ITSP_All_Symbol_Training_v2_0.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "ITSP-All-Symbol-Training-v2.0.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "bRHPp3uXxdvx",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 122
        },
        "outputId": "e16bfab4-27d7-414b-80cc-9452b7af9b01"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3aietf%3awg%3aoauth%3a2.0%3aoob&response_type=code&scope=email%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdocs.test%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive.photos.readonly%20https%3a%2f%2fwww.googleapis.com%2fauth%2fpeopleapi.readonly\n",
            "\n",
            "Enter your authorization code:\n",
            "··········\n",
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4reJFlQJxvF2",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import zipfile\n",
        "\n",
        "with zipfile.ZipFile('/content/drive/My Drive/mega.zip') as z:\n",
        "  z.extractall('/content/mega')"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "U_ju-IjNxxFN",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import os\n",
        "import sys\n",
        "import random\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hMxAsyrXxzW5",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "imlist = os.listdir('/content/mega/Mega Data/')"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HTh7GnCtx1Kg",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 71
        },
        "outputId": "805c13f8-4318-42b4-9aa6-9a18af293673"
      },
      "source": [
        "print(imlist)\n",
        "print(len(imlist))"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "['zero', 'LeftBrace', 'RightPar', 'phi', 'seven', 'z', 'mu', 'Sigma', 'RightBrace', 'infty', 'theta', 'six', 'gamma', 'Delta', 'plus', 'c', 'minus', 'beta', 'Psi', 'zeta', 'pi', 'three', 'n', 's', 'alpha', 'LeftPar', 't', 'nine', 'Gamma_upper', 'y', 'a', 'four', 'l', 'Pi_upper', 'eight', 'two', 'e', 'five', 'partial', 'prime', 'g', 'x', 'nabla', '.DS_Store', 'omega', 'one']\n",
            "46\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "h9kJk_Bf7Vwa",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "fa60f150-6602-4034-8fe7-70d6ec3248d9"
      },
      "source": [
        "TRAIN_PATH = '/content/mega/Mega Data/'\n",
        "train_dirs = next(os.walk(TRAIN_PATH))[1]\n",
        "\n",
        "total_imgs = 0\n",
        "\n",
        "for train_dir in train_dirs:\n",
        "\n",
        "  dir_path = TRAIN_PATH + train_dir\n",
        "  dir_files = next(os.walk(dir_path))[2]\n",
        "\n",
        "  for dir_file in dir_files:\n",
        "    \n",
        "    if dir_file == '.DS_Store':\n",
        "      continue\n",
        "       \n",
        "    total_imgs += 1\n",
        "\n",
        "print(total_imgs)"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "22707\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4O_EC4Kv877j",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "193b372f-f92b-4f68-82f6-01954b031754"
      },
      "source": [
        "import cv2\n",
        "\n",
        "from tqdm import tqdm\n",
        "\n",
        "X_train = np.zeros((total_imgs, 64, 64, 1), dtype = np.uint8)\n",
        "Y_train = np.zeros((total_imgs, 1))\n",
        "\n",
        "label_to_symbol_dict = {}\n",
        "\n",
        "label = 0\n",
        "index = 0\n",
        "for id_, train_dir in tqdm(enumerate(train_dirs), total = len(train_dirs)):\n",
        "\n",
        "  label_to_symbol_dict[label] = train_dir\n",
        "\n",
        "  dir_path = TRAIN_PATH + train_dir + '/'\n",
        "  dir_files = next(os.walk(dir_path))[2]\n",
        "\n",
        "  for dir_file in dir_files:\n",
        "\n",
        "    if dir_file == '.DS_Store':\n",
        "      continue\n",
        "\n",
        "    image_path = dir_path + dir_file\n",
        "    image = cv2.imread(image_path)\n",
        "    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)\n",
        "    gray = cv2.resize(gray, (64, 64))\n",
        "\n",
        "    gray = np.expand_dims(gray, axis = 2)\n",
