26_class_2layers.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fPCIU1KpzSJP",
        "outputId": "a8a064f1-5704-4790-e647-ba3ce9318500"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "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')\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "oK_6GUkt2Do-",
        "outputId": "4085a6ce-ece9-44c6-d8a0-99b5446eaf29"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "/content/drive/MyDrive/Colab Notebooks\n"
          ]
        }
      ],
      "source": [
        "%cd drive/MyDrive/Colab Notebooks"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "id": "XJavin0WykpJ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d8fca09f-efa9-48f3-fe29-783afed48bf9"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/tensorflow/python/compat/v2_compat.py:107: disable_resource_variables (from tensorflow.python.ops.variable_scope) is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "non-resource variables are not supported in the long term\n"
          ]
        }
      ],
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt \n",
        "import seaborn as sns\n",
        "import os\n",
        "import random\n",
        "import cv2\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "np.random.seed(0)\n",
        "\n",
        "from sklearn.metrics import confusion_matrix\n",
        "from sklearn.preprocessing import LabelBinarizer\n",
        "\n",
        "import tensorflow.compat.v1 as tf\n",
        "tf.disable_v2_behavior()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "id": "vwt_wr5hykpN"
      },
      "outputs": [],
      "source": [
        "# with tf.device('/CPU:0'):\n",
        "\n",
        "map_characters = {0: 'A', 1: 'B', 2: 'C', 3: 'D', 4: 'E', 5: 'F', 6: 'G', 7: 'H', 8: 'I', 9: 'J', \n",
        "                10: 'K', 11: 'L', 12: 'M', 13: 'N', 14: 'O', 15: 'P', 16: 'Q', 17: 'R', 18: 'S', \n",
        "                19: 'T', 20: 'U', 21: 'V', 22: 'W', 23: 'X', 24: 'Y', 25: 'Z'}\n",
        "\n",
        "\n",
        "labels_dict = {'A':0,'B':1,'C':2,'D':3,'E':4,'F':5,'G':6,'H':7,'I':8,'J':9,'K':10,'L':11,'M':12,\n",
        "                'N':13,'O':14,'P':15,'Q':16,'R':17,'S':18,'T':19,'U':20,'V':21,'W':22,'X':23,'Y':24,\n",
        "                'Z':25}\n",
        "\n",
        "\n",
        "order_list=('A',  'B',  'C', 'D', 'E',  'F', 'G', 'H',  'I',  'J', \n",
        "                'K',  'L',  'M',  'N',  'O',  'P',  'Q',  'R',  'S', \n",
        "                'T',  'U',  'V',  'W',  'X',  'Y',  'Z')\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "lb=LabelBinarizer()\n",
        "hotEncodedLabels=lb.fit_transform(order_list)\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "train_dir=\"asl_alphabet_train/asl_alphabet_train\"\n",
        "\n",
        "def load_data():\n",
        "    \"\"\"\n",
        "    Loads data and preprocess. Returns train and test data along with labels.\n",
        "    \"\"\"\n",
        "    images = []\n",
        "    labels = []\n",
        "    size = 64,64\n",
        "    print(\"LOADING DATA FROM : \",end = \"\")\n",
        "    for folder in os.listdir(train_dir):\n",
        "        print(folder, end = ' | ')\n",
        "        for image in os.listdir(train_dir + \"/\" + folder):\n",
        "                temp_img = cv2.imread(train_dir + '/' + folder + '/' + image, 0) # greyscale\n",
        "                temp_img = cv2.resize(temp_img, size)\n",
        "                temp_img = temp_img.flatten()\n",
        "                images.append(temp_img)\n",
        "                labels.append(hotEncodedLabels[labels_dict[folder]])\n",
        "            \n",
        "    \n",
        "    images = np.array(images)\n",
        "    images = images.astype('float32')/255.0\n",
        "    \n",
        "    \n",
        "    X_train, X_test, Y_train, Y_test = train_test_split(images, labels, test_size = 0.3, random_state=42)\n",
        "    X_test, X_validation, Y_test, Y_validation = train_test_split(X_test, Y_test, test_size = 0.8,random_state=42)\n",
        "    \n",
        "    \n",
        "    print()\n",
        "    print('Loaded', len(X_train),'images for training,','Train data shape =',X_train.shape)\n",
