keras.callbacks import ModelCheckpoint\n",
    "from keras.models import model_from_json\n",
    "from keras.utils import to_categorical\n",
    "from keras.preprocessing import image\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.model_selection import train_test_split\n",
    "from keras.utils import to_categorical\n",
    "from tqdm import tqdm\n",
    "from keras.constraints import maxnorm\n",
    "from keras.optimizers import SGD\n",
    "from keras.layers import Activation\n",
    "from keras.layers.convolutional import Conv2D, MaxPooling2D, ZeroPadding2D\n",
    "from keras.layers.normalization import BatchNormalization\n",
    "from keras.initializers import glorot_normal\n",
    "from keras.utils import np_utils\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Import the pregenerated labels for train and test dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 270
    },
    "colab_type": "code",
    "id": "yDozBX4S4oNT",
    "outputId": "2b9a6471-d8f2-48c8-921b-59ffed7d77c4"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>filename</th>\n",
       "      <th>abies_concolor</th>\n",
       "      <th>abies_nordmanniana</th>\n",
       "      <th>acer_campestre</th>\n",
       "      <th>acer_ginnala</th>\n",
       "      <th>acer_griseum</th>\n",
       "      <th>acer_negundo</th>\n",
       "      <th>acer_palmatum</th>\n",
       "      <th>acer_pensylvanicum</th>\n",
       "      <th>acer_platanoides</th>\n",
       "      <th>acer_pseudoplatanus</th>\n",
       "      <th>acer_rubrum</th>\n",
       "      <th>acer_saccharinum</th>\n",
       "      <th>acer_saccharum</th>\n",
       "      <th>aesculus_flava</th>\n",
       "      <th>aesculus_glabra</th>\n",
       "      <th>aesculus_hippocastamon</th>\n",
       "      <th>aesculus_pavi</th>\n",
       "      <th>ailanthus_altissima</th>\n",
       "      <th>albizia_julibrissin</th>\n",
       "      <th>amelanchier_arborea</th>\n",
       "      <th>amelanchier_canadensis</th>\n",
       "      <th>amelanchier_laevis</th>\n",
       "      <th>asimina_triloba</th>\n",
       "      <th>betula_alleghaniensis</th>\n",
       "      <th>betula_jacqemontii</th>\n",
       "      <th>betula_lenta</th>\n",
       "      <th>betula_nigra</th>\n",
       "      <th>betula_populifolia</th>\n",
       "      <th>broussonettia_papyrifera</th>\n",
       "      <th>carpinus_betulus</th>\n",
       "      <th>carpinus_caroliniana</th>\n",
       "      <th>carya_cordiformis</th>\n",
       "      <th>carya_glabra</th>\n",
       "      <th>carya_ovata</th>\n",
       "      <th>carya_tomentosa</th>\n",
       "      <th>castanea_dentata</th>\n",
       "      <th>catalpa_bignonioides</th>\n",
       "      <th>catalpa_speciosa</th>\n",
       "      <th>cedrus_atlantica</th>\n",
       "      <th>...</th>\n",
       "      <th>quercus_imbricaria</th>\n",
       "      <th>quercus_macrocarpa</th>\n",
       "      <th>quercus_marilandica</th>\n",
       "      <th>quercus_michauxii</th>\n",
       "      <th>quercus_montana</th>\n",
       "      <th>quercus_muehlenbergii</th>\n",
       "      <th>quercus_nigra</th>\n",
       "      <th>quercus_palustris</th>\n",
       "      <th>quercus_phellos</th>\n",
       "      <th>quercus_robur</th>\n",
       "      <th>quercus_rubra</th>\n",
       "      <th>quercus_shumardii</th>\n",
       "      <th>quercus_stellata</th>\n",
       "      <th>quercus_velutina</th>\n",
       "      <th>quercus_virginiana</th>\n",
       "      <th>robinia_pseudo-acacia</th>\n",
       "      <th>salix_babylonica</th>\n",
       "      <th>salix_caroliniana</th>\n",
       "      <th>salix_matsudana</th>\n",
       "      <th>salix_nigra</th>\n",
       "      <th>sassafras_albidum</th>\n",
       "      <th>staphylea_trifolia</th>\n",
       "      <th>stewartia_pseudocamellia</th>\n",
       "      <th>styrax_japonica</th>\n",
       "      <th>styrax_obassia</th>\n",
       "      <th>syringa_reticulata</th>\n",
       "      <th>taxodium_distichum</th>\n",
       "      <th>tilia_americana</th>\n",
       "      <th>tilia_cordata</th>\n",
       "      <th>tilia_europaea</th>\n",
       "      <th>tilia_tomentosa</th>\n",
       "      <th>toona_sinensis</th>\n",
       "      <th>tsuga_canadensis</th>\n",
       "      <th>ulmus_americana</th>\n",
       "      <th>ulmus_glabra</th>\n",
       "      <th>ulmus_parvifolia</th>\n",
       "      <th>ulmus_procera</th>\n",
       "      <th>ulmus_pumila</th>\n",
       "      <th>ulmus_rubra</th>\n",
