01. Intro To Deep Learning/02. Logistic Regression/02. Logistic Regression - MXNet.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "NtCzztnJHLIB"
   },
   "source": [
    "# Logistic Regression - MXNet (From Scratch)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 50
    },
    "colab_type": "code",
    "id": "XVQZ7D_1CkRr",
    "outputId": "3610e9ea-a37d-4f60-feab-8eadc3bdfef8"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "No of GPUs available on your system: 1\n",
      "Chosen device as your ctx: gpu(0)\n"
     ]
    }
   ],
   "source": [
    "import mxnet as mx\n",
    "from mxnet import nd, autograd, gluon\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "ctx = mx.gpu() if mx.context.num_gpus() > 0 else mx.cpu()\n",
    "print('No of GPUs available on your system:', mx.context.num_gpus())\n",
    "print('Chosen device as your ctx:', ctx)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 118
    },
    "colab_type": "code",
    "id": "S-5rDz25Cikv",
    "outputId": "87996158-3516-4d80-ee99-f207139a84a3"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train_x.shape: (10000, 2)\n",
      "train_y.shape: (10000,)\n",
      "test_x.shape: (1000, 2)\n",
      "test_y.shape: (1000,)\n",
      "train_y.max().asscalar(): 1.0\n",
      "train_y.min().asscalar(): 0.0\n"
     ]
    }
   ],
   "source": [
    "num_inputs = 2\n",
    "num_outputs = 1\n",
    "num_examples = 10000\n",
    "\n",
    "def function(x, num_examples = num_examples):\n",
    "    linear = 2 * x[:, 0] - 0.1 * x[:, 1] + 2.4\n",
    "    return nd.where(linear > 0, linear.ones_like(), linear.zeros_like())\n",
    "\n",
    "# loc is location(mean), and scale is standard deviation\n",
    "train_x = nd.random.normal(loc = 0, scale = 5, shape = (num_examples, num_inputs))\n",
    "train_y = function(train_x)\n",
    "\n",
    "test_x = nd.random.normal(loc = 0, scale = 5, shape = (1000, num_inputs))\n",
    "test_y= function(test_x)\n",
    "\n",
    "print('train_x.shape:', train_x.shape)\n",
    "print('train_y.shape:', train_y.shape)\n",
    "\n",
    "print('test_x.shape:', test_x.shape)\n",
    "print('test_y.shape:', test_y.shape)\n",
    "\n",
    "print('train_y.max().asscalar():', train_y.max().asscalar())\n",
    "print('train_y.min().asscalar():', train_y.min().asscalar())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "e8bHUIMae9nE"
   },
   "source": [
    "## Visualizing training dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 573
    },
    "colab_type": "code",
    "id": "D2eE-WJKc37I",
    "outputId": "45433814-c66f-4df6-e0ef-6c4aeee8b139"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter((train_x[:, 0] * (train_y == 0)).asnumpy(), (train_y * (train_y == 0)).asnumpy(), c = 'orange', label = 'Class 0')\n",
    "plt.scatter((train_x[:, 0] * (train_y == 1)).asnumpy(), (train_y * (train_y == 1)).asnumpy(), c = 'pink', label = 'Class 1')\n",
    "plt.title('Label vs Feature 1')\n",
    "plt.xlabel('Features 1')\n",
    "plt.ylabel('Labels')\n",
    "plt.legend()\n",
    "plt.show()\n",
    "\n",
    "plt.scatter((train_x[:, 1] * (train_y == 0)).asnumpy(), (train_y * (train_y == 0)).asnumpy(), c = 'orange', label = 'Class 0')\n",
    "plt.scatter((train_x[:, 1] * (train_y == 1)).asnumpy(), (train_y * (train_y == 1)).asnumpy(), c = 'pink', label = 'Class 1')\n",
    "plt.title('Label vs Feature 2')\n",
    "plt.xlabel('Features 2')\n",
    "plt.ylabel('Labels')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "2sc81uTYfvfz"
   },
   "source": [
    "## Visualizing testing dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 573
    },
    "colab_type": "code",