        "    print('Loaded', len(X_validation),'images for validation','validation data shape =',X_validation.shape)\n",
        "    print('Loaded', len(X_test),'images for testing','Test data shape =',X_test.shape)\n",
        "\n",
        "    \n",
        "    return X_train, X_test, Y_train, Y_test, X_validation, Y_validation\n",
        "\n",
        "def confussion_matrix(cf_matrix):\n",
        "    plt.figure(figsize=(10,10))\n",
        "    sns.heatmap(cf_matrix, annot=True, fmt='', cmap='Greens')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vD36xmfKykpP",
        "outputId": "1376921b-a36c-4bd1-8e9e-72f1893c2a36"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "LOADING DATA FROM : U | R | N | I | Z | G | T | S | F | H | M | A | O | C | J | V | D | P | Q | Y | B | X | K | L | E | W | \n",
            "Loaded 54614 images for training, Train data shape = (54614, 4096)\n",
            "Loaded 18726 images for validation validation data shape = (18726, 4096)\n",
            "Loaded 4681 images for testing Test data shape = (4681, 4096)\n",
            "54614\n",
            "18726\n",
            "4681\n"
          ]
        }
      ],
      "source": [
        "# with tf.device('/CPU:0'):\n",
        "X_train, X_test, Y_train, Y_test, X_validation, Y_validation= load_data()\n",
        "\n",
        "\n",
        "print(len(X_train))\n",
        "print(len(X_validation))\n",
        "print(len(X_test))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xZPciJJEykpR",
        "outputId": "f7afa097-93dd-4910-b86e-05758bcd5b11"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "26_class_5layers.ipynb  \u001b[0m\u001b[01;34masl_alphabet_train\u001b[0m/  Untitled0.ipynb\n",
            "\u001b[01;34masl_alphabet_test\u001b[0m/      Untitled\n"
          ]
        }
      ],
      "source": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "4IBk_kevykpR"
      },
      "outputs": [],
      "source": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "TTqkUHZoykpS"
      },
      "outputs": [],
      "source": [
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "oI8mXQ68ykpT"
      },
      "outputs": [],
      "source": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "R-g4lT2vykpT"
      },
      "outputs": [],
      "source": [
        "# with tf.device('/CPU:0'):\n",
        "#     y_tr=[]\n",
        "#     for i in Y_train:\n",
        "#         y_tr.append(i.tolist())\n",
        "#     y_train_t=tf.convert_to_tensor(\n",
        "#         y_tr, dtype=None, dtype_hint=None, name=None\n",
        "#     )\n",
        "\n",
        "#     X_tr=[]\n",
        "#     for i in X_train:\n",
        "#         X_tr.append(i.tolist())\n",
        "#     X_train_t=tf.convert_to_tensor(\n",
        "#         X_tr, dtype=None, dtype_hint=None, name=None\n",
        "#     )\n",
        "\n",
        "#     y_vald=[]\n",
        "#     for i in Y_validation:\n",
        "#         y_vald.append(i.tolist())\n",
        "#     y_vald_t=tf.convert_to_tensor(\n",
        "#         y_vald, dtype=None, dtype_hint=None, name=None\n",
        "#     )"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dnLL1a2cykpU",
        "outputId": "f31bf851-5b56-4fc7-92ac-f64903d6e662"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "start\n",
            "0  this epoch accuracy:  0.042828154  best accuracy seen when checked  0.042828154\n",
            "1000  this epoch accuracy:  0.23032148  best accuracy seen when checked  0.23032148\n",
            "2000  this epoch accuracy:  0.3168322  best accuracy seen when checked  0.3168322\n",
            "3000  this epoch accuracy:  0.37119514  best accuracy seen when checked  0.37119514\n",
            "4000  this epoch accuracy:  0.4201111  best accuracy seen when checked  0.4201111\n",
            "5000  this epoch accuracy:  0.45994875  best accuracy seen when checked  0.45994875\n",
            "6000  this epoch accuracy:  0.49012068  best accuracy seen when checked  0.49012068\n",
            "7000  this epoch accuracy:  0.51778275  best accuracy seen when checked  0.51778275\n",
            "8000  this epoch accuracy:  0.543202  best accuracy seen when checked  0.543202\n",