       "      <th>zelkova_serrata</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 186 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   filename  abies_concolor  ...  ulmus_rubra  zelkova_serrata\n",
       "0         0               0  ...            0                0\n",
       "1         1               0  ...            0                0\n",
       "2         2               0  ...            0                0\n",
       "3         3               0  ...            0                0\n",
       "4         4               0  ...            0                0\n",
       "\n",
       "[5 rows x 186 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df=pd.read_csv(\"/content/drive/My Drive/leaf_classification/dataset/one_hot_labels_train.csv\")\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 270
    },
    "colab_type": "code",
    "id": "IIoN_OtDHMkl",
    "outputId": "8daa5c07-e457-47e9-c8bf-a624b40b8a36"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>filename</th>\n",
       "      <th>abies_concolor</th>\n",
       "      <th>abies_nordmanniana</th>\n",
       "      <th>acer_campestre</th>\n",
       "      <th>acer_ginnala</th>\n",
       "      <th>acer_griseum</th>\n",
       "      <th>acer_negundo</th>\n",
       "      <th>acer_palmatum</th>\n",
       "      <th>acer_pensylvanicum</th>\n",
       "      <th>acer_platanoides</th>\n",
       "      <th>acer_pseudoplatanus</th>\n",
       "      <th>acer_rubrum</th>\n",
       "      <th>acer_saccharinum</th>\n",
       "      <th>acer_saccharum</th>\n",
       "      <th>aesculus_flava</th>\n",
       "      <th>aesculus_glabra</th>\n",
       "      <th>aesculus_hippocastamon</th>\n",
       "      <th>aesculus_pavi</th>\n",
       "      <th>ailanthus_altissima</th>\n",
       "      <th>albizia_julibrissin</th>\n",
       "      <th>amelanchier_arborea</th>\n",
       "      <th>amelanchier_canadensis</th>\n",
       "      <th>amelanchier_laevis</th>\n",
       "      <th>asimina_triloba</th>\n",
       "      <th>betula_alleghaniensis</th>\n",
       "      <th>betula_jacqemontii</th>\n",
       "      <th>betula_lenta</th>\n",
       "      <th>betula_nigra</th>\n",
       "      <th>betula_populifolia</th>\n",
       "      <th>broussonettia_papyrifera</th>\n",
       "      <th>carpinus_betulus</th>\n",
       "      <th>carpinus_caroliniana</th>\n",
       "      <th>carya_cordiformis</th>\n",
       "      <th>carya_glabra</th>\n",
       "      <th>carya_ovata</th>\n",
       "      <th>carya_tomentosa</th>\n",
       "      <th>castanea_dentata</th>\n",
       "      <th>catalpa_bignonioides</th>\n",
       "      <th>catalpa_speciosa</th>\n",
       "      <th>cedrus_atlantica</th>\n",
       "      <th>...</th>\n",
       "      <th>quercus_imbricaria</th>\n",
       "      <th>quercus_macrocarpa</th>\n",
       "      <th>quercus_marilandica</th>\n",
       "      <th>quercus_michauxii</th>\n",
       "      <th>quercus_montana</th>\n",
       "      <th>quercus_muehlenbergii</th>\n",
       "      <th>quercus_nigra</th>\n",
       "      <th>quercus_palustris</th>\n",
       "      <th>quercus_phellos</th>\n",
       "      <th>quercus_robur</th>\n",
       "      <th>quercus_rubra</th>\n",
       "      <th>quercus_shumardii</th>\n",
       "      <th>quercus_stellata</th>\n",
       "      <th>quercus_velutina</th>\n",
       "      <th>quercus_virginiana</th>\n",
       "      <th>robinia_pseudo-acacia</th>\n",
       "      <th>salix_babylonica</th>\n",
       "      <th>salix_caroliniana</th>\n",
       "      <th>salix_matsudana</th>\n",
       "      <th>salix_nigra</th>\n",
       "      <th>sassafras_albidum</th>\n",
       "      <th>staphylea_trifolia</th>\n",
       "      <th>stewartia_pseudocamellia</th>\n",
       "      <th>styrax_japonica</th>\n",
       "      <th>styrax_obassia</th>\n",
       "      <th>syringa_reticulata</th>\n",
       "      <th>taxodium_distichum</th>\n",
       "      <th>tilia_americana</th>\n",
       "      <th>tilia_cordata</th>\n",
       "      <th>tilia_europaea</th>\n",
       "      <th>tilia_tomentosa</th>\n",
       "      <th>toona_sinensis</th>\n",
       "      <th>tsuga_canadensis</th>\n",
       "      <th>ulmus_americana</th>\n",
       "      <th>ulmus_glabra</th>\n",
       "      <th>ulmus_parvifolia</th>\n",
       "      <th>ulmus_procera</th>\n",
       "      <th>ulmus_pumila</th>\n",
       "      <th>ulmus_rubra</th>\n",
       "      <th>zelkova_serrata</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 186 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   filename  abies_concolor  ...  ulmus_rubra  zelkova_serrata\n",