        "    X_train[index] = gray\n",
        "    Y_train[index] = label\n",
        "\n",
        "    index += 1\n",
        "\n",
        "  label += 1"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "100%|██████████| 45/45 [00:02<00:00, 18.39it/s]\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KxvXbzk9ABpC",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 302
        },
        "outputId": "eac97d4e-bc86-4019-8b0d-0836c6c83e4c"
      },
      "source": [
        "image_test = X_train[0].squeeze()\n",
        "print(image_test.shape)\n",
        "\n",
        "plt.imshow(image_test)"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(64, 64)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7fd7277e37f0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Vfteq-E5Dlx3",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "5c59c72d-79d3-434f-93c9-220374337f0e"
      },
      "source": [
        "print(Y_train[0])"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[0.]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PqFZScbwDE5H",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "858c0da0-91e8-4936-e7d4-f52b971f19f8"
      },
      "source": [
        "from tensorflow.keras.utils import to_categorical\n",
        "\n",
        "Y_train = to_categorical(Y_train, num_classes = 45)\n",
        "\n",
        "print(Y_train.shape)"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(22707, 45)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "usR-r6HWCbMz",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "X_train, X_val, Y_train, Y_val = train_test_split(X_train, Y_train, test_size = 0.1, random_state = 42)"
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "50dmK1qnyEc5",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import numpy as np # linear algebra\n",
        "import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n",
        "\n",
        "from tensorflow.keras.layers import Conv2D, BatchNormalization, Dropout, MaxPool2D\n",
        "from tensorflow.keras.layers import Flatten, Dense\n",
        "from tensorflow.keras import Model\n",
        "from tensorflow.keras.optimizers import Adam\n",
        "from tensorflow.keras.models import Sequential"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ihk5Jgh9yHGu",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model = Sequential()\n",
        "\n",
        "model.add(Conv2D(filters = 32, kernel_size = (3,3), padding = 'Same', \n",
        "                 activation ='relu', input_shape = (64,64,1)))\n",
        "model.add(BatchNormalization(axis = 3))\n",
        "model.add(Conv2D(filters = 32, kernel_size = (3,3), padding = 'Same', \n",
        "                 activation ='relu'))\n",
        "model.add(BatchNormalization(axis = 3))\n",
        "model.add(Dropout(0.25))\n",
        "\n",
        "model.add(Conv2D(filters = 32, kernel_size = (3,3), padding = 'Same', \n",
        "                 activation ='relu', input_shape = (64,64,1)))\n",
        "model.add(BatchNormalization(axis = 3))\n",
        "model.add(Conv2D(filters = 32, kernel_size = (3,3), padding = 'Same', \n",
        "                 activation ='relu'))\n",
        "model.add(BatchNormalization(axis = 3))\n",
        "model.add(Dropout(0.25))\n",
        "\n",
        "model.add(MaxPool2D(pool_size=(2,2)))\n",
        "\n",
        "\n",
        "model.add(Conv2D(filters = 64, kernel_size = (5,5), padding = 'Same', \n",
        "                 activation ='relu'))\n",
        "model.add(BatchNormalization(axis = 3))\n",
        "model.add(Conv2D(filters = 64, kernel_size = (5,5), padding = 'Same', \n",
        "                 activation ='relu'))\n",
        "model.add(BatchNormalization(axis = 3))\n",
        "model.add(Dropout(0.25))\n",
        "\n",
        "model.add(Conv2D(filters = 64, kernel_size = (5,5), padding = 'Same', \n",
        "                 activation ='relu'))\n",
        "model.add(BatchNormalization(axis = 3))\n",
        "model.add(Conv2D(filters = 64, kernel_size = (5,5), padding = 'Same', \n",
