tensorflow.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "be60cb5f-d3d4-41e0-889f-1104e89f04f5",
   "metadata": {},
   "source": [
    "# Tensorflow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "e76e13e0-ac3f-4060-a35a-ff82f9ce4b14",
   "metadata": {},
   "outputs": [],
   "source": [
    "# import\n",
    "import numpy as np\n",
    "import tensorflow as tf"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d6a3b292-b0ae-40b5-98b3-967cea367b18",
   "metadata": {},
   "source": [
    "0-d Tensors:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "047fdb8e-241d-44ac-81b8-c758c0d487a1",
   "metadata": {},
   "outputs": [],
   "source": [
    "sport = tf.constant('Tennis', tf.string)\n",
    "pi = tf.constant(3.141592, tf.float64)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "119360b0-7cca-4fd5-ae8e-e341d62f4b56",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "sport is a 0\n",
      "pi is a 0\n"
     ]
    }
   ],
   "source": [
    "print(f'sport is a {tf.rank(sport).numpy()}')\n",
    "print(f'pi is a {tf.rank(pi).numpy()}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "eda49fd5-0687-4dd8-bf4c-d759839980f2",
   "metadata": {},
   "source": [
    "1-d Tensors:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "05cbc192-db0d-45d9-9eca-d6007d28fc5d",
   "metadata": {},
   "outputs": [],
   "source": [
    "sports = tf.constant(['Tennis', 'Basketball'], tf.string)\n",
    "numbers = tf.constant([3.141592, 1.414213, 2.71821], tf.float64)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "79844d1d-7bef-413c-b585-451d1746a65c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "sport is a 1\n",
      "pi is a 1\n"
     ]
    }
   ],
   "source": [
    "print(f'sport is a {tf.rank(sports).numpy()}')\n",
    "print(f'pi is a {tf.rank(numbers).numpy()}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4bd3b177-7b68-4233-826e-a370bcaccb3a",
   "metadata": {},
   "source": [
    "## Computations on Tensors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "e0470d5f-096d-4e24-883b-0c63ff1ae3a0",
   "metadata": {},
   "outputs": [],
   "source": [
    "a = tf.constant(27)\n",
    "b = tf.constant(32)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "455c0201-f447-42aa-9730-83e70bc7d785",
   "metadata": {},
   "source": [
    "Add"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "31a2248a-e3f2-47d6-adbb-31eb30fdb1b6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tf.Tensor(59, shape=(), dtype=int32)\n"
     ]
    }
   ],
   "source": [
    "c1 = tf.add(a, b)\n",
    "print(c1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "5dcf1681-a275-4d18-b593-fe8c07108d3a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tf.Tensor(59, shape=(), dtype=int32)\n"
     ]
    }
   ],
   "source": [
    "c2 = a + b\n",
    "print(c2)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0fb1ee4a-29e1-4431-a914-d46e950143f2",
   "metadata": {},
   "source": [
    "Tensor computation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "48c1bad6-a978-49e8-9b59-578094e10d7a",
   "metadata": {},
   "outputs": [],
   "source": [
    "def func(a, b):\n",
    "    return (a + b) * (b - 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "70545a0a-956f-4ada-bb9b-1cd4255d467a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "6.0\n"
     ]
    }
   ],
   "source": [
    "a, b = 1.5, 2.5\n",
    "c1 = func(a, b)\n",
    "print(c1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a8bf23c4-d72f-498c-ad52-0d4d74302de2",
   "metadata": {},
   "source": [
    "## Neural networks in TensorFlow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "b422801f-c9c8-4521-a096-4739c94dc95e",
   "metadata": {},
   "outputs": [],
   "source": [
    "class Dense(tf.keras.layers.Layer):\n",
    "    def __init__(self, n_output_nodes):\n",
    "        super(Dense, self).__init__()\n",
    "        self.n_output_nodes = n_output_nodes\n",
    "        \n",
    "    def build(self, input_shape):\n",
    "        d = int(input_shape[-1])\n",
    "        # initialize parameters (weights and biases)\n",
    "        self.W = self.add_weight('weights', shape=[d, self.n_output_nodes])\n",
    "        self.b = self.add_weight('biases', shape=[1, self.n_output_nodes])\n",
    "        \n",
    "    def call(self, x):\n",
    "        z = tf.matmul(x, self.W) + b\n",
    "        \n",
    "        a = tf.sigmoid(z)\n",
    "        \n",
    "        return a"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "72f81b16-0d7c-42fa-a0ae-6d942abf4102",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0.14959235 0.28079373]]\n"
     ]
    }
   ],
   "source": [
    "layer = Dense(2)\n",
    "layer.build((1, 3))\n",
    "x_input = tf.constant([[1., 2., 3.]], shape=(1, 3))\n",
