feeding_data_totf_part1.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## <font color=brown>Simple Graph Execution</font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "## import necessary stuff\n",
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "import os,sys\n",
    "from tensorflow.keras.datasets import mnist\n",
    "\n",
    "import time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "## load data of mnist from keras. alternatively you can load data from tf.contrib.learn.datasets.\n",
    "(x_train,y_train),(x_test,y_test) = mnist.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train = x_train.reshape(-1,784)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Building simple DNN Model and feeding the numpy/pandas data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Simple MNIST Model of Dense layers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Using default config.\n",
      "WARNING:tensorflow:Using temporary folder as model directory: /tmp/tmpap27xr75\n",
      "INFO:tensorflow:Using config: {'_model_dir': '/tmp/tmpap27xr75', '_tf_random_seed': None, '_save_summary_steps': 100, '_save_checkpoints_steps': None, '_save_checkpoints_secs': 600, '_session_config': None, '_keep_checkpoint_max': 5, '_keep_checkpoint_every_n_hours': 10000, '_log_step_count_steps': 100, '_train_distribute': None, '_device_fn': None, '_service': None, '_cluster_spec': <tensorflow.python.training.server_lib.ClusterSpec object at 0x7fa4160c9080>, '_task_type': 'worker', '_task_id': 0, '_global_id_in_cluster': 0, '_master': '', '_evaluation_master': '', '_is_chief': True, '_num_ps_replicas': 0, '_num_worker_replicas': 1}\n"
     ]
    }
   ],
   "source": [
    "## defining the type of features columns to be used on model.\n",
    "feature_column = [tf.feature_column.numeric_column(key='image',shape=(784,))]\n",
    "\n",
    "##defining the model\n",
    "model = tf.estimator.DNNClassifier([100,100],n_classes=10,feature_columns=feature_column)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Defining the dataset flow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''parse function to be used. This function is needed to do \n",
    "the preprocessing of data like reshaping ,\n",
    "converting to tensors from numpy arrays ,one-hot encoding ,etc.'''\n",
    "def _parse_(x,y):\n",
    "    x = tf.reshape(x,(784,)) #reshape to 784 as expected.\n",
    "    x = tf.cast(x,tf.float32) #cast to float32 as the weights are float32.\n",
    "    # y = tf.one_hot(y,10) #should be one hot encoded but tensorflow is throwing error .idky??\n",
    "    y = tf.cast(y,tf.int32) #cast to tensor of int32\n",
    "    return (dict({'image':x}),y) #return tuple of dict of feature name with value and label."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "##define the function that feeds the data to the model .\n",
    "def train_input_fn(x_train,y_train,batch_size=64):\n",
    "    ##Here we are using dataset API.\n",
    "    '''\n",
    "    take the data from tensor_slices i.e. an array of datapoints in simple words.\n",
    "    '''\n",
    "    dataset = tf.data.Dataset.from_tensor_slices((x_train,y_train)) \n",
    "    \n",
    "    \n",
    "    dataset = dataset.map(lambda x,y:_parse_(x,y)).shuffle(buffer_size=128) \\\n",
    "                .batch(batch_size).make_one_shot_iterator()\n",
    "    \n",
    "    return dataset.get_next()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Training the model by feeding data through the above defined flow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Calling model_fn.\n",
      "INFO:tensorflow:Done calling model_fn.\n",
      "INFO:tensorflow:Create CheckpointSaverHook.\n",
      "INFO:tensorflow:Graph was finalized.\n",
      "INFO:tensorflow:Running local_init_op.\n",
      "INFO:tensorflow:Done running local_init_op.\n",
      "INFO:tensorflow:Saving checkpoints for 0 into /tmp/tmpap27xr75/model.ckpt.\n",
      "INFO:tensorflow:loss = 8941.259, step = 0\n",
      "INFO:tensorflow:global_step/sec: 240.438\n",
      "INFO:tensorflow:loss = 101.23705, step = 100 (0.417 sec)\n",
      "INFO:tensorflow:Saving checkpoints for 150 into /tmp/tmpap27xr75/model.ckpt.\n",
      "INFO:tensorflow:Loss for final step: 125.124916.\n",
      "---------18.832788705825806\n"
     ]
    }
   ],
   "source": [
    "t1 = time.time()\n",
    "model.train(input_fn=lambda:train_input_fn(x_train,y_train,64),steps=150)\n",
    "t2= time.time()\n",
    "print('---------{}'.format(t2 - t1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "t1 = time.time()\n",
    "model.train(input_fn=lambda:train_input_fn(x_train,y_train,64),steps=251)\n",
    "t2= time.time()\n",
    "print('---------{}'.format(t2 - t1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Train using the numpy input function**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Calling model_fn.\n",
      "INFO:tensorflow:Done calling model_fn.\n",
      "INFO:tensorflow:Create CheckpointSaverHook.\n",
      "INFO:tensorflow:Graph was finalized.\n",
      "INFO:tensorflow:Restoring parameters from /tmp/tmpbqg7sg1j/model.ckpt-102\n",
      "INFO:tensorflow:Running local_init_op.\n",
      "INFO:tensorflow:Done running local_init_op.\n",
      "INFO:tensorflow:Saving checkpoints for 102 into /tmp/tmpbqg7sg1j/model.ckpt.\n",
      "INFO:tensorflow:loss = 174.48169, step = 102\n",
      "INFO:tensorflow:global_step/sec: 642.66\n",
      "INFO:tensorflow:loss = 149.84337, step = 202 (0.157 sec)\n",
      "INFO:tensorflow:Saving checkpoints for 203 into /tmp/tmpbqg7sg1j/model.ckpt.\n",
      "INFO:tensorflow:Loss for final step: 149.84337.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.estimator.canned.dnn.DNNClassifier at 0x7f5f2706c208>"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.train(input_fn=tf.estimator.inputs.numpy_input_fn(dict({'image':x_train}),\n",
    "                                                        np.array(y_train,np.int32),shuffle=True),steps=101)"
   ]
  },
  {
   "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}