B2.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "B2.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true,
      "mount_file_id": "13sbCHt_uGN8c2vDgWX7Bob3COTePHqiI",
      "authorship_tag": "ABX9TyNFBDQn7MNq/X3eresndpQW"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "51CDvBXAB_oZ"
      },
      "source": [
        "import tensorflow as tf\n",
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aWYoSwMmIVDF"
      },
      "source": [
        "data = './'"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uYKicRJxDvlm",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "2b85ac15-26b9-4692-8f0d-f7828ac331ae"
      },
      "source": [
        "!pip install wget\n",
        "import wget\n",
        "link = 'https://d17h27t6h515a5.cloudfront.net/topher/2017/February/5898cd6f_traffic-signs-data/traffic-signs-data.zip'\n",
        "wget.download(link)\n",
        "\n",
        "!unzip -q traffic-signs-data.zip -d $data"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting wget\n",
            "  Downloading https://files.pythonhosted.org/packages/47/6a/62e288da7bcda82b935ff0c6cfe542970f04e29c756b0e147251b2fb251f/wget-3.2.zip\n",
            "Building wheels for collected packages: wget\n",
            "  Building wheel for wget (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for wget: filename=wget-3.2-cp36-none-any.whl size=9682 sha256=0094cf573bce77cc7797ca5a712e21cd45c3555de06bf6f5cd440ec0fbc28dc0\n",
            "  Stored in directory: /root/.cache/pip/wheels/40/15/30/7d8f7cea2902b4db79e3fea550d7d7b85ecb27ef992b618f3f\n",
            "Successfully built wget\n",
            "Installing collected packages: wget\n",
            "Successfully installed wget-3.2\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kQNtGJZEDy7N"
      },
      "source": [
        "import pickle\n",
        "\n",
        "train_link = data + 'train.p'\n",
        "valid_link = data + 'valid.p'\n",
        "test_link = data + 'test.p'\n",
        "\n",
        "with open(train_link, mode='rb') as f:\n",
        "    train = pickle.load(f)\n",
        "with open(valid_link, mode='rb') as f:\n",
        "    valid = pickle.load(f)\n",
        "with open(test_link, mode='rb') as f:\n",
        "    test = pickle.load(f)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MwA81k8WFOwi"
      },
      "source": [
        "from sklearn.utils import shuffle\n",
        "\n",
        "train_X, train_y = shuffle(train['features'], train['labels'])\n",
        "valid_X, valid_y = valid['features'], valid['labels']\n",
        "test_X, test_y = test['features'], test['labels']"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9dPp8nfJM8lI",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4026f579-9a78-4269-d9bb-ac79200baa2a"
      },
      "source": [
        "train_X.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(34799, 32, 32, 3)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GEwd-89oGScv"
      },
      "source": [
        "classNames = {0: 'Speed limit (20km/h)',\n",
        " 1: 'Speed limit (30km/h)',\n",
        " 2: 'Speed limit (50km/h)',\n",
        " 3: 'Speed limit (60km/h)',\n",
        " 4: 'Speed limit (70km/h)',\n",
        " 5: 'Speed limit (80km/h)',\n",
        " 6: 'End of speed limit (80km/h)',\n",
        " 7: 'Speed limit (100km/h)',\n",
        " 8: 'Speed limit (120km/h)',\n",
        " 9: 'No passing',\n",
        " 10: 'No passing for vehicles over 3.5 metric tons',\n",
        " 11: 'Right-of-way at the next intersection',\n",
        " 12: 'Priority road',\n",
        " 13: 'Yield',\n",
        " 14: 'Stop',\n",
        " 15: 'No vehicles',\n",
        " 16: 'Vehicles over 3.5 metric tons prohibited',\n",
        " 17: 'No entry',\n",
        " 18: 'General caution',\n",
        " 19: 'Dangerous curve to the left',\n",
        " 20: 'Dangerous curve to the right',\n",
        " 21: 'Double curve',\n",
        " 22: 'Bumpy road',\n",
        " 23: 'Slippery road',\n",
        " 24: 'Road narrows on the right',\n",
        " 25: 'Road work',\n",
        " 26: 'Traffic signals',\n",
        " 27: 'Pedestrians',\n",
        " 28: 'Children crossing',\n",
        " 29: 'Bicycles crossing',\n",
        " 30: 'Beware of ice/snow',\n",
        " 31: 'Wild animals crossing',\n",
        " 32: 'End of all speed and passing limits',\n",
        " 33: 'Turn right ahead',\n",
        " 34: 'Turn left ahead',\n",
        " 35: 'Ahead only',\n",
        " 36: 'Go straight or right',\n",
        " 37: 'Go straight or left',\n",
        " 38: 'Keep right',\n",
        " 39: 'Keep left',\n",
        " 40: 'Roundabout mandatory',\n",
        " 41: 'End of no passing',\n",
        " 42: 'End of no passing by vehicles over 3.5 metric tons'}"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bBAMoGSbHAY_"
      },
      "source": [
        "train_X = train_X.astype(\"float\") / 255.0\n",
        "valid_X = valid_X.astype(\"float\") / 255.0\n",
        "test_X = test_X.astype(\"float\") / 255.0"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8Yn3D0vZFWXA"
      },
      "source": [
        "from sklearn.preprocessing import LabelBinarizer\n",
        "\n",
        "lb = LabelBinarizer()\n",
        "train_y = lb.fit_transform(train_y)\n",
        "valid_y = lb.fit_transform(valid_y)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_JHT0AcUL5sY"
      },
      "source": [