        "                 activation ='relu'))\n",
        "model.add(BatchNormalization(axis = 3))\n",
        "model.add(Dropout(0.25))\n",
        "\n",
        "model.add(MaxPool2D(pool_size=(2,2)))\n",
        "\n",
        "\n",
        "model.add(Conv2D(filters = 32, kernel_size = (3,3), padding = 'Same', \n",
        "                 activation ='relu'))\n",
        "model.add(BatchNormalization(axis = 3))\n",
        "model.add(Conv2D(filters = 32, kernel_size = (3,3), padding = 'Same', \n",
        "                 activation ='relu'))\n",
        "model.add(BatchNormalization(axis = 3))\n",
        "model.add(Dropout(0.25))\n",
        "\n",
        "model.add(Conv2D(filters = 32, kernel_size = (3,3), padding = 'Same', \n",
        "                 activation ='relu'))\n",
        "model.add(BatchNormalization(axis = 3))\n",
        "model.add(Conv2D(filters = 32, kernel_size = (3,3), padding = 'Same', \n",
        "                 activation ='relu'))\n",
        "model.add(BatchNormalization(axis = 3))\n",
        "model.add(Dropout(0.25))\n",
        "\n",
        "model.add(MaxPool2D(pool_size=(2,2)))\n",
        "\n",
        "\n",
        "model.add(Flatten())\n",
        "model.add(Dense(1024, activation = \"relu\"))\n",
        "model.add(BatchNormalization(axis = 1))\n",
        "model.add(Dropout(0.5))\n",
        "model.add(Dense(45, activation = \"softmax\"))"
      ],
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DsE49Tishi-m",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "0bd327a2-7bb4-4f79-d657-0a862fd913eb"
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_2\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d_20 (Conv2D)           (None, 64, 64, 32)        320       \n",
            "_________________________________________________________________\n",
            "batch_normalization_22 (Batc (None, 64, 64, 32)        128       \n",
            "_________________________________________________________________\n",
            "conv2d_21 (Conv2D)           (None, 64, 64, 32)        9248      \n",
            "_________________________________________________________________\n",
            "batch_normalization_23 (Batc (None, 64, 64, 32)        128       \n",
            "_________________________________________________________________\n",
            "dropout_12 (Dropout)         (None, 64, 64, 32)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_22 (Conv2D)           (None, 64, 64, 32)        9248      \n",
            "_________________________________________________________________\n",
            "batch_normalization_24 (Batc (None, 64, 64, 32)        128       \n",
            "_________________________________________________________________\n",
            "conv2d_23 (Conv2D)           (None, 64, 64, 32)        9248      \n",
            "_________________________________________________________________\n",
            "batch_normalization_25 (Batc (None, 64, 64, 32)        128       \n",
            "_________________________________________________________________\n",
            "dropout_13 (Dropout)         (None, 64, 64, 32)        0         \n",
            "_________________________________________________________________\n",
            "max_pooling2d_5 (MaxPooling2 (None, 32, 32, 32)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_24 (Conv2D)           (None, 32, 32, 64)        51264     \n",
            "_________________________________________________________________\n",
            "batch_normalization_26 (Batc (None, 32, 32, 64)        256       \n",
            "_________________________________________________________________\n",
            "conv2d_25 (Conv2D)           (None, 32, 32, 64)        102464    \n",
            "_________________________________________________________________\n",
            "batch_normalization_27 (Batc (None, 32, 32, 64)        256       \n",
            "_________________________________________________________________\n",
