Nepali_Digits_Prediction.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Nepali Digits Prediction.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/bhuwanaryal19/nepali_digit_recognizer/blob/master/Nepali_Digits_Prediction.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JcQLIcFPEmDE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Change runtime to GPU for faster compuatation"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "S3sQCofLEyyT",
        "colab_type": "text"
      },
      "source": [
        "# Prepare the dataset\n",
        "Download dataset from [this url](https://archive.ics.uci.edu/ml/machine-learning-databases/00389/). "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xwAARA2yE1Ot",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# get dataset from url\n",
        "! wget -qq https://archive.ics.uci.edu/ml/machine-learning-databases/00389/DevanagariHandwrittenCharacterDataset.zip\n",
        "\n",
        "# unzip zip files\n",
        "!unzip -q DevanagariHandwrittenCharacterDataset.zip\n",
        "\n",
        "# rename directory\n",
        "!mv DevanagariHandwrittenCharacterDataset/ data/ \n",
        "# remove .zip file\n",
        "!rm *.zip"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gnoTXwAZDrB8",
        "colab_type": "text"
      },
      "source": [
        "#Load libraries "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "64bnuj27mS1n",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# import libraries\n",
        "import matplotlib.pyplot as plt\n",
        "from PIL import Image\n",
        "import numpy as np\n",
        "import random\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras import layers, callbacks, optimizers, models"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AWFRyjYXFZy3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# for this problem i'll consider only digits\n",
        "class_names = [f'digit_{i}' for i in range(10)]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gjQr7MMHF8V7",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "train_datagen = ImageDataGenerator(rescale=1./255,)\n",
        "test_datagen = ImageDataGenerator(rescale=1./255)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HlfUmaOBCYlB",
        "colab_type": "code",
        "outputId": "64794ca6-74b1-4acb-846e-838d11fed9f1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "train_digits_loader = train_datagen.flow_from_directory(\n",
        "        'data/Train',\n",
        "        target_size=(32, 32),\n",
        "        classes = class_names,\n",
        "        batch_size=32,\n",
        "        color_mode = 'grayscale',\n",
        "        class_mode='categorical')\n",
        "\n",
        "test_digits_loader = test_datagen.flow_from_directory(\n",
        "        'data/Test',\n",
        "        target_size=(32, 32),\n",
        "        classes = class_names,\n",
        "        color_mode = 'grayscale',\n",
        "        batch_size=1,\n",
        "        shuffle = False,\n",
        "        seed = 2076,\n",
        "        class_mode='categorical')"
      ],
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 17000 images belonging to 10 classes.\n",
            "Found 3000 images belonging to 10 classes.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9GU2g1VfEIl6",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OnFMP_UuEz89",
        "colab_type": "text"
      },
      "source": [
        "# Visualize Data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "L79AV3hxnU0Y",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def plot_image(img_path_index):\n",
        "  img_path = train_digits_loader.filepaths[img_path_index]\n",
        "  a = Image.open(img_path)\n",
        "  arr = np.array(a)\n",
        "  plt.imshow(arr,cmap= 'gray')\n",
        "  plt.grid(False)\n",
        "  plt.xticks([])\n",
        "  plt.yticks([])\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ebmwYsannITS",
        "colab_type": "code",
        "outputId": "53a4fdbc-0851-4023-bdc3-ddbe67b49d7f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 535
        }
      },
      "source": [
        "# plot 20 random images from our training set\n",
        "num_rows = 5\n",
        "num_cols = 4\n",
        "num_images = num_rows*num_cols\n",
        "plt.figure(figsize=(2*num_cols, 2*num_rows))\n",
        "for i in range(num_images):\n",
        "  plt.subplot(num_rows, 2 *num_cols, 2*i + 1)\n",
        "  rand_index = random.randint(1,17000)\n",