    "id": "ZZVGDoXCc38y",
    "outputId": "31faa9df-2d18-4358-a987-df0e10efea2a"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nWcahYGZmmf8CfSUUuqYZCCkAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter((test_x[:, 0] * (test_y == 0)).asnumpy(), (test_y * (test_y == 0)).asnumpy(), c = 'orange', label = 'Class 0')\n",
    "plt.scatter((test_x[:, 0] * (test_y == 1)).asnumpy(), (test_y * (test_y == 1)).asnumpy(), c = 'pink', label = 'Class 1')\n",
    "plt.title('Label vs Feature 1')\n",
    "plt.xlabel('Features 1')\n",
    "plt.ylabel('Labels')\n",
    "plt.legend()\n",
    "plt.show()\n",
    "\n",
    "plt.scatter((test_x[:, 1] * (test_y == 0)).asnumpy(), (test_y * (test_y == 0)).asnumpy(), c = 'orange', label = 'Class 0')\n",
    "plt.scatter((test_x[:, 1] * (test_y == 1)).asnumpy(), (test_y * (test_y == 1)).asnumpy(), c = 'pink', label = 'Class 1')\n",
    "plt.title('Label vs Feature 2')\n",
    "plt.xlabel('Features 2')\n",
    "plt.ylabel('Labels')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "9ZiJKr2OD-Rw"
   },
   "source": [
    "## Defining data iterator for training dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "GL9QkuoICinF",
    "outputId": "ffc9ace8-39ab-47c7-91a1-ccc1bc534a96"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "No of batches for batch size 16 is 625\n"
     ]
    }
   ],
   "source": [
    "batch_size = 16\n",
    "dataset = gluon.data.DataLoader(gluon.data.ArrayDataset(train_x, train_y), batch_size = batch_size, shuffle = True)\n",
    "\n",
    "print('No of batches for batch size {} is {}'.format(batch_size, len(dataset)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "TEpA8r5KEvjp"
   },
   "source": [
    "## Building logistic regression model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "IxOcYaxcLXGV"
   },
   "source": [
    "Below we are initializing our model parameters for our model, and attaching gradient to them to convert them into differentiable parameters.(Note that we are initializing our parameters for context ctx, which we have initialized above)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "WX4QIEyFba5o"
   },
   "source": [
    "## Logistic or Sigmoid function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "skP3j_6ac4Lb"
   },
   "outputs": [],
   "source": [
    "def sigmoid(linear):\n",
    "    return 1/(1 + nd.exp(-linear))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "2pz8NB2ec5Gr"
   },
   "source": [
    "## Visualizing sigmoid function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 295
    },
    "colab_type": "code",
    "id": "Gv7vaI-7c9cJ",
    "outputId": "06dd60a2-68d8-4976-e90f-37b56afef15d"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nlBRERKSVkoKIiLRSUhARkVb/H/SoRqTVNim6AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "x = nd.arange(-10, 10, 0.1)\n",
    "y = sigmoid(x)\n",
    "\n",
    "plt.plot(x.asnumpy(), y.asnumpy())\n",
    "plt.xlabel('x-axis')\n",
    "plt.ylabel('y-axis')\n",
    "plt.title('Sigmoid Function')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "-JRoqBaGCiq0"
   },
   "outputs": [],
   "source": [
    "weight = nd.random.normal(shape = (num_inputs, num_outputs), ctx = ctx) * 0.01\n",
    "bias = nd.random.normal(shape = num_outputs, ctx = ctx) * 0.01\n",
    "params = [weight, bias]\n",
    "\n",
    "for param in params:\n",
    "    param.attach_grad()\n",
    "\n",
    "def logisticregression(inputs):\n",
    "    return sigmoid(nd.dot(inputs, weight) + bias)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "HW4Jzb46EyyD"
   },
   "source": [
    "## Defining our objective function\n",
    "\n",
