{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "\n",
    "import tensorflow as tf\n",
    "import gradio"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "n_classes = 10\n",
    "(x_train, y_train),(x_test, y_test) = tf.keras.datasets.mnist.load_data()\n",
    "x_train, x_test = x_train.reshape(-1, 784) / 255.0, x_test.reshape(-1, 784) / 255.0\n",
    "y_train = tf.keras.utils.to_categorical(y_train, n_classes).astype(float)\n",
    "y_test = tf.keras.utils.to_categorical(y_test, n_classes).astype(float)\n",
    "\n",
    "learning_rate = 0.5\n",
    "epochs = 5\n",
    "batch_size = 100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From C:\\Users\\ALI\\Anaconda3\\lib\\site-packages\\tensorflow\\python\\framework\\op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Colocations handled automatically by placer.\n"
     ]
    }
   ],
   "source": [
    "x = tf.placeholder(tf.float32, [None, 784], name=\"x\")\n",
    "y = tf.placeholder(tf.float32, [None, 10], name=\"y\")\n",
    "\n",
    "W1 = tf.Variable(tf.random_normal([784, 300], stddev=0.03), name='W1')\n",
    "b1 = tf.Variable(tf.random_normal([300]), name='b1')\n",
    "W2 = tf.Variable(tf.random_normal([300, 10], stddev=0.03), name='W2')\n",
    "hidden_out = tf.add(tf.matmul(x, W1), b1)\n",
    "hidden_out = tf.nn.relu(hidden_out)\n",
    "y_ = tf.matmul(hidden_out, W2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "probs = tf.nn.softmax(y_)\n",
    "cross_entropy = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits_v2(logits=y_, labels=y))\n",
    "optimizer = tf.train.GradientDescentOptimizer(learning_rate=learning_rate).minimize(cross_entropy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "init_op = tf.global_variables_initializer()\n",
    "correct_prediction = tf.equal(tf.argmax(y, 1), tf.argmax(y_, 1))\n",
    "accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "sess = tf.Session()\n",
    "sess.run(init_op)\n",
    "total_batch = int(len(y_train) / batch_size)\n",
    "for epoch in range(epochs):\n",
    "    avg_cost = 0\n",
    "    for start, end in zip(range(0, len(y_train), batch_size), range(batch_size, len(y_train)+1, batch_size)):        \n",
    "        batch_x = x_train[start: end]\n",
    "        batch_y = y_train[start: end]\n",
    "        _, c = sess.run([optimizer, cross_entropy], feed_dict={x: batch_x, y: batch_y})\n",
    "        avg_cost += c / total_batch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def predict(inp):\n",
    "    return sess.run(probs, feed_dict={x:inp})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "inp = gradio.inputs.Sketchpad(flatten=True)\n",
    "io = gradio.Interface(inputs=inp, outputs=\"label\", model_type=\"pyfunc\", model=predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "No validation samples for this interface... skipping validation.\n",
      "NOTE: Gradio is in beta stage, please report all bugs to: a12d@stanford.edu\n",
      "Model is running locally at: http://localhost:7860/interface.html\n",
      "Model available publicly for 8 hours at: https://share.gradio.app/25d5d472\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "\n",
       "        <iframe\n",
       "            width=\"1000\"\n",
       "            height=\"500\"\n",
       "            src=\"http://localhost:7860/interface.html\"\n",
       "            frameborder=\"0\"\n",
       "            allowfullscreen\n",
       "        ></iframe>\n",
       "        "
      ],
      "text/plain": [
       "<IPython.lib.display.IFrame at 0x229f9dade80>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "(<gradio.networking.serve_files_in_background.<locals>.HTTPServer at 0x229f97553c8>,\n",
       " 'http://localhost:7860/',\n",
       " 'http://25d5d472.ngrok.io')"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "io.launch(share=True)"
   ]
  },
  {
   "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}