            "9000  this epoch accuracy:  0.5634412  best accuracy seen when checked  0.5634412\n",
            "10000  this epoch accuracy:  0.582292  best accuracy seen when checked  0.582292\n",
            "11000  this epoch accuracy:  0.5996475  best accuracy seen when checked  0.5996475\n",
            "12000  this epoch accuracy:  0.61374557  best accuracy seen when checked  0.61374557\n",
            "13000  this epoch accuracy:  0.62875146  best accuracy seen when checked  0.62875146\n",
            "14000  this epoch accuracy:  0.6429563  best accuracy seen when checked  0.6429563\n",
            "15000  this epoch accuracy:  0.6545445  best accuracy seen when checked  0.6545445\n",
            "16000  this epoch accuracy:  0.66634625  best accuracy seen when checked  0.66634625\n",
            "17000  this epoch accuracy:  0.6768664  best accuracy seen when checked  0.6768664\n",
            "18000  this epoch accuracy:  0.6865321  best accuracy seen when checked  0.6865321\n",
            "19000  this epoch accuracy:  0.69571716  best accuracy seen when checked  0.69571716\n",
            "20000  this epoch accuracy:  0.7047955  best accuracy seen when checked  0.7047955\n",
            "21000  this epoch accuracy:  0.71349996  best accuracy seen when checked  0.71349996\n",
            "22000  this epoch accuracy:  0.71974796  best accuracy seen when checked  0.71974796\n",
            "23000  this epoch accuracy:  0.7277048  best accuracy seen when checked  0.7277048\n",
            "24000  this epoch accuracy:  0.7341664  best accuracy seen when checked  0.7341664\n",
            "25000  this epoch accuracy:  0.73950654  best accuracy seen when checked  0.73950654\n",
            "26000  this epoch accuracy:  0.74650216  best accuracy seen when checked  0.74650216\n",
            "27000  this epoch accuracy:  0.7525366  best accuracy seen when checked  0.7525366\n",
            "28000  this epoch accuracy:  0.75787675  best accuracy seen when checked  0.75787675\n",
            "29000  this epoch accuracy:  0.7638043  best accuracy seen when checked  0.7638043\n"
          ]
        }
      ],
      "source": [
        "# with tf.device('/CPU:0'):\n",
        "acc=0\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "(hidden1_size, hidden2_size) = (1040, 520)\n",
        "x = tf.placeholder(tf.float32, [None, 4096])\n",
        "y_ = tf.placeholder(tf.float32, [None, 26])\n",
        "\n",
        "W1 = tf.Variable(tf.truncated_normal([4096, hidden1_size], stddev=0.1))\n",
        "b1 = tf.Variable(tf.constant(0.1, shape=[hidden1_size]))\n",
        "z1 = tf.nn.relu(tf.matmul(x,W1)+b1)\n",
        "\n",
        "W2 = tf.Variable(tf.truncated_normal([hidden1_size, hidden2_size], stddev=0.1))\n",
        "b2 = tf.Variable(tf.constant(0.1, shape=[hidden2_size]))\n",
        "z2 = tf.nn.relu(tf.matmul(z1,W2)+b2)\n",
        "\n",
        "W3 = tf.Variable(tf.truncated_normal([hidden2_size, 26], stddev=0.1))\n",
        "b3 = tf.Variable(tf.constant(0.1, shape=[26]))\n",
        "\n",
        "\n",
        "\n",
        "pred = tf.nn.softmax(tf.matmul(z2, W3)+b3)\n",
        "cross_entropy = tf.reduce_mean(-tf.reduce_sum(y_ * tf.log(pred), reduction_indices=[1]))\n",
        "train_step = tf.train.GradientDescentOptimizer(0.001).minimize(cross_entropy)\n",
        "\n",
        "\n",
        "data_x=np.array(X_train)\n",
        "data_y=np.array(Y_train)  \n",
        "data_arr_x=[data_x[0::5],data_x[1::5],data_x[2::5],data_x[3::5],data_x[4::5]]\n",
        "data_arr_y=[data_y[0::5],data_y[1::5],data_y[2::5],data_y[3::5],data_y[4::5]]\n",
        "\n",
        "\n",
        "sess1 = tf.Session()\n",
        "sess1.run(tf.global_variables_initializer())\n",
        "\n",
        "correct_prediction = tf.equal(tf.argmax(pred,1), tf.argmax(y_,1))\n",
        "accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))\n",
        "\n",
        "print(\"start\")\n",
        "for i in range (30000):\n",
        "    sess1.run(train_step, feed_dict={x: X_train, y_: Y_train})\n",
        "    if i % 1000 == 0 :\n",
        "    # print(i, sess1.run(accuracy, feed_dict={x: X_validation, y_: Y_validation}))\n",
        "      acc1=sess1.run(accuracy, feed_dict={x: X_validation, y_: Y_validation})\n",