            "dropout_14 (Dropout)         (None, 32, 32, 64)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_26 (Conv2D)           (None, 32, 32, 64)        102464    \n",
            "_________________________________________________________________\n",
            "batch_normalization_28 (Batc (None, 32, 32, 64)        256       \n",
            "_________________________________________________________________\n",
            "conv2d_27 (Conv2D)           (None, 32, 32, 64)        102464    \n",
            "_________________________________________________________________\n",
            "batch_normalization_29 (Batc (None, 32, 32, 64)        256       \n",
            "_________________________________________________________________\n",
            "dropout_15 (Dropout)         (None, 32, 32, 64)        0         \n",
            "_________________________________________________________________\n",
            "max_pooling2d_6 (MaxPooling2 (None, 16, 16, 64)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_28 (Conv2D)           (None, 16, 16, 32)        18464     \n",
            "_________________________________________________________________\n",
            "batch_normalization_30 (Batc (None, 16, 16, 32)        128       \n",
            "_________________________________________________________________\n",
            "conv2d_29 (Conv2D)           (None, 16, 16, 32)        9248      \n",
            "_________________________________________________________________\n",
            "batch_normalization_31 (Batc (None, 16, 16, 32)        128       \n",
            "_________________________________________________________________\n",
            "dropout_16 (Dropout)         (None, 16, 16, 32)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_30 (Conv2D)           (None, 16, 16, 32)        9248      \n",
            "_________________________________________________________________\n",
            "batch_normalization_32 (Batc (None, 16, 16, 32)        128       \n",
            "_________________________________________________________________\n",
            "conv2d_31 (Conv2D)           (None, 16, 16, 32)        9248      \n",
            "_________________________________________________________________\n",
            "batch_normalization_33 (Batc (None, 16, 16, 32)        128       \n",
            "_________________________________________________________________\n",
            "dropout_17 (Dropout)         (None, 16, 16, 32)        0         \n",
            "_________________________________________________________________\n",
            "max_pooling2d_7 (MaxPooling2 (None, 8, 8, 32)          0         \n",
            "_________________________________________________________________\n",
            "flatten_2 (Flatten)          (None, 2048)              0         \n",
            "_________________________________________________________________\n",
            "dense_4 (Dense)              (None, 1024)              2098176   \n",
            "_________________________________________________________________\n",
            "batch_normalization_34 (Batc (None, 1024)              4096      \n",
            "_________________________________________________________________\n",
            "dropout_18 (Dropout)         (None, 1024)              0         \n",
            "_________________________________________________________________\n",
            "dense_5 (Dense)              (None, 45)                46125     \n",
            "=================================================================\n",
            "Total params: 2,583,373\n",
            "Trainable params: 2,580,301\n",
            "Non-trainable params: 3,072\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NPy0fNKh2V11",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from tensorflow.keras.callbacks import ReduceLROnPlateau, EarlyStopping"
      ],
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "d6ftouA81KGe",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 935
        },
        "outputId": "f037acec-50f5-49b1-9385-2a2bef3169ea"
      },
      "source": [