       "0         0               0  ...            0                0\n",
       "1         1               0  ...            0                0\n",
       "2         2               0  ...            0                0\n",
       "3         3               0  ...            0                0\n",
       "4         4               0  ...            0                0\n",
       "\n",
       "[5 rows x 186 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_df=pd.read_csv(\"/content/drive/My Drive/leaf_classification/dataset/one_hot_labels_test.csv\")\n",
    "test_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "aH67x30MHWZ7"
   },
   "outputs": [],
   "source": [
    "y_test=test_df[test_df.columns[1:]]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### All images were stacked and converted to numpy array and stored for easy uploading of dataset to drive"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "hCIyq22X7xgD"
   },
   "outputs": [],
   "source": [
    "x_npz = np.load(path+\"train_images_arrays.npz\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "bnz6qB52GfCl"
   },
   "outputs": [],
   "source": [
    "test_npz=np.load(path+\"test_images_arrays.npz\")\n",
    "X_test=test_npz['arr_0']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "e0M740dBzwAe"
   },
   "outputs": [],
   "source": [
    "X=x_npz['arr_0']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "3XpKLgVZ7zQd",
    "outputId": "de81280e-2ed0-40c5-88f9-8e328e875ecb"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(22224, 64, 64, 3) (22224, 185) (2470, 64, 64, 3) (2470, 185)\n"
     ]
    }
   ],
   "source": [
    "X_train, X_val, y_train, y_val = train_test_split(X, df[df.columns[1:]], test_size=0.1, random_state=420)\n",
    "print(X_train.shape,y_train.shape,X_val.shape,y_val.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Preprocessing and Data Augmentation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "zA915dNMKC9r"
   },
   "outputs": [],
   "source": [
    "from keras.preprocessing.image import ImageDataGenerator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "SA1f6qFZg3He"
   },
   "outputs": [],
   "source": [
    "def preproc(image):\n",
    "    kernel = np.array([[-1,-1,-1,-1,-1],\n",
    "                       [-1,2,2,2,-1],\n",
    "                       [-1,2,8,2,-1],\n",
    "                       [-1,2,2,2,-1],\n",
    "                       [-1,-1,-1,-1,-1]])*(1/8)\n",
    "    \n",
    "    sharpened =  cv2.filter2D(image,ddepth = -1,kernel = kernel)\n",
    "    gblur = cv2.GaussianBlur(sharpened,(5,5),0)\n",
    "    return gblur"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Some data augmentation techniques were not found to perform poorly and hence only flips, rotations and normalization was used"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "_5L31PZgKE76"
   },
   "outputs": [],
   "source": [
    "#datagen = ImageDataGenerator(horizontal_flip=True,vertical_flip=True,rotation_range=90,\n",
    "#                                                               zoom_range=0.2)\n",
    "datagen = ImageDataGenerator(horizontal_flip=True,vertical_flip=True,rotation_range=90,samplewise_center=True,samplewise_std_normalization=True)\n",
    "#datagen = ImageDataGenerator(rotation_range=90,horizontal_flip=True,vertical_flip=True,preprocessing_function=preproc,)\n",
    "datagen.fit(X_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "qcCQGkErKmgC"
   },
   "source": [
    "## Modelling"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### We tried various models given below and found that a basic 3 layer convolutional network gave the best tradeoff between accuracy and training time"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### No transfer learning is used and all training is done from scratch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "NrZEfJM6pI-d"
   },
   "outputs": [],
   "source": [
    "#leafnetmodel\n",
    "input_shape = (X_train.shape[1], X_train.shape[2], X_train.shape[3]) #(64, 64, 3)\n",
    "num_classes = 185\n",
    "\n",
    "model = Sequential()\n",
    "model.add(Conv2D(32, kernel_size=(5, 5), strides=(1, 1), input_shape=input_shape, activation='relu'))\n",
    "#model.add(BatchNormalization())\n",
    "model.add(MaxPooling2D(pool_size=(2, 2), strides=(2, 2)))\n",
    "model.add(Conv2D(64, (5, 5), activation='relu'))\n",
    "#model.add(BatchNormalization())\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "model.add(Dropout(0.7))\n",