        "      acc=max(acc,acc1)\n",
        "      print(i,\" this epoch accuracy: \",acc1,\" best accuracy seen when checked \", acc)\n",
        "\n",
        "\n",
        "        \n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "H4GGVWO8ykpV"
      },
      "outputs": [],
      "source": [
        "# # with tf.device('/CPU:0'):\n",
        "# y_tr=[]\n",
        "# for i in Y_train:\n",
        "#     y_tr.append(i.tolist())\n",
        "# y_train_t=tf.convert_to_tensor(\n",
        "#     y_tr, dtype=None, dtype_hint=None, name=None\n",
        "# )\n",
        "\n",
        "# X_tr=[]\n",
        "# for i in X_train:\n",
        "#     X_tr.append(i.tolist())\n",
        "# X_train_t=tf.convert_to_tensor(\n",
        "#     X_tr, dtype=None, dtype_hint=None, name=None\n",
        "# )\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 609
        },
        "id": "BAQTyg4DykpV",
        "outputId": "9e5206c2-4a40-4c62-cd68-72098cfb5250"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0.7754753\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x720 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "from sklearn.metrics import confusion_matrix\n",
        "import seaborn as sns\n",
        "import matplotlib.pyplot as plt\n",
        "# function to make pretty confusion matrix\n",
        "def confussion_matrix(cf_matrix):\n",
        "    plt.figure(figsize=(10,10))\n",
        "    sns.heatmap(cf_matrix, annot=True, fmt='', cmap='Greens')\n",
        "\n",
        "# make a tensor with the y validation \n",
        "y_vald=[]\n",
        "for i in Y_test:\n",
        "    y_vald.append(i.tolist())\n",
        "y_vald_t=tf.convert_to_tensor(\n",
        "    y_vald, dtype=None, dtype_hint=None, name=None\n",
        ")\n",
        "\n",
        "\n",
        "# get the prediction foe x validation\n",
        "\n",
        "mypred=sess1.run(pred,feed_dict={x: X_test, y_: Y_test})\n",
        "\n",
        "\n",
        "# check accuracy\n",
        "correct_prediction1 = tf.equal(tf.argmax(mypred,1), tf.argmax(y_vald_t,1))\n",
        "accuracy1 = tf.reduce_mean(tf.cast(correct_prediction1, tf.float32))\n",
        "print(sess1.run(accuracy1))\n",
        "\n",
        "\n",
        "# take mypred change to a list then to a tensor\n",
        "mypred_l=[]\n",
        "for i in mypred:\n",
        "    mypred_l.append(i.tolist())\n",
        "\n",
        "mypred_t=tf.convert_to_tensor(\n",
        "    mypred_l, dtype=None, dtype_hint=None, name=None\n",
        ")\n",
        "\n",
        "\n",
        "# create a list with the pred class\n",
        "pred_confussion_matrix=[]\n",
        "for i in sess1.run(mypred_t):\n",
        "    pred_confussion_matrix.append( np.argmax(i))\n",
        "\n",
        "# create a list with the actual class\n",
        "Y_validation_confussion_matrix=[]\n",
        "for i in sess1.run(y_vald_t):\n",
        "    Y_validation_confussion_matrix.append( np.argmax(i))\n",
        "\n",
        "# create the cunfusion matrix\n",
        "confussion_matrix(confusion_matrix(Y_validation_confussion_matrix,pred_confussion_matrix))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "id": "XDqIz4seykpV",
        "outputId": "29e2ea43-bead-4ff7-bfac-15db754797b6"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'26class_2layers_30000iter'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 11
        }
      ],
      "source": [
        "saver = tf.train.Saver()\n",
        "saver.save(sess1, \"26class_2layers_30000iter\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Mb_LKBpOykpb"
      },
      "outputs": [],
      "source": [
        "# 26 classes , 2 layers , 1040 , 520 , 1000 iterations , accuracy 0.228\n",
        "# 26 classes , 2 layers , 1040 , 520 , 5000 iterations , accuracy 0.463"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "provenance": []
    },
    "gpuClass": "premium",
    "kernelspec": {
      "display_name": "Python 3.10.8 64-bit",
      "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": "3.10.8"
    },
    "vscode": {
      "interpreter": {
        "hash": "bcc40836f47ef7b4493f57e3a45135d915df5deffbc353e6c0f28b20bc9e4f44"
      }
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}