        "  plot_image(rand_index)\n",
        "plt.show()"
      ],
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZoAAAIGCAYAAABtSknGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRAKlQCMiIoFSoBERkUAp0IiISKD+H/2tkR5wZ7hSAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 576x720 with 20 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "onZIXt48E8JZ",
        "colab_type": "text"
      },
      "source": [
        "# Model Building and training"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "12USDlgcGlxP",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# let's create a base model\n",
        "def cnn_model():\n",
        "  model = models.Sequential()\n",
        "  model.add(layers.Conv2D(32, (3,3), activation='relu', input_shape=(32,32,1)))\n",
        "  model.add(layers.MaxPooling2D((2,2)))\n",
        "  model.add(layers.Conv2D(64, (3,3), activation='relu'))\n",
        "  model.add(layers.MaxPooling2D((2,2)))\n",
        "  model.add(layers.Conv2D(64,(3,3), activation='relu'))\n",
        "  model.add(layers.Flatten())\n",
        "  model.add(layers.Dense(64, activation='relu'))\n",
        "  model.add(layers.Dense(10, activation='softmax'))\n",
        "  opt = optimizers.Adam(lr=0.0015, beta_1=0.9, beta_2=0.99, epsilon=None, decay=0.0, amsgrad=False)\n",
        "  model.compile(optimizer=opt, loss='categorical_crossentropy', metrics=['accuracy'])\n",
        "  return model\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1fhElWooqQpU",
        "colab_type": "code",
        "outputId": "3f4edd6b-62b1-45a4-c0a6-b05d6aa08c7c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 408
        }
      },
      "source": [
        "model = cnn_model()\n",
        "model.summary()"
      ],
      "execution_count": 56,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d_4 (Conv2D)            (None, 30, 30, 32)        320       \n",
            "_________________________________________________________________\n",
            "max_pooling2d_3 (MaxPooling2 (None, 15, 15, 32)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_5 (Conv2D)            (None, 13, 13, 64)        18496     \n",
            "_________________________________________________________________\n",
            "max_pooling2d_4 (MaxPooling2 (None, 6, 6, 64)          0         \n",
            "_________________________________________________________________\n",
            "conv2d_6 (Conv2D)            (None, 4, 4, 64)          36928     \n",
            "_________________________________________________________________\n",
            "flatten_2 (Flatten)          (None, 1024)              0         \n",
            "_________________________________________________________________\n",
            "dense_3 (Dense)              (None, 64)                65600     \n",
            "_________________________________________________________________\n",
            "dense_4 (Dense)              (None, 10)                650       \n",
            "=================================================================\n",
            "Total params: 121,994\n",
            "Trainable params: 121,994\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wPZzBTINHnuW",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# callbacks for saving models\n",
        "filepath = 'cnn-{val_acc: 5.3f}.h5'\n",
        "model_save = callbacks.ModelCheckpoint(filepath, monitor= 'val_acc', save_best_only=True)\n",
        "early_stop = callbacks.EarlyStopping(monitor='val_loss', patience =4)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vFEAKoplI4kP",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def plot_acc(history):\n",
        "  plt.plot(history.epoch, history.history['val_acc'] )\n",
        "  plt.plot(history.epoch, history.history['acc'])\n",
        "  plt.legend(['val_acc', 'train acc'])\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BrWkIv_xI5UP",
        "colab_type": "code",
        "outputId": "124837e3-9d70-4c62-d8ef-656f7f9c64cc",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 445
        }
      },
      "source": [
        "sample_size = train_digits_loader.n\n",
        "batch_size = 32\n",
        "\n",
        "hist = model.fit_generator(\n",
        "    train_digits_loader,\n",
        "    steps_per_epoch=sample_size // batch_size,\n",
        "    epochs=50,\n",
        "    validation_data=test_digits_loader,\n",
        "    validation_steps=3000 ,\n",
        "    callbacks=[model_save, early_stop]\n",
        ")"