    "model.add(Flatten())\n",
    "model.add(Dense(1000, activation='relu'))\n",
    "#model.add(BatchNormalization())\n",
    "model.add(Dropout(0.7))\n",
    "model.add(Dense(num_classes, activation='softmax'))\n",
    "\n",
    "\n",
    "model.compile(optimizer=keras.optimizers.Adam(lr=0.0003, beta_1=0.9, beta_2=0.999, epsilon=None, decay=1e-8, amsgrad=False),\n",
    "              loss='categorical_crossentropy',\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "FMlk1s4-Kt4N"
   },
   "outputs": [],
   "source": [
    "# #leafnetmodel test\n",
    "# input_shape = (X_train.shape[1], X_train.shape[2], X_train.shape[3]) #(64, 64, 3)\n",
    "# num_classes = 185\n",
    "\n",
    "# model = Sequential()\n",
    "# model.add(Conv2D(32, kernel_size=(5, 5), strides=(1, 1), input_shape=input_shape, activation='relu'))\n",
    "# model.add(BatchNormalization())\n",
    "# model.add(MaxPooling2D(pool_size=(2, 2), strides=(2, 2)))\n",
    "# model.add(Conv2D(64, (5, 5), activation='relu'))\n",
    "# model.add(BatchNormalization())\n",
    "# model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "# model.add(Dropout(0.7))\n",
    "# model.add(Conv2D(128, (5, 5), activation='relu'))\n",
    "# model.add(BatchNormalization())\n",
    "# model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "# model.add(Dropout(0.7))\n",
    "# model.add(Flatten())\n",
    "# model.add(Dense(1000, activation='relu'))\n",
    "# model.add(BatchNormalization())\n",
    "# model.add(Dropout(0.7))\n",
    "# model.add(Dense(num_classes, activation='softmax'))\n",
    "\n",
    "\n",
    "# model.compile(optimizer=keras.optimizers.Adam(lr=0.0003, beta_1=0.9, beta_2=0.999, epsilon=None, decay=1e-8, amsgrad=False),\n",
    "#               loss='categorical_crossentropy',\n",
    "#               metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "nh9qUUS31-N_"
   },
   "outputs": [],
   "source": [
    "# #input_shape = (X_train.shape[1], X_train.shape[2], X_train.shape[3]) #(64, 64, 3)\n",
    "# input_shape=(224,224,3)\n",
    "# num_classes = 185\n",
    "\n",
    "# model = Sequential()\n",
    "# model.add(Conv2D(32, kernel_size=(5, 5), strides=(1, 1), input_shape=input_shape))\n",
    "# model.add(Activation('relu'))\n",
    "# model.add(MaxPooling2D(pool_size=(2, 2), strides=(2, 2)))\n",
    "# model.add(Conv2D(64, (5, 5)))\n",
    "# model.add(Activation('relu'))\n",
    "# model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "# model.add(Dropout(0.7))\n",
    "# model.add(Flatten())\n",
    "# model.add(Dense(1000))\n",
    "# model.add(Activation('relu'))\n",
    "# model.add(Dropout(0.7))\n",
    "# model.add(Dense(num_classes, activation='softmax'))\n",
    "\n",
    "\n",
    "# model.compile(optimizer=keras.optimizers.Adam(lr=0.0003, beta_1=0.9, beta_2=0.999, epsilon=None, decay=1e-8, amsgrad=False),\n",
    "#               loss='categorical_crossentropy',\n",
    "#               metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "yiPbRX3xc4xD"
   },
   "outputs": [],
   "source": [
    "# num_classes=185\n",
    "# model = Sequential()\n",
    "\n",
    "# model.add(ZeroPadding2D((1, 1), input_shape=(64,64,3)))\n",
    "# model.add(Conv2D(64, (3, 3)))\n",
    "# model.add(Activation('relu'))\n",
    "# model.add(ZeroPadding2D((1, 1)))\n",
    "# model.add(Conv2D(64, (3, 3)))\n",
    "# model.add(Activation('relu'))\n",
    "# model.add(MaxPooling2D(pool_size=(2, 2), strides=(2, 2)))\n",
    "\n",
    "# model.add(ZeroPadding2D((1, 1)))\n",
    "# model.add(Conv2D(128, (3, 3)))\n",
    "# model.add(Activation('relu'))\n",
    "# model.add(ZeroPadding2D((1, 1)))\n",
    "# model.add(Conv2D(128, (3, 3)))\n",
    "# model.add(Activation('relu'))\n",
    "# model.add(MaxPooling2D(pool_size=(2, 2), strides=(2, 2)))\n",
    "\n",
    "# model.add(ZeroPadding2D((1, 1)))\n",
    "# model.add(Conv2D(256, (3, 3)))\n",
    "# model.add(Activation('relu'))\n",
    "# model.add(ZeroPadding2D((1, 1)))\n",
    "# model.add(Conv2D(256, (3, 3)))\n",
    "# model.add(Activation('relu'))\n",
    "# model.add(ZeroPadding2D((1, 1)))\n",
    "# model.add(Conv2D(256, (3, 3)))\n",
    "# model.add(Activation('relu'))\n",
    "# model.add(MaxPooling2D(pool_size=(2, 2), strides=(2, 2)))\n",
    "\n",
    "# model.add(ZeroPadding2D((1, 1)))\n",
    "# model.add(Conv2D(512, (3, 3)))\n",
    "# model.add(Activation('relu'))\n",
    "# model.add(ZeroPadding2D((1, 1)))\n",
    "# model.add(Conv2D(512, (3, 3)))\n",
    "# model.add(Activation('relu'))\n",
    "# model.add(ZeroPadding2D((1, 1)))\n",