        "from tensorflow.keras.models import Model\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import BatchNormalization\n",
        "from tensorflow.keras.layers import AveragePooling2D\n",
        "from tensorflow.keras.layers import MaxPooling2D\n",
        "from tensorflow.keras.layers import Conv2D\n",
        "from tensorflow.keras.layers import Activation\n",
        "from tensorflow.keras.layers import Dropout\n",
        "from tensorflow.keras.layers import Flatten\n",
        "from tensorflow.keras.layers import Input\n",
        "from tensorflow.keras.layers import Dense\n",
        "from tensorflow.keras.layers import concatenate\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
        "from tensorflow.keras.optimizers import SGD"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YLym8eDsND3C"
      },
      "source": [
        "# Khởi tạo model\n",
        "model = Sequential()\n",
        "\n",
        "width = 32\n",
        "height = 32\n",
        "classes = 43\n",
        "\n",
        "shape = (width, height, 3)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EJTpFt3WoEo3"
      },
      "source": [
        "model.add(Conv2D(32, (3, 3), padding=\"same\", input_shape=shape))\n",
        "model.add(Activation(\"relu\"))\n",
        "model.add(BatchNormalization())"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6QUHSQ2tpK0X"
      },
      "source": [
        "model.add(Conv2D(32, (3, 3), padding=\"same\"))\n",
        "model.add(Activation(\"relu\"))\n",
        "model.add(BatchNormalization())"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Lig0XYh0pg93"
      },
      "source": [
        "model.add(MaxPooling2D(pool_size=(2, 2)))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4dRBvpDGqKbf"
      },
      "source": [
        "model.add(Conv2D(64, (3, 3), padding=\"same\"))\n",
        "model.add(Activation(\"relu\"))\n",
        "model.add(BatchNormalization())"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Hubwb4ttrG5d"
      },
      "source": [
        "model.add(Conv2D(64, (3, 3), padding=\"same\"))\n",
        "model.add(Activation(\"relu\"))\n",
        "model.add(BatchNormalization())"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yZeR1K0LrwIF"
      },
      "source": [
        "model.add(MaxPooling2D(pool_size=(2, 2)))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XnalvAV4sg02"
      },
      "source": [
        "model.add(Flatten())"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NcwlZJ5PszpF"
      },
      "source": [
        "model.add(Dense(512))\n",
        "\n",
        "# Cần có hàm activation sau hàm Dense\n",
        "model.add(Activation(\"relu\"))\n",
        "model.add(BatchNormalization())"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8sp4wCMitJaU"
      },
      "source": [
        "model.add(Dense(classes))\n",
        "model.add(Activation(\"softmax\"))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_3D4BrifoMzW",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f5c3593a-cf85-432c-dade-89367aa8a191"
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d (Conv2D)              (None, 32, 32, 32)        896       \n",
            "_________________________________________________________________\n",
            "activation (Activation)      (None, 32, 32, 32)        0         \n",
            "_________________________________________________________________\n",
            "batch_normalization (BatchNo (None, 32, 32, 32)        128       \n",
            "_________________________________________________________________\n",
            "conv2d_1 (Conv2D)            (None, 32, 32, 32)        9248      \n",
            "_________________________________________________________________\n",
            "activation_1 (Activation)    (None, 32, 32, 32)        0         \n",
            "_________________________________________________________________\n",
            "batch_normalization_1 (Batch (None, 32, 32, 32)        128       \n",
            "_________________________________________________________________\n",
            "max_pooling2d (MaxPooling2D) (None, 16, 16, 32)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_2 (Conv2D)            (None, 16, 16, 64)        18496     \n",
            "_________________________________________________________________\n",
            "activation_2 (Activation)    (None, 16, 16, 64)        0         \n",
            "_________________________________________________________________\n",
            "batch_normalization_2 (Batch (None, 16, 16, 64)        256       \n",
            "_________________________________________________________________\n",
            "conv2d_3 (Conv2D)            (None, 16, 16, 64)        36928     \n",
            "_________________________________________________________________\n",
            "activation_3 (Activation)    (None, 16, 16, 64)        0         \n",
            "_________________________________________________________________\n",
            "batch_normalization_3 (Batch (None, 16, 16, 64)        256       \n",
            "_________________________________________________________________\n",
            "max_pooling2d_1 (MaxPooling2 (None, 8, 8, 64)          0         \n",
            "_________________________________________________________________\n",
            "flatten (Flatten)            (None, 4096)              0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 512)               2097664   \n",