      ],
      "execution_count": 59,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "531/531 [==============================] - 35s 66ms/step - loss: 0.2625 - acc: 0.9131 - val_loss: 0.0544 - val_acc: 0.9860\n",
            "Epoch 2/50\n",
            "531/531 [==============================] - 35s 67ms/step - loss: 0.0509 - acc: 0.9852 - val_loss: 0.0447 - val_acc: 0.9883\n",
            "Epoch 3/50\n",
            "531/531 [==============================] - 37s 69ms/step - loss: 0.0274 - acc: 0.9915 - val_loss: 0.0333 - val_acc: 0.9917\n",
            "Epoch 4/50\n",
            "531/531 [==============================] - 35s 65ms/step - loss: 0.0202 - acc: 0.9935 - val_loss: 0.0398 - val_acc: 0.9907\n",
            "Epoch 5/50\n",
            "531/531 [==============================] - 36s 68ms/step - loss: 0.0129 - acc: 0.9963 - val_loss: 0.0334 - val_acc: 0.9913\n",
            "Epoch 6/50\n",
            "531/531 [==============================] - 35s 66ms/step - loss: 0.0089 - acc: 0.9971 - val_loss: 0.0295 - val_acc: 0.9937\n",
            "Epoch 7/50\n",
            "531/531 [==============================] - 36s 68ms/step - loss: 0.0078 - acc: 0.9975 - val_loss: 0.0290 - val_acc: 0.9930\n",
            "Epoch 8/50\n",
            "531/531 [==============================] - 34s 64ms/step - loss: 0.0072 - acc: 0.9980 - val_loss: 0.0237 - val_acc: 0.9950\n",
            "Epoch 9/50\n",
            "531/531 [==============================] - 36s 68ms/step - loss: 0.0055 - acc: 0.9984 - val_loss: 0.0286 - val_acc: 0.9937\n",
            "Epoch 10/50\n",
            "531/531 [==============================] - 34s 65ms/step - loss: 0.0084 - acc: 0.9972 - val_loss: 0.0340 - val_acc: 0.9937\n",
            "Epoch 11/50\n",
            "531/531 [==============================] - 35s 65ms/step - loss: 0.0048 - acc: 0.9984 - val_loss: 0.0359 - val_acc: 0.9937\n",
            "Epoch 12/50\n",
            "531/531 [==============================] - 37s 69ms/step - loss: 0.0044 - acc: 0.9983 - val_loss: 0.0515 - val_acc: 0.9917\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mzi-zWvDJti_",
        "colab_type": "code",
        "outputId": "a1331358-cde7-4c64-af11-8562a7d7661c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        }
      },
      "source": [
        "plot_acc(hist)"
      ],
      "execution_count": 60,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAD8CAYAAACb4nSYAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "F8EtAKy6GR58",
        "colab_type": "text"
      },
      "source": [
        "# Model loading and evaluation"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9vnOypMa_86I",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "classifier = models.load_model('cnn- 0.995.h5')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yoHv8ilDtER2",
        "colab_type": "code",
        "outputId": "8de327ab-8be9-4585-9c5b-7db7c66d58a7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "classifier.evaluate_generator(test_digits_loader, steps = 3000)"
      ],
      "execution_count": 63,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.023661077855737612, 0.995]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 63
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_BUHlXF_b_vl",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# lets test for our test data\n",
        "test_digits_loader.reset()\n",
        "true_labels = test_digits_loader.classes\n",
        "predictions_with_prob = classifier.predict_generator(test_digits_loader, steps = 3000)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "51AV2A10ixJe",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "predictions = np.argmax(predictions_with_prob, axis= 1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oRqPF5NNxl_G",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "incorrect_predictions = np.nonzero(predictions - true_labels)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qunG_lX4iV2f",
        "colab_type": "code",
        "outputId": "37850465-5f2c-4cfa-e5a0-56b470d676dd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "incorrect_predictions"
      ],
      "execution_count": 76,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([ 428,  752,  910,  949, 1047, 1177, 1208, 1445, 1771, 1965, 1967,\n",
              "        2090, 2116, 2709, 2993]),)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 76
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "G1IufAekJZj0",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 187