    "# model.add(Conv2D(512, (3, 3)))\n",
    "# model.add(Activation('relu'))\n",
    "# model.add(MaxPooling2D(pool_size=(2, 2), strides=(2, 2)))\n",
    "\n",
    "# model.add(ZeroPadding2D((1, 1)))\n",
    "# model.add(Conv2D(512, (3, 3)))\n",
    "# model.add(Activation('relu'))\n",
    "# model.add(ZeroPadding2D((1, 1)))\n",
    "# model.add(Conv2D(512, (3, 3)))\n",
    "# model.add(Activation('relu'))\n",
    "# model.add(ZeroPadding2D((1, 1)))\n",
    "# model.add(Conv2D(512, (3, 3)))\n",
    "# model.add(Activation('relu'))\n",
    "# model.add(MaxPooling2D(pool_size=(2, 2), strides=(2, 2)))\n",
    "\n",
    "# model.add(Flatten())\n",
    "# model.add(Dense(4096))\n",
    "# model.add(Activation('relu'))\n",
    "# model.add(Dropout(0.5))\n",
    "# model.add(Dense(4096))\n",
    "# model.add(Activation('relu'))\n",
    "# model.add(Dropout(0.5))\n",
    "# model.add(Dense(num_classes))\n",
    "# model.add(Activation('softmax'))\n",
    "# sgd = SGD(lr=0.1, decay=1e-6, nesterov=True)\n",
    "# model.compile(loss='categorical_crossentropy', optimizer=sgd, metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 459
    },
    "colab_type": "code",
    "id": "LFAdEIbd466y",
    "outputId": "d68c5244-b6ce-4fd7-bf1e-ffacddb99735"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_4\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_9 (Conv2D)            (None, 60, 60, 32)        2432      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_9 (MaxPooling2 (None, 30, 30, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_10 (Conv2D)           (None, 26, 26, 64)        51264     \n",
      "_________________________________________________________________\n",
      "max_pooling2d_10 (MaxPooling (None, 13, 13, 64)        0         \n",
      "_________________________________________________________________\n",
      "dropout_9 (Dropout)          (None, 13, 13, 64)        0         \n",
      "_________________________________________________________________\n",
      "flatten_4 (Flatten)          (None, 10816)             0         \n",
      "_________________________________________________________________\n",
      "dense_7 (Dense)              (None, 1000)              10817000  \n",
      "_________________________________________________________________\n",
      "dropout_10 (Dropout)         (None, 1000)              0         \n",
      "_________________________________________________________________\n",
      "dense_8 (Dense)              (None, 185)               185185    \n",
      "=================================================================\n",
      "Total params: 11,055,881\n",
      "Trainable params: 11,055,881\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Mz0Yx2qNMbJv"
   },
   "outputs": [],
   "source": [
    "model_json = model.to_json()\n",
    "with open(\"/content/drive/My Drive/leaf_classification/dataset/leafnetmodel2.json\", \"w\") as json_file:\n",
    "    json_file.write(model_json)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "sOKVaHCXjsjX"
   },
   "outputs": [],
   "source": [
    "model=load_model(path+\"leafnetmodeltestaug3.h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "colab_type": "code",
    "id": "4YsDGSvVlBrL",
    "outputId": "5d6b444e-5ac4-4596-ca1f-d199931b1a4d"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2470/2470 [==============================] - 1s 597us/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.15802784422751864, 0.9445344129554656]"
      ]
     },
     "execution_count": 97,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.evaluate(X_val,y_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "colab_type": "code",
    "id": "1yGQA8-9HCSb",
    "outputId": "2ac27c06-4ea2-42db-ea3f-ac5ccd031fef"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3090/3090 [==============================] - 1s 238us/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.08239633732484382, 0.9809061488673139]"
      ]
     },
     "execution_count": 98,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.evaluate(X_test,y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "SF-0J074FLMS"
   },
   "source": [
    "The various accuracies we got with different models\n",
    "leafnetmodel 0.81\n",
    "leafnetmodel2 0.83\n",
    "leafnetmodeltest 0.78\n",
    "leafnetmodeltestaug 0.84 0.91 0.98\n",
    "leafnetmodeltestaug2 0.88 Done val\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "uElvZcjRtQJl",
    "outputId": "a8068823-ca9d-4f29-fff7-c2c4cef46b09"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mounted at /content/drive\n"
