Tensorflow-Keras.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Tensorflow Graph "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "## https://developers.google.com/machine-learning/crash-course/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "a"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Tensorflow Eager Execution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import __future__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TensorFlow version: 1.8.0\n",
      "Eager execution: True\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import tensorflow as tf\n",
    "import tensorflow.contrib.eager as tfe\n",
    "\n",
    "tf.enable_eager_execution()\n",
    "\n",
    "print(\"TensorFlow version: {}\".format(tf.VERSION))\n",
    "print(\"Eager execution: {}\".format(tf.executing_eagerly()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "x = [[1], [2], [3]]\n",
    "y = [[2], [3], [4]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "m = tf.matmul(y, tf.transpose(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tf.Tensor: id=10, shape=(3, 3), dtype=int32, numpy=\n",
       "array([[ 2,  4,  6],\n",
       "       [ 3,  6,  9],\n",
       "       [ 4,  8, 12]], dtype=int32)>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = [\n",
    "     [1, 2, 3], \n",
    "     [4, 5, 6], \n",
    "     [7, 8, 9], \n",
    "     [10, 11, 12]\n",
    "    ]\n",
    "# x shape is 4 * 3\n",
    "# assuming we need convert X to X' which shape is 4 * 5\n",
    "# then we need a weights 3 * 5\n",
    "\n",
    "weights = [\n",
    "    [1, 1, 1], \n",
    "    [2, 2, 2], \n",
    "    [3, 3, 3]\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "bias = [1, 2, 3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "new_X = tf.matmul(X, weights) + bias"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tf.Tensor: id=16, shape=(4, 3), dtype=int32, numpy=\n",
       "array([[15, 16, 17],\n",
       "       [33, 34, 35],\n",
       "       [51, 52, 53],\n",
       "       [69, 70, 71]], dtype=int32)>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "w = tf.contrib.eager.Variable([1.0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "with tf.GradientTape() as tape: \n",
    "    loss = w * w"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "grad = tape.gradient(loss, [w])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<tf.Tensor: id=35, shape=(1,), dtype=float32, numpy=array([ 2.], dtype=float32)>]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grad"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<img src='https://www.tensorflow.org/images/opt1.gif'></img>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "file_src = '../dataset/'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "example features: tf.Tensor([ 5.80000019  2.5999999   4.          1.20000005], shape=(4,), dtype=float32)\n",
      "example label: tf.Tensor(1, shape=(), dtype=int32)\n"
     ]
    }
   ],
   "source": [
    "train_dataset_url = '{}iris_training.csv'.format(file_src)\n",
    "\n",
    "train_dataset = tf.data.TextLineDataset(train_dataset_url)\n",
    "train_dataset = train_dataset.skip(1)             # skip the first header row\n",
    "train_dataset = train_dataset.map(parse_csv)      # parse each row\n",
    "train_dataset = train_dataset.shuffle(buffer_size=1000)  # randomize\n",
    "train_dataset = train_dataset.batch(32)\n",
    "\n",
    "# View a single example entry from a batch\n",
    "features, label = iter(train_dataset).next()\n",
    "print(\"example features:\", features[0])\n",
    "print(\"example label:\", label[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "def parse_csv(line):\n",
    "    example_defaults = [[0.], [0.], [0.], [0.], [0]]  # sets field types\n",
    "    parsed_line = tf.decode_csv(line, example_defaults)\n",
    "    # First 4 fields are features, combine into single tensor\n",
    "    features = tf.reshape(parsed_line[:-1], shape=(4,))\n",
    "    # Last field is the label\n",
    "    label = tf.reshape(parsed_line[-1], shape=())\n",
    "    return features, label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = tf.keras.Sequential([\n",
    "  tf.keras.layers.Dense(10, activation=\"relu\", input_shape=(4,)),  # input shape required\n",
    "  tf.keras.layers.Dense(10, activation=\"relu\"),\n",
    "  tf.keras.layers.Dense(3)\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "def loss(model, x, y):\n",
    "    y_ = model(x)\n",
    "    return tf.losses.sparse_softmax_cross_entropy(labels=y, logits=y_)\n",
    "\n",
    "def grad(model, inputs, targets):\n",
    "    with tf.GradientTape() as tape:\n",
    "        loss_value = loss(model, inputs, targets)\n",
    "    return tape.gradient(loss_value, model.variables)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "optimizer = tf.train.GradientDescentOptimizer(learning_rate=0.01)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 000: Loss: 1.013, Accuracy: 53.333%\n",
      "Epoch 050: Loss: 0.436, Accuracy: 80.000%\n",
      "Epoch 100: Loss: 0.298, Accuracy: 95.833%\n",
      "Epoch 150: Loss: 0.201, Accuracy: 96.667%\n",
      "Epoch 200: Loss: 0.155, Accuracy: 97.500%\n"
     ]
    }
   ],
   "source": [
    "# keep results for plotting\n",
    "train_loss_results = []\n",
    "train_accuracy_results = []\n",
    "\n",
    "num_epochs = 201\n",
    "\n",
    "for epoch in range(num_epochs):\n",
    "    epoch_loss_avg = tfe.metrics.Mean()\n",
    "    epoch_accuracy = tfe.metrics.Accuracy()\n",
    "\n",
    "    # Training loop - using batches of 32\n",
    "    for x, y in train_dataset:\n",
    "    # Optimize the model\n",
    "        grads = grad(model, x, y)\n",
    "        optimizer.apply_gradients(zip(grads, model.variables),\n",
    "                                  global_step=tf.train.get_or_create_global_step())\n",
    "\n",
    "        # Track progress\n",
    "        epoch_loss_avg(loss(model, x, y))  # add current batch loss\n",
    "        # compare predicted label to actual label\n",
    "        epoch_accuracy(tf.argmax(model(x), axis=1, output_type=tf.int32), y)\n",
    "\n",
    "        # end epoch\n",
    "        train_loss_results.append(epoch_loss_avg.result())\n",
    "        train_accuracy_results.append(epoch_accuracy.result())\n",
    "\n",
    "    if epoch % 50 == 0:\n",
    "        print(\"Epoch {:03d}: Loss: {:.3f}, Accuracy: {:.3%}\".format(epoch,\n",
    "                                                                epoch_loss_avg.result(),\n",
    "                                                                epoch_accuracy.result()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axes = plt.subplots(2, sharex=True, figsize=(12, 8))\n",
    "fig.suptitle('Training Metrics')\n",
    "\n",
    "axes[0].set_ylabel(\"Loss\", fontsize=14)\n",
    "axes[0].plot(train_loss_results)\n",
    "\n",
    "axes[1].set_ylabel(\"Accuracy\", fontsize=14)\n",
    "axes[1].set_xlabel(\"Epoch\", fontsize=14)\n",
    "axes[1].plot(train_accuracy_results)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1. tensorflow\n",
    "    : eager\n",
    "    : graph\n",
    "    : estimator\n",
    "    : self defined function\n",
    "    : serving\n",
    "    : distributation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 改"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 改"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1.  背景、处理什么问题\n",
    "## 2.  以前方法缺点， 它怎么牛逼？ \n",
    "## 3.  理论基础 理论假设\n",
    "## 4.  实验如何设计以及如何支持(support ) 作者的假设\n",
    "## 5.  他这个东西， 对我有什么帮助， 和我有什么关系\n",
    "## 6.  他说的对吗？ 有没有什么问题？ "
   ]
  }
 ],
 "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": "3.6.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}