    "$\\text{Binary cross entropy} = -\\large\\frac{1}{n}\\sum_{i = 1}^{n}\\left(y_{i} * \\log{\\hat{y}_{i}} + (1 - y_{i}) * \\log\\left(1 - {\\hat{y}_{i}}\\right)\\right)$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "avO1uf5fItMV"
   },
   "outputs": [],
   "source": [
    "def binarycrossentropy(predictions, labels):\n",
    "    return -nd.mean((labels * nd.log(predictions)) +((1 - labels) * nd.log(1 - predictions)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "aKSx43yShlhM"
   },
   "source": [
    "## Defining our parameters optimizer(Gradient Decent)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "hvFKa4WuI9c-"
   },
   "outputs": [],
   "source": [
    "def sgd(params, lr):\n",
    "    for param in params:\n",
    "        param[:] = param - lr * param.grad"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "UlkEn5XxFPkd"
   },
   "source": [
    "## Training the  model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 353
    },
    "colab_type": "code",
    "id": "kVao6lbWCizF",
    "outputId": "4cdd1963-d851-494b-9b1c-3412609682b2"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 0 Loss: 0.61941044921875\n",
      "Epoch: 1 Loss: 0.490379345703125\n",
      "Epoch: 2 Loss: 0.420066943359375\n",
      "Epoch: 3 Loss: 0.3760067138671875\n",
      "Epoch: 4 Loss: 0.3455659912109375\n",
      "Epoch: 5 Loss: 0.3230663330078125\n",
      "Epoch: 6 Loss: 0.3056268798828125\n",
      "Epoch: 7 Loss: 0.2916218994140625\n",
      "Epoch: 8 Loss: 0.2800612548828125\n",
      "Epoch: 9 Loss: 0.2703139404296875\n",
      "Epoch: 10 Loss: 0.2619516357421875\n",
      "Epoch: 11 Loss: 0.2546765625\n",
      "Epoch: 12 Loss: 0.2482675048828125\n",
      "Epoch: 13 Loss: 0.242564794921875\n",
      "Epoch: 14 Loss: 0.237446142578125\n",
      "Epoch: 15 Loss: 0.232815478515625\n",
      "Epoch: 16 Loss: 0.2285983154296875\n",
      "Epoch: 17 Loss: 0.2247368408203125\n",
      "Epoch: 18 Loss: 0.22117978515625\n",
      "Epoch: 19 Loss: 0.217888671875\n"
     ]
    }
   ],
   "source": [
    "epochs = 20\n",
    "num_batches = len(dataset)\n",
    "learning_rate = .0001\n",
    "losses = []\n",
    "\n",
    "for epoch in range(epochs):\n",
    "    cumulative_loss = 0\n",
    "    for features, labels in dataset:\n",
    "        features = features.as_in_context(ctx) # .as_in_context put features stored in cpu to ctx(which we have initialized above)\n",
    "        labels = labels.as_in_context(ctx).reshape((-1, 1)) # .as_in_context put labels stored in cpu to ctx(which we have initialized above)\n",
    "        with autograd.record():\n",
    "            output = logisticregression(features)\n",
    "            loss = binarycrossentropy(output, labels)\n",
    "        loss.backward()\n",
    "        sgd(params, learning_rate)\n",
    "        cumulative_loss += loss\n",
    "\n",
    "    print('Epoch: {} Loss: {}'.format(epoch, cumulative_loss.asscalar()/num_batches))\n",
    "    losses.append(cumulative_loss.asscalar()/num_batches)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "qeUqn19hCi2h"
   },
   "outputs": [],
   "source": [
    "def plot(losses, model = logisticregression, sample_size=100):\n",
    "    xs = list(range(len(losses)))\n",
    "    plt.plot(xs, losses, '-r')\n",
    "    plt.title('Loss during training')\n",
    "    plt.xlabel('Epoch')\n",
    "    plt.ylabel('Loss')\n",
    "    plt.show()\n",
    "    \n",
    "    ####################################\n",
    "    # Visualizing training predictions #\n",
    "    ####################################\n",
    "    plt.plot(train_x[:sample_size, 0].asnumpy(),\n",
    "             model(train_x[:sample_size, :].as_in_context(ctx)).asnumpy(), 'or', label='Estimated')\n",