     ]
    }
   ],
   "source": [
    "from google.colab import drive\n",
    "drive.mount('/content/drive')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training code"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### We used keras callbacks to save the best model in case of overfitting\n",
    "#### Learning rate decay is also used with a patience of 5 epochs"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### The model was repeatedly trained for a total of more than 400 epochs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "id": "o2jwmzJfp5P6",
    "outputId": "3004bd46-23f9-40af-c1a2-fb6c84653425"
   },
   "outputs": [],
   "source": [
    "lrr = ReduceLROnPlateau(monitor='val_acc', \n",
    "                        patience=5, \n",
    "                        verbose=1, \n",
    "                        factor=0.4,\n",
    "                        min_lr=0.00001)\n",
    "\n",
    "# autosave best Model\n",
    "best_model_file = path+\"leafnetmodeltestaug4.h5\"\n",
    "best_model = ModelCheckpoint(best_model_file, monitor='val_loss', verbose=1, save_best_only=True)\n",
    "\n",
    "print('Training model...')\n",
    "results = model.fit(X_test, y_test, epochs=50, batch_size=64, validation_data=(X_val, y_val), callbacks=[best_model])\n",
    "#model.fit_generator(datagen.flow(X_train, y_train, batch_size=64),steps_per_epoch=len(X_train) / 64, epochs=100, callbacks=[best_model],validation_data=(X_val, y_val))\n",
    "print('Traing finished.')\n",
    "\n",
    "\n",
    "\n",
    "print('Loading the best model...')\n",
    "model = load_model(best_model_file)\n",
    "print('Best Model loaded!')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## We got a final test accuracy of 88% which is pretty good considering the minimal preprocessing, noisy training data and the large number of target classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 136
    },
    "colab_type": "code",
    "id": "hByd5TuM81Bj",
    "outputId": "951e3798-f7b9-4827-a6ff-9cb0e043ffd6"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[12, 10,  0, ...,  0,  0,  0],\n",
       "       [ 0,  0,  0, ...,  0,  0,  0],\n",
       "       [ 0,  0,  4, ...,  0,  0,  0],\n",
       "       ...,\n",
       "       [ 0,  0,  0, ..., 15,  0,  2],\n",
       "       [ 0,  0,  0, ...,  0,  6,  0],\n",
       "       [ 0,  0,  0, ...,  0,  0,  4]])"
      ]
     },
     "execution_count": 27,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "+from sklearn.metrics import confusion_matrix\n",
    "yv=model.predict(X_val)\n",
    "cnf_matrix = confusion_matrix(np.argmax(yv,axis=-1), np.argmax(np.array(y_val),axis=-1))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 373
    },
    "colab_type": "code",
    "id": "3McsoLBxFc6w",
    "outputId": "5799023f-06aa-498c-f5ce-64c32f41e9a5"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x1440 with 0 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXMAAAFTCAYAAADLFuykAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import seaborn as sns\n",
    "plt.figure(figsize = (20,20))\n",
    "fig, ax = plt.subplots(1)\n",
    "\n",
    "ax = sns.heatmap(cnf_matrix[:10,:10], ax=ax, cmap=plt.cm.Greens, annot=True)\n",
    "ax.set_xticklabels(df.columns[1:11])\n",
    "ax.set_yticklabels(df.columns[1:11])\n",
    "plt.title('Confusion Matrix')\n",
    "plt.ylabel('True class')\n",
    "plt.xlabel('Predicted class')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "Cwe5foF1aPiZ",
    "outputId": "b2d9316c-610c-4e84-9c05-9d0eae2f855a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(2470,) (2470, 185)\n"
     ]
    }
   ],
   "source": [
    "y_prob = model.predict(X_val) \n",
    "y_classes = y_prob.argmax(axis=-1)\n",
    "y_truths = np.array(y_val).argmax(axis=-1)\n",
    "print(y_classes.shape,y_val.shape)\n",
    "incorrects = np.nonzero(y_classes.reshape((-1,)) != y_truths.reshape((-1,)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 286
    },
    "colab_type": "code",
    "id": "7lu8y74icDY4",
    "outputId": "e95db7bf-b373-40bd-cddb-538ea9e20658"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7facf21c0278>"
      ]
     },