    "y = layer.call(x_input)\n",
    "\n",
    "print(y.numpy())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a0901d24-41d4-487c-98a8-cceddf8c3645",
   "metadata": {},
   "source": [
    "### Keras API"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "de2573db-13b7-4aa1-a973-6193e15323df",
   "metadata": {},
   "outputs": [],
   "source": [
    "# import\n",
    "from tensorflow.keras import Sequential\n",
    "from tensorflow.keras.layers import Dense"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aef93421-952a-499d-a783-251d4c8253e0",
   "metadata": {},
   "source": [
    "Sequential Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "aef1d668-c78e-4c81-9cfe-97905f0de4a3",
   "metadata": {},
   "outputs": [],
   "source": [
    "# build model\n",
    "model = Sequential()\n",
    "model.add(Dense(units=3, activation='sigmoid'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "7af781f1-d4b3-45d2-990a-653e686900fb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tf.Tensor: shape=(1, 3), dtype=float32, numpy=array([[0.4231848 , 0.20146888, 0.00948974]], dtype=float32)>"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# predict\n",
    "model.call(x_input)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b151f783-5e91-4dd7-b84d-ae1c89af10b1",
   "metadata": {},
   "source": [
    "Subclassing Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "d89b28ad-80c2-49fe-9552-4290aa2b0d7d",
   "metadata": {},
   "outputs": [],
   "source": [
    "class Model(tf.keras.Model):\n",
    "    def __init__(self, n_output_layers):\n",
    "        super(Model, self).__init__()\n",
    "        self.dense_layer = Dense(units=n_output_layers, activation='sigmoid')\n",
    "        \n",
    "    def call(self, inputs):\n",
    "        return self.dense_layer(inputs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "d80915e0-2d5f-44c5-abac-917a99c9352e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tf.Tensor: shape=(1, 3), dtype=float32, numpy=array([[0.4231848 , 0.20146888, 0.00948974]], dtype=float32)>"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# num of output layers\n",
    "n = 3\n",
    "model = Model(n)\n",
    "model.call(x_input)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "34cfcb39-976d-42c8-a6bc-0f3be552fb58",
   "metadata": {},
   "source": [
    "## Automatic differentiation in Tensorflow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "6261b5f3-44ed-4c4a-87db-108d1c4941a7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tf.Tensor(8.0, shape=(), dtype=float32)\n"
     ]
    }
   ],
   "source": [
    "# y = x^2\n",
    "x = tf.Variable(4.0)\n",
    "\n",
    "with tf.GradientTape() as tape:\n",
    "    y = x * x\n",
    "\n",
    "dy_dx = tape.gradient(y, x)\n",
    "print(dy_dx)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a04f3a24-9f28-47d3-b395-0e1b86621744",
   "metadata": {},
   "source": [
    "More about GradientTape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "249b6a95-67e5-4221-9b9a-3aa5f7a0f047",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Initializing x=[[-0.883817]]\n"
     ]
    }
   ],
   "source": [
    "x = tf.Variable([tf.random.normal([1])])\n",
    "print(\"Initializing x={}\".format(x.numpy()))\n",
    "\n",
    "# learning rate\n",
    "lr = 1e-2\n",
    "hist = []\n",
    "\n",
    "# target value\n",
    "x_f = 5\n",
    "\n",
    "# run SGD for a number of itr\n",
    "# compute the derivative of the loss with respect to x and perform the sgd update\n",
    "for i in range(500):\n",
    "    with tf.GradientTape() as tape:\n",
    "        loss = (x - x_f) * (x - x_f)\n",
    "        \n",
    "    grad = tape.gradient(loss, x)  # compute the derivative of the loss respect to x\n",
    "    new_x = x - lr * grad          # update \n",
    "    x.assign(new_x)\n",
    "    hist.append(x.numpy()[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "43e9f142-da53-4816-8d55-f20a12e7cb71",
   "metadata": {},
   "source": [
    "Visualize"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "14065a07-f5d5-4b66-a982-4f9719ee528a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'x value')"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.plot(hist)\n",
    "plt.plot([0, 500], [x_f, x_f])\n",
    "plt.legend(('Predicted', 'True'))\n",
    "plt.xlabel('Iteration')\n",
    "plt.ylabel('x value')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0cefc28f-dbe4-4151-919c-dbac9a86550f",
   "metadata": {},
   "source": [
    "Source: https://github.com/aamini/introtodeeplearning/"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "one4all",
   "language": "python",
   "name": "one4all"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}