    "    plt.plot(train_x[:sample_size, 0].asnumpy(),\n",
    "             function(train_x[:sample_size, :]).asnumpy(), '*g', label='Real')\n",
    "    plt.legend()\n",
    "    plt.xlabel('First Features')\n",
    "    plt.ylabel('Labels')\n",
    "    plt.title('Prediction on training data as viewed from feature 1')\n",
    "    plt.show()\n",
    "    \n",
    "    plt.plot(train_x[:sample_size, 1].asnumpy(),\n",
    "             model(train_x[:sample_size, :].as_in_context(ctx)).asnumpy(), 'or', label='Estimated')\n",
    "    plt.plot(train_x[:sample_size, 1].asnumpy(),\n",
    "             function(train_x[:sample_size, :]).asnumpy(), '*g', label='Real')\n",
    "    plt.legend()\n",
    "    plt.xlabel('Second Features')\n",
    "    plt.ylabel('Labels')\n",
    "    plt.title('Prediction on training data as viewed from feature 2')\n",
    "    plt.show()\n",
    "    \n",
    "    ####################################\n",
    "    # Visualizing testing predictions #\n",
    "    ####################################\n",
    "    plt.plot(test_x[:sample_size, 0].asnumpy(),\n",
    "             model(test_x[:sample_size, :].as_in_context(ctx)).asnumpy(), 'or', label='Estimated')\n",
    "    plt.plot(test_x[:sample_size, 0].asnumpy(),\n",
    "             function(test_x[:sample_size, :]).asnumpy(), '*g', label='Real')\n",
    "    plt.legend()\n",
    "    plt.xlabel('First Features')\n",
    "    plt.ylabel('Labels')\n",
    "    plt.title('Prediction on testing data as viewed from feature 1')\n",
    "    plt.show()\n",
    "    \n",
    "    plt.plot(test_x[:sample_size, 1].asnumpy(),\n",
    "             model(test_x[:sample_size, :].as_in_context(ctx)).asnumpy(), 'or', label='Estimated')\n",
    "    plt.plot(test_x[:sample_size, 1].asnumpy(),\n",
    "             function(test_x[:sample_size, :]).asnumpy(), '*g', label='Real')\n",
    "    plt.legend()\n",
    "    plt.xlabel('Second Features')\n",
    "    plt.ylabel('Labels')\n",
    "    plt.title('Prediction on testing data as viewed from feature 2')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "id": "o3juufOWCi4a",
    "outputId": "119ac065-c0d5-4f7b-a20b-f2572d3ade5c"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAEWCAYAAACXGLsWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nymFhZmatcliYmVmr/j94iYcdWg3dhwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nT6y+dwISZ70AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYcAAAEWCAYAAACNJFuYAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n42AYhmHEYMbBMAzDiOH/A0rHz/h/JaKzAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nREQkQUFBREQS/j/+wC04ZuM1qgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot(losses)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "CIKATfHUBmt3"
   },
   "source": [
    "## Lets look out our weight and bias value after training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 134
    },
    "colab_type": "code",
    "id": "tpFOl4yVBrwj",
    "outputId": "c1b4e987-559f-46a7-9d48-a8044dfb8eb5"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Weight: \n",
      "[[ 0.61694986]\n",
      " [-0.02959069]]\n",
      "<NDArray 2x1 @gpu(0)>\n",
      "Bias: \n",
      "[0.10943359]\n",
      "<NDArray 1 @gpu(0)>\n"
     ]
    }
   ],
   "source": [
    "print(f'Weight: {weight}')\n",
    "print(f'Bias: {bias}')"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "Logistic Regression - MXNet (From Scratch).ipynb",
   "provenance": [],
   "version": "0.3.2"
  },
  "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": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}