     "execution_count": 48,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP4AAAD8CAYAAABXXhlaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAFYpJREFUeJzt3W2MHdV9BvDn2bveF7+sX/BiHAwY\nsAuhUjHRAgZTkphA3UADHxAKjRKrteQvtCJqqhRSKW1akOBLCB+qUKvQuBEJEAI1QhRijN2oqmtY\ngh2MjcG4pvbWy66xjQ32vv/74c7eOWe69+7svTP32j7PT1rtmZlzZ87ee/8758w5c4ZmBhEJS1Oj\nCyAi9afAFwmQAl8kQAp8kQAp8EUCpMAXCZACXyRANQU+yVUk95DcS/K+rAolIvlitQN4SBYAvAfg\nZgAHAbwB4G4z25Vd8UQkD801vPYaAHvNbB8AkHwKwO0Aygb+/PnzbfHixTUcUkQq2b9/Pw4fPszJ\n8tUS+OcDOOAsHwRwbaUXLF68GN3d3TUcUkQq6erqSpUv94t7JNeS7CbZ3d/fn/fhRCSFWgK/B8AF\nzvKiaJ3HzNaZWZeZdXV2dtZwOBHJSi2B/waApSQvJtkC4OsAXsimWCKSp6rb+GY2QvLPALwCoADg\nCTN7J7OSiUhuarm4BzN7CcBLGZVFROpEI/dEAqTAFwmQAl8kQAp8kQAp8EUCpMAXCZACXyRACnyR\nACnwRQKkwBcJkAJfJEAKfJEAKfBFAqTAFwmQAl8kQAp8kQAp8EUCpMAXCZACXyRACnyRACnwRQKk\nwBcJkAJfJEAKfJEAKfBFAjRp4JN8gmQfyZ3OunkkN5J8P/o9N99iikiW0pzxfwJgVWLdfQA2mdlS\nAJuiZRE5Q0wa+Gb2awBHEqtvB7A+Sq8HcEfG5RKRHFXbxl9gZoeidC+ABRmVR0TqoOaLe2ZmAKzc\ndpJrSXaT7O7v76/1cCKSgWoD/yOSCwEg+t1XLqOZrTOzLjPr6uzsrPJwIpKlagP/BQCro/RqABuy\nKY6I1EOa7ryfA9gK4DKSB0muAfAQgJtJvg/gK9GyiJwhmifLYGZ3l9l0U8ZlEZE60cg9kQAp8EUC\npMAXCZACXyRACnyRACnwRQKkwBcJkAJfJEAKfJEAKfBFAqTAFwmQAl8kQAp8kQAp8EUCpMAXCZAC\nXyRACnyRACnwRQKkwBcJkAJfJEAKfJEAKfBFAqTAFwmQAl8kQAp8kQCleYTWBSQ3k9xF8h2S90br\n55HcSPL96Pfc/IsrIllIc8YfAfAdM7sCwHIA95C8AsB9ADaZ2VIAm6JlETkDTBr4ZnbIzH4TpU8A\n2A3gfAC3A1gfZVsP4I68Ciki2ZpSG5/kYgBXAdgGYIGZHYo29QJYkGnJRCQ3qQOf5EwAvwTwbTM7\n7m4zMwNgZV63lmQ3ye7+/v6aCisi2UgV+CSnoRj0T5rZc9Hqj0gujLYvBNA30WvNbJ2ZdZlZV2dn\nZxZlFpEapbmqTwCPA9htZj90Nr0AYHWUXg1gQ/bFE5E8NKfIswLANwG8TXJ7tO57AB4C8AzJNQA+\nBHBXPkUUkaxNGvhm9h8AWGbzTdkWR0TqQSP3RAKkwBcJkAJfJEAKfJEAKfBFAqTAFwmQAl8kQGkG\n8MgkxsbGvOWmpqay20ZHR0vpQqFQ9nWu4eFhb3natGlVlVNknM74IgFS4IsESFX9DJSrok+0rXgH\nc1Hx/qfJpc0nkpbO+CIBUuCLBEiBLxIgtfHrTO11OR3ojC8SIAW+SIBU1c+A20WXx+vUPJCs6Ywv\nEiAFvkiAVNXPQKUqe7KaXmmUX7lRfcmbeURqpTO+SIAU+CIBUuCLBEht/AxkcQeeSD2leXZeG8nX\nSe4g+Q7JH0TrLya5jeRekk+TbMm/uCKShTRV/UEAK83sSgDLAKwiuRzAwwAeMbMlAI4CWJNfMUUk\nS5MGvhV9Gi1Oi34MwEoAz0br1wO4I5cSnoFIln4qMTPvZ3h4eMIfkaylurhHshA9KbcPwEYAHwA4\nZmYjUZaDAM7Pp4gikrVUgW9mo2a2DMAiANcAuDztAUiuJdlNsru/v7/KYopIlqbUnWdmxwBsBnAd\ngDkkx3sFFgHoKfOadWbWZWZdnZ2dNRVWRLIxaXceyU4Aw2Z2jGQ7gJtRvLC3GcCdAJ4CsBrAhjwL\neiZJzqXvctv9yfb74OBgqv21trbWUDqRdP34CwGsJ1lAsYbwjJm9SHIXgKdIPgDgLQCP51hOEcnQ\npIFvZr8FcNUE6/eh2N4XkTOMRu7loNIdeK7kXXfNzfHHMTQ0VEp/+umnXj53m/sawH+8VnJbOe5j\nvZLLTYV4XFalmwTHSh080evo3rEYN1WGRga8fIVm5y5E+AdIPJjMSSX/rvh1R4994m2ZO2duKe3+\nmc2B3/CosfoiAVLgiwRIVf0MJK+6u1fuK43eS1b129raSmm3yp7c/6lTp0rpZE/AwEBclXaPnXzC\nrtszkGwSuOVyK+zDw36TYHgkPjbpb2tti5sIAwNxU8Xg52tpnl5Kj8JvLhScr+fQyJDzmuTXNn5/\n3Ko9AAwOxj0nra16yvA4nfFFAqTAFwmQAl8kQGrjn0bcNnmlrjh3W3L0n9vGd7v9kvnc5eRIQO96\nAAvO+kIiX5uz5Lfd3Xb3od54NPemTa96uXbt2llKX3TRRd6266+/oZS+tmv5hPsGgFFn4pMC/XOZ\n26532/ttgbf3dcYXCZACXyRAqupnIO1IPaD6x22VO16ymt7SEnejuVV9N51cPn78uLfNHbk3/9xz\n4uMmvi5Hjsa3WX/22Qlv28ZX/62Ufv7550rpDz5438t3zvy4++21zX4zYMeOHfH+v/VZKb3yizd5\n+cbG4vd0YMgfGTijvaOUbg67du/RGV8kQAp8kQAp8EUCpDZ+zpJtene52sdku23w5LBfN6/b/k9e\nC3Db+Mlhv25X37vv7iqlL11yiZfv5ZdfKqX/5adPeNu2bv3PUnrp71xcSp+/aIGX7/Dh+DrBjBnt\n3rZ398Rdfd//m++V0uf9o3+sKy773VK6qc2f5X1gKL420Nbi7z9kOuOLBEiBLxIgVfVzkPax2ckq\nfNrHbbl36yWPlXYfbrefm07uc+68WaX0s8894+V78MG/L6Xbp/v7+PznL4vL1BQ3TY4c+djL19Ye\nv250zB9dOGN6fOyenv8tpR944O+8fD/76VOldFPi729uds9ttXelni10xhcJkAJfJECq6mcg7XTa\nEy1XI+1ceq60zY/ksjk3xBQKfj53Io557bO8bUPD8Qi6T0/EIwNHRv0RhG3t8ci93t5eb9vnPhf/\nnUuWXFpK//uvN3v5XnFGCa5cebO3bVpT3JsxOBxPYNI2LexhfDrjiwRIgS8SIAW+SIDUxj+NlGuH\nV9tll/ZYlUYXjiJux588edLLd+JEfEdeR8cMv4xOF547sYd7XQAAPvkkngd/0aJF3ra+vsOl9NBg\nPBHn0qWXevkee+yxUvoPvvJVb9unJ+OJPmdOnwkpSn3Gjx6V/RbJF6Pli0luI7mX5NMkWybbh4ic\nHqZS1b8XwG5n+WEAj5jZEgBHAazJsmAikp9UVX2SiwDcCuBBAH/BYl1zJYA/jrKsB/C3AH6cQxlP\ne1OZiKOSclX4LKr21e5zdDT+22bN8rvsvPn3E82Fo0eOldLnnRc/Hj35qK3R0Xi0XnJCEPc5A+57\n/Mkn/qQfR4/Ey30f93nbzj3nPGdJl7TGpX0nfgTgu4hnOTwHwDGz0qd4EMD5GZdNRHIyaeCTvA1A\nn5m9Wc0BSK4l2U2yu7+/f/IXiEju0pzxVwD4Gsn9AJ5CsYr/KIA5JMebCosA9Ez0YjNbZ2ZdZtbV\n2dk5URYRqbNJ2/hmdj+A+wGA5JcA/KWZfYPkLwDcieI/g9UANuRYTmkYd2JPfyKLQlPcTXfypD/J\nZcesOaX0yEjctXfqpN+dN3fe7FK6r89vn8+dE0/0+dlncVfidGcCTQAYHov3v2P72962G2+M92EW\nX9dob/EnJglNLVc7/grFC317UWzzP55NkUQkb1MawGNmWwBsidL7AFyTfZFEJG8auScVNRfi6vx1\n163wtq1aFY+S27LlNW/bwEB8Fx4ZV6tHR/07GQ/3Hy2lZ3fM87bNnh03F/Z98D+ldBPbvHwjw3FX\n4tat27xt119/Yyk9o10j98apY1MkQAp8kQCpqi8VEXFVv2PmXG/bDSu+WEq//fY73rZ9+94rpY87\n37KZM6d7+dypwnsOfuRtO3H8lJPPmQdwznwv3/z58ei8lml+zwO9r7i+7uN0xhcJkAJfJEAKfJEA\nqdEjU+CfJ771zT8tpW+55RZv29b/ih+htWVL/Pjrj/oOefl6e+PlFSsu9Lade278uK0V1/9+Kb18\n+Q1evvMWLCylx8b8Mhaa4jsIh4birsTWlrDPeWH/9SKBUuCLBEhVfUntuNO9BgAdHXHX2XkL/Pny\nbrv1j0rpW2+9tZRuSTwTYGQsHuHX3ORvO3Isnsxj1qz4Zh5LPMbArd67VXsAcHoL0RJ49d6ld0Ik\nQAp8kQAp8EUCpDa+VObMoTlzhj8c1m0/Dw35E2ycGoivB8yb606c4U/KOTwc76S51Z+hvaMjvjuv\n2Zn0AwX/a1vhsYAYOBVP7llwXtfePlHucOiMLxIgBb5IgFTVl9SSjw8YcabIb2/357Dzl+P+t+Mn\njnn5Opy5+odG/EdotzTH+xhxJvBo9nvsMOg8Xqut1f9Kz5ihr/hEdMYXCZACXyRAqgdJRZWetOU8\nBHcS8fll9qx5ZXO1Npff4bRC2U1ob9PXeKp0xhcJkAJfJEAKfJEAKfBFApTqqkj0wMwTAEYBjJhZ\nF8l5AJ4GsBjAfgB3mdnRcvsQkdPHVM