        },
        "outputId": "3ff110fd-7dd7-4151-ac34-5f4910f301fd"
      },
      "source": [
        "for _ in range(10):\n",
        "  rand_index = random.randint(0,3000)\n",
        "  print(f'True label: {true_labels[rand_index]} , Predicted labe: {predictions[rand_index]} for index {rand_index}')"
      ],
      "execution_count": 82,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "True label: 8 , Predicted labe: 8 for index 2481\n",
            "True label: 3 , Predicted labe: 3 for index 993\n",
            "True label: 5 , Predicted labe: 5 for index 1706\n",
            "True label: 3 , Predicted labe: 3 for index 963\n",
            "True label: 7 , Predicted labe: 7 for index 2376\n",
            "True label: 4 , Predicted labe: 4 for index 1480\n",
            "True label: 2 , Predicted labe: 2 for index 768\n",
            "True label: 3 , Predicted labe: 3 for index 1032\n",
            "True label: 6 , Predicted labe: 6 for index 1862\n",
            "True label: 9 , Predicted labe: 9 for index 2931\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hhGgmdrMxQ_-",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# function to plot image with predictions percentages\n",
        "def plot_predicted_image(i, predictions_array, true_labels, img_generator):\n",
        "  predicted_label, true_label, img = predictions_array[i], true_labels[i], img_generator[i][0]\n",
        "  plt.grid(False)\n",
        "  plt.xticks([])\n",
        "  plt.yticks([])\n",
        "  reshaped_img = img.reshape(32,32)\n",
        "  plt.imshow(reshaped_img, cmap=plt.cm.binary)\n",
        "  plt.title(f' Predicted {true_label } for {predicted_label}  digit')\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_SZRnYXkRNDn",
        "colab_type": "code",
        "outputId": "8532bbb5-a51f-47ad-a24c-82bfc18e5707",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 268
        }
      },
      "source": [
        "plot_predicted_image(100, predictions, true_labels, test_digits_loader)"
      ],
      "execution_count": 91,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAO4AAAD7CAYAAABt9agKAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "neSiyqwKApNj",
        "colab_type": "code",
        "outputId": "b1da27b7-073e-47f8-fd5b-2f3cfeeeb964",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 485
        }
      },
      "source": [
        "num_rows = 4\n",
        "num_cols = 3\n",
        "num_images = num_rows*num_cols\n",
        "plt.figure(figsize=(2*2*num_cols, 2*num_rows))\n",
        "for i in range(num_images):\n",
        "  plt.subplot(num_rows, 2 *num_cols, 2*i + 1)\n",
        "  rand_idx = random.randint(0,3000)\n",
        "  plot_predicted_image(rand_idx, predictions, true_labels, test_digits_loader)\n",
        "plt.show()"
      ],
      "execution_count": 99,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmQAAAHUCAYAAACUMzRUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nh4yIiIgoYP8PhEtC82CO2f8AAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 864x576 with 12 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NL2Yw5n5EjOF",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TwOqKg7KRaCL",
        "colab_type": "text"
      },
      "source": [
        "# Let's plot our incorrect prediction"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pwFXV_sXRiFh",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 594
        },
        "outputId": "daf5cc3c-f3e4-4e08-a03e-3453cc4ed839"
      },
      "source": [
        "num_rows = 5\n",
        "num_cols = 3\n",
        "num_images = num_rows*num_cols\n",
        "plt.figure(figsize=(2*2*num_cols, 2*num_rows))\n",
        "for i in range(num_images):\n",
        "  plt.subplot(num_rows, 2 *num_cols, 2*i + 1)\n",
        "  plot_predicted_image(incorrect_predictions[0][i], predictions, true_labels, test_digits_loader)\n",
        "plt.show()"
      ],
      "execution_count": 101,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmQAAAJBCAYAAADocKk5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\niIiIiELGEzIiIiKikP0/RgD5RGp+FHUAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 864x720 with 15 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AkhRvINaSAUU",
        "colab_type": "text"
      },
      "source": [
        "It looks like our dataset has some mislabels. We can remove those to get better models."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sJ1xZ2G1R0rl",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}