        "optimizer = Adam(learning_rate = 0.001)\n",
        "\n",
        "model.compile(optimizer = optimizer, loss = 'categorical_crossentropy', metrics = ['accuracy'])\n",
        "\n",
        "lr_scheduler = ReduceLROnPlateau(monitor = 'loss', patience = 1, verbose = 1, factor = 0.1)\n",
        "\n",
        "early_stopper = EarlyStopping(monitor = 'loss', patience = 12, verbose = 1)\n",
        "\n",
        "history = model.fit(x = X_train, y = Y_train, batch_size = 32, epochs = 20, validation_data = (X_val, Y_val), callbacks = [lr_scheduler, early_stopper], verbose = 2, validation_batch_size = 64)"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/20\n",
            "639/639 - 54s - loss: 0.2602 - accuracy: 0.9336 - val_loss: 0.2534 - val_accuracy: 0.9212 - lr: 0.0010\n",
            "Epoch 2/20\n",
            "639/639 - 53s - loss: 0.0310 - accuracy: 0.9927 - val_loss: 0.1117 - val_accuracy: 0.9679 - lr: 0.0010\n",
            "Epoch 3/20\n",
            "639/639 - 53s - loss: 0.0243 - accuracy: 0.9943 - val_loss: 0.0683 - val_accuracy: 0.9819 - lr: 0.0010\n",
            "Epoch 4/20\n",
            "639/639 - 53s - loss: 0.0225 - accuracy: 0.9939 - val_loss: 0.0488 - val_accuracy: 0.9908 - lr: 0.0010\n",
            "Epoch 5/20\n",
            "639/639 - 53s - loss: 0.0221 - accuracy: 0.9941 - val_loss: 0.0287 - val_accuracy: 0.9916 - lr: 0.0010\n",
            "Epoch 6/20\n",
            "639/639 - 53s - loss: 0.0186 - accuracy: 0.9954 - val_loss: 0.0276 - val_accuracy: 0.9947 - lr: 0.0010\n",
            "Epoch 7/20\n",
            "\n",
            "Epoch 00007: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
            "639/639 - 53s - loss: 0.0207 - accuracy: 0.9946 - val_loss: 0.0214 - val_accuracy: 0.9938 - lr: 0.0010\n",
            "Epoch 8/20\n",
            "639/639 - 53s - loss: 0.0097 - accuracy: 0.9977 - val_loss: 0.0126 - val_accuracy: 0.9978 - lr: 1.0000e-04\n",
            "Epoch 9/20\n",
            "639/639 - 53s - loss: 0.0058 - accuracy: 0.9985 - val_loss: 0.0131 - val_accuracy: 0.9978 - lr: 1.0000e-04\n",
            "Epoch 10/20\n",
            "639/639 - 53s - loss: 0.0037 - accuracy: 0.9989 - val_loss: 0.0103 - val_accuracy: 0.9974 - lr: 1.0000e-04\n",
            "Epoch 11/20\n",
            "\n",
            "Epoch 00011: ReduceLROnPlateau reducing learning rate to 1.0000000474974514e-05.\n",
            "639/639 - 53s - loss: 0.0051 - accuracy: 0.9988 - val_loss: 0.0113 - val_accuracy: 0.9974 - lr: 1.0000e-04\n",
            "Epoch 12/20\n",
            "\n",
            "Epoch 00012: ReduceLROnPlateau reducing learning rate to 1.0000000656873453e-06.\n",
            "639/639 - 53s - loss: 0.0042 - accuracy: 0.9988 - val_loss: 0.0101 - val_accuracy: 0.9974 - lr: 1.0000e-05\n",
            "Epoch 13/20\n",
            "639/639 - 53s - loss: 0.0028 - accuracy: 0.9992 - val_loss: 0.0100 - val_accuracy: 0.9974 - lr: 1.0000e-06\n",
            "Epoch 14/20\n",
            "639/639 - 53s - loss: 0.0022 - accuracy: 0.9992 - val_loss: 0.0108 - val_accuracy: 0.9974 - lr: 1.0000e-06\n",
            "Epoch 15/20\n",
            "639/639 - 53s - loss: 0.0020 - accuracy: 0.9995 - val_loss: 0.0099 - val_accuracy: 0.9974 - lr: 1.0000e-06\n",
            "Epoch 16/20\n",
            "\n",
            "Epoch 00016: ReduceLROnPlateau reducing learning rate to 1.0000001111620805e-07.\n",
            "639/639 - 53s - loss: 0.0027 - accuracy: 0.9993 - val_loss: 0.0107 - val_accuracy: 0.9974 - lr: 1.0000e-06\n",
            "Epoch 17/20\n",
            "\n",
            "Epoch 00017: ReduceLROnPlateau reducing learning rate to 1.000000082740371e-08.\n",
            "639/639 - 53s - loss: 0.0028 - accuracy: 0.9992 - val_loss: 0.0114 - val_accuracy: 0.9974 - lr: 1.0000e-07\n",
            "Epoch 18/20\n",
            "\n",
            "Epoch 00018: ReduceLROnPlateau reducing learning rate to 1.000000082740371e-09.\n",
            "639/639 - 53s - loss: 0.0024 - accuracy: 0.9995 - val_loss: 0.0098 - val_accuracy: 0.9974 - lr: 1.0000e-08\n",
            "Epoch 19/20\n",
            "\n",