74XzazZWbWFS3fB2CTmS0FsClaFpEzQC1V/dsBrI/S6wHc\nUXtxRKQe0ga+AfgVyTdJro3WLTCz8ZkSewEsmPilInK6STvy4QYz6yF5LoCNJN91N5qZkZzw5sjo\nH8VaALjwwgsnyiIidZbqjG9mPdHvPgDPo/h47I9ILgSA6HdfmdeuM7MuM+vq7OzMptQiUpNJA5/k\nDJKzxtMAbgGwE8ALAFZH2VYD2JBXIUUkW2mq+gsAPM/ioO1mAD8zs5dJvgHgGZJrAHwI4K78iiki\nWZo08M1sH4ArJ1j/MYCb8iiUiORLI/dEAqTAFwmQAl8kQAp8kQAp8EUCpMAXCZACXyRACnyRACnw\nRQKkwBcJkAJfJEAKfJEAKfBFAqTAFwlQ8A8dGx0d9ZaHhuJHNTc3+2/PtDIPixsbG5tw/URY6WF0\nKZlNOMvZ/9t/FsdKvj9pj5XFsU8Xld7vtPmaks8YT/G65Pq05UhDZ3yRACnwRQJ01lb1jx8/7i27\nVS23yt7a2urla29vL7tPt0rvNgna2tqqLmc1qqlGV1tNLBTi51NPZR9u3rSvq1QdbqS073elfNW8\n/3k2l07Pd1pEcqXAFwmQAl8kQGdtG7+jo6PstuHh4VI62faq1K5y26Buu34q3S7VbEuuL9etWOl1\nyS7HSm3wtO3ztPt3t2XZJVWLSt2USVl0502ly3dcNW38kZGRVPl0xhcJkAJfJEBnbVW/p6fHWz5w\n4EAp3dvbW0onq3xudb7SNrfb78SJE2XLMZVqdNoq9sDAQKrjVapip61+u1XHqZS33P4r7aOa6nC1\n0laJk6ptxqX929KOvCy3rb+/P9VxUp3xSc4h+SzJd0nuJnkdyXkkN5J8P/o9N9URRaTh0lb1HwXw\nspldjuLjtHYDuA/AJjNbCmBTtCwiZwBOdsWS5GwA2wFcYk5mknsAfMnMDkWPyd5iZpdV2ldXV5d1\nd3dnUOypc/9Od9RdcuSea3Bw0Ft2q/4tLS2ltDu6LXmsSuWotqrvHjtvWVzRPl2u5LuquWmmlrxp\nb6yqtD7NVf6rr74a3d3dk2ZMc8a/GEA/gH8m+RbJf4oel73AzA5FeXpRfKquiJwB0gR+M4AvAPix\nmV0F4DMkqvVRTWDCf2kk15LsJtmd9sKDiOQrTeAfBHDQzLZFy8+i+I/go6iKj+h330QvNrN1ZtZl\nZl2dnZ1ZlFlEajRpd56Z9ZI8QPIyM9sD4CYAu6Kf1QAein5vyLWkU5TsrnGXkxNslJNsS7tdMsl2\nvSvvSSjybjNXM5nH2TTxxtn0t5STth//zwE8SbIFwD4Af4JibeEZkmsAfAjgrnyKKCJZSxX4ZrYd\nQNcEm27KtjgiUg9n7ci9ZHXeXXar7MkRVZWqueWq91O54SMLlZoZZ7J6dvtVe6xqu++qmWQkz/dD\nY/VFAqTAFwmQAl8kQGdtG78St+2ebOOnbYu5w37rOYS2WtUOqU37fmQxZLeek22e6V12tbb/dcYX\nCZACXyRAk96dl+nByH4UB/vMB3C4bgee2OlQBkDlSFI5fFMtx0VmNunY+LoGfumgZLeZTTQgKKgy\nqBwqR6PKoaq+SIAU+CIBalTgr2vQcV2nQxkAlSNJ5fDlUo6GtPFFpLFU1RcJUF0Dn+QqkntI7iVZ\nt1l5ST5Bso/kTmdd3acHJ3kByc0kd5F8h+S9jSgLyTaSr5PcEZXjB9H6i0luiz6fp6P5F3JHshDN\n5/hio8pBcj/Jt0luJ9kdrWvEd6QuU9nXLfBJFgD8A4A/BHAFgLtJXlGnw/8EwKrEukZMDz4C4Dtm\ndgWA5QDuid6DepdlEMBKM7sSwDIAq0guB/AwgEfMbAmAowDW5FyOcfeiOGX7uEaV48tmtszpPmvE\nd6Q+U9mbWV1+AFwH4BVn+X4A99fx+IsB7HSW9wBYGKUXAthTr7I4ZdgA4OZGlgXAdAC/AXAtigNF\nmif6vHI8/qLoy7wSwIsA2KBy7AcwP7Gurp8LgNkA/hvRtbc8y1HPqv75AA44ywejdY3S0OnBSS4G\ncBWAbY0oS1S93o7iJKkbAXwA4JiZjU9OWK/P50cAvgtg/G6pcxpUDgPwK5Jvklwbrav351K3qex1\ncQ+VpwfPA8mZAH4J4NtmdrwRZTGzUTNbhuIZ9xoAl+d9zCSStwHoM7M3633sCdxgZl9AsSl6D8kb\n3Y11+lxqmsp+KuoZ+D0ALnCWF0XrGiXV9OBZIzkNxaB/0syea2RZAMDMjgHYjGKVeg7J8Vu16/H5\nrADwNZL7ATyFYnX/0QaUA2bWE/3uA/A8iv8M6/251DSV/VTUM/DfALA0umLbAuDrAF6o4/GTXkBx\nWnCgTtODs3gT+OMAdpvZDxtVFpKdJOdE6XYUrzPsRvEfwJ31KoeZ3W9mi8xsMYrfh9fM7Bv1LgfJ\nGSRnjacB3AJgJ+r8uZhZL4ADJMcfRTc+lX325cj7okniIsVXAbyHYnvyr+t43J8DOARgGMX/qmtQ\nbEtuAvA+gFcBzKtDOW5AsZr2WxSfR7g9ek/qWhYAvwfgragcOwF8P1p/CYDXAewF8AsArXX8jL4E\n4MVGlCM63o7o553x72aDviPLAHRHn82/ApibRzk0ck8kQLq4JxIgBb5IgBT4IgFS4IsESIEvEiAF\nvkiAFPgiAVLgiwTo/wCbC6Ub6MSxsAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(X_val[2468])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "lR9FUZ2ye9RJ",
    "outputId": "51022cdf-4d4d-4538-d580-0de61b55fbea"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([178])"
      ]
     },
     "execution_count": 52,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.argmax(model.predict(X_val[2468].reshape((1,64,64,3))),axis=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "wbXux_r4gE5O",
    "outputId": "54b3ddba-e571-4345-d2d5-d9a20d96d3d6"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'ulmus_americana'"
      ]
     },
     "execution_count": 53,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_val.columns[178]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 170
    },
    "colab_type": "code",
    "id": "MbJ8jQGefD81",
    "outputId": "33ea8846-2529-4269-ea2c-16b13c8bcc11"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.6/dist-packages/numpy/core/fromnumeric.py:56: FutureWarning: \n",
      "The current behaviour of 'Series.argmax' is deprecated, use 'idxmax'\n",
      "instead.\n",
      "The behavior of 'argmax' will be corrected to return the positional\n",
      "maximum in the future. For now, use 'series.values.argmax' or\n",
      "'np.argmax(np.array(values))' to get the position of the maximum\n",
      "row.\n",
      "  return getattr(obj, method)(*args, **kwds)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'populus_tremuloides'"
      ]
     },
     "execution_count": 54,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.argmax(y_val.iloc[2468])"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "machine_shape": "hm",
   "name": "leaf_classification.ipynb",
   "provenance": []
  },
  "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": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}