tf2.0/setup-and-first-code.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Cài đặt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "https://www.tensorflow.org/guide/effective_tf2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# !pip3 install tensorflow==2.0.0-rc0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow import keras"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Kiểm tra vesion"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.0.0-rc0\n"
     ]
    }
   ],
   "source": [
    "print(tf.__version__)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "mnist = tf.keras.datasets.fashion_mnist"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "(x_train, y_train), (x_test, y_test) = mnist.load_data()\n",
    "x_train, x_test = x_train / 255.0, x_test / 255.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Chú ý:  Nhập command bên dưới vào terminal trong trường hợp không load được dữ liệu về\n",
    "\n",
    "# /Applications/Python\\ 3.7/Install\\ Certificates.command"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 28, 28)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_train.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<img src=\"https://github.com/zalandoresearch/fashion-mnist/blob/master/doc/img/fashion-mnist-sprite.png?raw=true\" width=400/>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Có thể nhận thấy features là một ma trận 3 chiều tương ứng với 60000 cái ảnh 2 chiều (28 * 28)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000,)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Tương ứng ta có 60000 nhãn với 10 loại từ 0 -> 9 đaị diện cho 10 loại quần áo khác nhau."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "min(y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "9"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "max(y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "| Nhãn | Miêu tả |\n",
    "| --- | --- |\n",
    "| 0 | T-shirt/top |\n",
    "| 1 | Trouser |\n",
    "| 2 | Pullover |\n",
    "| 3 | Dress |\n",
    "| 4 | Coat |\n",
    "| 5 | Sandal |\n",
    "| 6 | Shirt |\n",
    "| 7 | Sneaker |\n",
    "| 8 | Bag |\n",
    "| 9 | Ankle boot |"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Sử dụng dictionary để lưu ánh xạ dữ liệu"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "labels = {0: \"T-shirt/top\", 1: \"Trouser\", 2 : \"Pullover\", 3:\"Dress\", 4: \"Coat\", 5: \"Sandal\", 6: \"Shirt\", 7: \"Sneaker\", 8: \"Bag\", 9: \"Ankle boot\"}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{0: 'T-shirt/top',\n",
       " 1: 'Trouser',\n",
       " 2: 'Pullover',\n",
       " 3: 'Dress',\n",
       " 4: 'Coat',\n",
       " 5: 'Sandal',\n",
       " 6: 'Shirt',\n",
       " 7: 'Sneaker',\n",
       " 8: 'Bag',\n",
       " 9: 'Ankle boot'}"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labels"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Nhiệm vụ của chúng ta là phải train được model có thể mappting từ features đến những nhãn này"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Ví dụ đây là một feature"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.colorbar.Colorbar at 0x151ab9f90>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure()\n",
    "plt.imshow(x_train[20])\n",
    "plt.colorbar()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Với nhãn là:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train[20]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Theo bảng trên ta có thể thấy đây là một cái váy 3 -> Dress"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Train Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras import Sequential"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.layers import Flatten, Dense"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1. Khởi tạo một Neural network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Sequential()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "2. Thêm lớp nhận đầu vào có kích cỡ (28*28)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.add(Flatten(input_shape = (28, 28)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Flatten ở đây có nhiệm vụ duỗi ảnh từ ma trận (28 * 28) thành (784 * 1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "3. Thêm một lớp ẩn có cỡ (28 * 1). Hàm phi tuyến ReLU"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.add(Dense(28, activation=\"relu\"))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "4. Thêm lớp softmax để thu về đầu ra là 10 lóp như mong muốn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.add(Dense(10, activation=\"softmax\"))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "5. Kiểm tra lại model trước khi train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "flatten (Flatten)            (None, 784)               0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 28)                21980     \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 10)                290       \n",
      "=================================================================\n",
      "Total params: 22,270\n",
      "Trainable params: 22,270\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "6. Loss function: Hàm mất mát thể hiện độ sai lệch dự đoán model với nhãn thực tế"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(optimizer='adam', loss=\"sparse_categorical_crossentropy\", metric=['accuracy'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "7. Tiến hành train model trong 10 vòng lặp để tối ưu hoá loss function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 60000 samples\n",
      "Epoch 1/10\n",
      "60000/60000 [==============================] - 3s 55us/sample - loss: 0.5608\n",
      "Epoch 2/10\n",
      "60000/60000 [==============================] - 2s 41us/sample - loss: 0.4150\n",
      "Epoch 3/10\n",
      "60000/60000 [==============================] - 2s 41us/sample - loss: 0.3816\n",
      "Epoch 4/10\n",
      "60000/60000 [==============================] - 2s 40us/sample - loss: 0.3644\n",
      "Epoch 5/10\n",
      "60000/60000 [==============================] - 2s 41us/sample - loss: 0.3479\n",
      "Epoch 6/10\n",
      "60000/60000 [==============================] - 2s 41us/sample - loss: 0.3365\n",
      "Epoch 7/10\n",
      "60000/60000 [==============================] - 3s 48us/sample - loss: 0.3263\n",
      "Epoch 8/10\n",
      "60000/60000 [==============================] - 3s 43us/sample - loss: 0.3190\n",
      "Epoch 9/10\n",
      "60000/60000 [==============================] - 3s 46us/sample - loss: 0.3123\n",
      "Epoch 10/10\n",
      "60000/60000 [==============================] - 3s 44us/sample - loss: 0.3037\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x125a13110>"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(x_train, y_train, epochs = 10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "8. Kiểm tra độ chính xác hiện tại"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import accuracy_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_predicted = model.predict_classes(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8618"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy_score(y_predicted, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Bản chất y_predicted là một vector có chiều bằng y_test. y_predicted bao gồm các nhãn được dự đoán bởi mô hình đã được train. y_test bao gồm các nhãn chính xác của data. Sự so sánh độ tương đồng của 2 vector này thể hiện độ chính xác của mô hình."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<img src=\"./1.inference.jpg\" />"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "9. Quá trình inference (Truyền data không được train vào mô hình)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_inference = x_test[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.colorbar.Colorbar at 0x133fd6890>"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure()\n",
    "plt.imshow(x_inference)\n",
    "plt.colorbar()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_predicted = model.predict_classes(np.array([x_inference]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "9"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_predicted[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Ankle boot'"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labels[y_predicted[0]]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Vậy ta có thể thấy với độ chính xác như trên mô hình có thể dự đoán chính xác được hình ảnh này là một đôi giầy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}