            "_________________________________________________________________\n",
            "activation_4 (Activation)    (None, 512)               0         \n",
            "_________________________________________________________________\n",
            "batch_normalization_4 (Batch (None, 512)               2048      \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 43)                22059     \n",
            "_________________________________________________________________\n",
            "activation_5 (Activation)    (None, 43)                0         \n",
            "=================================================================\n",
            "Total params: 2,188,107\n",
            "Trainable params: 2,186,699\n",
            "Non-trainable params: 1,408\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UFqyUz9KoO3J"
      },
      "source": [
        "# Làm giàu ảnh\n",
        "aug = ImageDataGenerator(rotation_range=0.18, zoom_range=0.15, width_shift_range=0.2, height_shift_range=0.2, horizontal_flip=True)\n",
        "learning_rate = 0.01\n",
        "epochs = 10\n",
        "batch_size = 64\n",
        "opt = SGD(learning_rate=learning_rate, momentum=0.9)\n",
        "\n",
        "model.compile(optimizer=opt, loss='categorical_crossentropy', metrics=['accuracy'])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "U50MmK4sNurm",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ec5cbfed-aa95-4858-be5f-fc3506a657bc"
      },
      "source": [
        "# Training\n",
        "H = model.fit_generator(aug.flow(train_X, train_y, batch_size=batch_size), validation_data=(valid_X, valid_y), \n",
        "                        steps_per_epoch=train_X.shape[0]//batch_size, epochs=epochs, verbose=1)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:From <ipython-input-22-03ec743228e6>:3: Model.fit_generator (from tensorflow.python.keras.engine.training) is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Please use Model.fit, which supports generators.\n",
            "Epoch 1/10\n",
            "543/543 [==============================] - 24s 45ms/step - loss: 2.0550 - accuracy: 0.4024 - val_loss: 1.7803 - val_accuracy: 0.4760\n",
            "Epoch 2/10\n",
            "543/543 [==============================] - 24s 44ms/step - loss: 0.9049 - accuracy: 0.6973 - val_loss: 1.2180 - val_accuracy: 0.6231\n",
            "Epoch 3/10\n",
            "543/543 [==============================] - 24s 44ms/step - loss: 0.4755 - accuracy: 0.8426 - val_loss: 0.9270 - val_accuracy: 0.7288\n",
            "Epoch 4/10\n",
            "543/543 [==============================] - 24s 44ms/step - loss: 0.2685 - accuracy: 0.9119 - val_loss: 0.9639 - val_accuracy: 0.7481\n",
            "Epoch 5/10\n",
            "543/543 [==============================] - 24s 44ms/step - loss: 0.1736 - accuracy: 0.9436 - val_loss: 0.8498 - val_accuracy: 0.8005\n",
            "Epoch 6/10\n",
            "543/543 [==============================] - 24s 44ms/step - loss: 0.1227 - accuracy: 0.9607 - val_loss: 0.7656 - val_accuracy: 0.7961\n",
            "Epoch 7/10\n",
            "543/543 [==============================] - 24s 44ms/step - loss: 0.1037 - accuracy: 0.9674 - val_loss: 0.5537 - val_accuracy: 0.8431\n",
            "Epoch 8/10\n",
            "543/543 [==============================] - 24s 44ms/step - loss: 0.0802 - accuracy: 0.9739 - val_loss: 0.5808 - val_accuracy: 0.8494\n",
            "Epoch 9/10\n",
            "543/543 [==============================] - 24s 43ms/step - loss: 0.0672 - accuracy: 0.9784 - val_loss: 0.4714 - val_accuracy: 0.8655\n",
            "Epoch 10/10\n",
            "543/543 [==============================] - 24s 44ms/step - loss: 0.0599 - accuracy: 0.9808 - val_loss: 0.5417 - val_accuracy: 0.8653\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ah1POKO1O9yp"
      },
      "source": [
        "# Lưu model\n",
        "model.save('B2.h5') # h5 -- keras"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TCj7dmviPQ9N"
      },
      "source": [
        "# Sử dụng model có sẵn\n",
        "saved_model = tf.keras.models.load_model(data + 'B2.h5')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3IrQsJA9PpPc"
      },
      "source": [
        "result = saved_model.predict(test_X[1:2])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9njvtOt7QZmg"
      },
      "source": [
        "import numpy as np\n",
        "\n",
        "final = np.argmax(result)\n",
        "final = classNames[final]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "becFV3GnQjHZ",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "904810e0-da14-4760-b479-df1e8c0e605c"
      },
      "source": [
        "final"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'Speed limit (30km/h)'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 28
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MGp4djw3QytA",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 283
        },
        "outputId": "289406b4-28bf-4873-fa03-f340b3f807af"
      },
      "source": [
        "plt.imshow(test['features'][1])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7efbc4607128>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 29
        },
        {
          "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": "6NKl2m1KSMih"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}