            "Epoch 00019: ReduceLROnPlateau reducing learning rate to 1.000000082740371e-10.\n",
            "639/639 - 53s - loss: 0.0024 - accuracy: 0.9996 - val_loss: 0.0097 - val_accuracy: 0.9974 - lr: 1.0000e-09\n",
            "Epoch 20/20\n",
            "639/639 - 53s - loss: 0.0018 - accuracy: 0.9994 - val_loss: 0.0109 - val_accuracy: 0.9974 - lr: 1.0000e-10\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QxIpZMKV21lE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model.save(\"monster_model.h5\")"
      ],
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "q5fvuuYv2-i0",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import json\n",
        "\n",
        "with open('mappings_mega.json', 'w') as fp:\n",
        "    json.dump(label_to_symbol_dict, fp)"
      ],
      "execution_count": 25,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zcsrDGvM3CJP",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "with open('mappings_mega.json', 'r') as fp:\n",
        "    mappings = json.load(fp)"
      ],
      "execution_count": 26,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "B8EHxWR6FHNp",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 54
        },
        "outputId": "c769e32d-0e91-4242-ea85-ca5bf732c0a8"
      },
      "source": [
        "print(mappings)"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "{'0': 'zero', '1': 'LeftBrace', '2': 'RightPar', '3': 'phi', '4': 'seven', '5': 'z', '6': 'mu', '7': 'Sigma', '8': 'RightBrace', '9': 'infty', '10': 'theta', '11': 'six', '12': 'gamma', '13': 'Delta', '14': 'plus', '15': 'c', '16': 'minus', '17': 'beta', '18': 'Psi', '19': 'zeta', '20': 'pi', '21': 'three', '22': 'n', '23': 's', '24': 'alpha', '25': 'LeftPar', '26': 't', '27': 'nine', '28': 'Gamma_upper', '29': 'y', '30': 'a', '31': 'four', '32': 'l', '33': 'Pi_upper', '34': 'eight', '35': 'two', '36': 'e', '37': 'five', '38': 'partial', '39': 'prime', '40': 'g', '41': 'x', '42': 'nabla', '43': 'omega', '44': 'one'}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1QozThyB3Eaj",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 268
        },
        "outputId": "d3dd74f5-2919-478f-ee4a-e29b15f82b3e"
      },
      "source": [
        "#from PIL import Image\n",
        "\n",
        "test_img_path = '/content/mega/Mega Data/one/0.png'\n",
        "\n",
        "test_img = cv2.imread(test_img_path)\n",
        "test_img = cv2.cvtColor(test_img, cv2.COLOR_BGR2GRAY)\n",
        "test_img = np.array(test_img)\n",
        "test_img = cv2.resize(test_img,(64,64))\n",
        "\n",
        "img = np.array(test_img)\n",
        "plt.imshow(img, cmap = 'gray')\n",
        "plt.show()"
      ],
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VaUdJJSq3K17",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "8a992cda-0446-4636-fa7a-f2340ac1df98"
      },
      "source": [
        "print(test_img.shape)"
      ],
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(64, 64)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VUnw5MTw3XF0",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "test_img = np.expand_dims(test_img, axis = 2)"
      ],
      "execution_count": 36,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZjL-Qq4O3ffo",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "41d54444-e1e7-46d2-ab7b-4f74df257342"
      },
      "source": [
        "print(test_img.shape)"
      ],
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(64, 64, 1)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ymmL-o383g5n",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "test = np.zeros((1, 64, 64, 1), dtype = np.uint8)\n",
        "\n",
        "test[0] = test_img"
      ],
      "execution_count": 38,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IoqJaoez3oXd",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "8002d6ec-bac3-40ea-fbdf-445d281f6d08"
      },
      "source": [
        "mappings[str(np.argmax(model.predict(test)))]"
      ],
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'one'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 39
        }
      ]
    }
  ]
}