{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:53:52.257836Z",
     "start_time": "2020-11-03T23:53:52.100727Z"
    }
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import IPython.display as ipd  # for display and clear_output\n",
    "import time  # for sleep"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Last time we discussed using pytorch to construct, train, and use neural networks as regression models.  \"Regression\" refers to the output of continuous values, like rainfall amounts or stock prices.\n",
    "\n",
    "Today we will modify the code in order to make classification models.  These are models that output discrete values, or categorical values, representing class labels for each sample, such as \"mask\" or \"no mask\" from images of people, and \"excited\" or \"calm\" from speech signals."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# From Regression to Classification\n",
    "\n",
    "The primary changes we need to consider is that we use a different output function for the network, a different loss function for classification, and our target outputs are now class labels."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Network Output\n",
    "\n",
    "For regression, we just output the weighted sum of inputs coming into the output layer as our prediction of the correct output for each sample. \n",
    "\n",
    "For classification, we want to convert these output values into class probabilities, or the probabilities that a given input sample should be classified as each of the possible classes.  So, if we have 2 classes, like \"cat\" or \"dog\", we need 2 outputs.  For these two outputs to be probabilities, we want each one to be between 0 and 1 and for a given sample we want the two values to sum to 1.\n",
    "\n",
    "We will accomplish this by passing the outputs of the neural network through a \"softmax\" function.  Let's call the two outputs of our network for input sample $n$, $p_{n,0}$ and $p_{n,1}$. where\n",
    "\n",
    "$$\\begin{align*}\n",
    "0 \\le p_{n,0} \\le 1\\\\\n",
    "0 \\le p_{n,1} \\le 1\\\\\n",
    "p_{n,0} + p_{n,1} = 1\n",
    "\\end{align*}$$\n",
    "\n",
    "The softmax function that converts our network outputs, $y_{n,0}$, and $y_{n, 1}$, to class probabilities uses each $y$ as an exponent of base $e$ and\n",
    "dividing each result by the sum of these exponentiated values:\n",
    "\n",
    "$$ \\begin{align*}\n",
    "p_{n,i} = \\frac{e^{y_{n,i}}}{\\sum_{j=0}^K e^{y_{n,j}}}\n",
    "\\end{align*}$$\n",
    "\n",
    "where $K$ is the number of classes.\n",
    "\n",
    "Let's do this in python."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:53:53.060023Z",
     "start_time": "2020-11-03T23:53:53.057602Z"
    }
   },
   "outputs": [],
   "source": [
    "def softmax(Y):\n",
    "    '''Y is n_samples X n_classes'''\n",
    "    expY = np.exp(Y)\n",
    "    P = expY / np.sum(expY, axis=1)\n",
    "    return P"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:53:53.456139Z",
     "start_time": "2020-11-03T23:53:53.405745Z"
    }
   },
   "outputs": [],
   "source": [
    "softmax?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:53:53.664452Z",
     "start_time": "2020-11-03T23:53:53.660663Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-2.3,  8.2]])"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Y = np.array([[-2.3, 8.2]])\n",
    "Y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:53:53.849206Z",
     "start_time": "2020-11-03T23:53:53.845946Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[2.75356911e-05, 9.99972464e-01]])"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "P = softmax(Y)\n",
    "P"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:53:54.443761Z",
     "start_time": "2020-11-03T23:53:54.440225Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(P)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Does this work for multiple samples, in rows of `Y`?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:53:58.075302Z",
     "start_time": "2020-11-03T23:53:58.072168Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-9.25841623,  4.20027187, -5.06507049],\n",
       "       [ 2.15564981, -7.69344863, -1.59212194],\n",
       "       [-5.78538943, -0.60967184,  4.57905291],\n",
       "       [ 4.54206011,  7.35250898, -0.59140415],\n",
       "       [ 3.51171099, -0.94854703, -9.98066263]])"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "n_samples = 5\n",
    "n_classes = 3\n",
    "Y = np.random.uniform(-10, 10, size=(n_samples, n_classes))\n",
    "Y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:53:59.016073Z",
     "start_time": "2020-11-03T23:53:58.979307Z"
    }
   },
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "operands could not be broadcast together with shapes (5,3) (5,) ",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-8-3461d9de5ee0>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mP\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msoftmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mY\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0mP\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m<ipython-input-2-c4371ee16ee1>\u001b[0m in \u001b[0;36msoftmax\u001b[0;34m(Y)\u001b[0m\n\u001b[1;32m      2\u001b[0m     \u001b[0;34m'''Y is n_samples X n_classes'''\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m     \u001b[0mexpY\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexp\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mY\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 4\u001b[0;31m     \u001b[0mP\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mexpY\u001b[0m \u001b[0;34m/\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msum\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mexpY\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      5\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0mP\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mValueError\u001b[0m: operands could not be broadcast together with shapes (5,3) (5,) "
     ]
    }
   ],
   "source": [
    "P = softmax(Y)\n",
    "P"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "How can we fix this?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:54:00.111697Z",
     "start_time": "2020-11-03T23:54:00.107593Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-10.12321484],\n",
       "       [ -7.12992075],\n",
       "       [ -1.81600836],\n",
       "       [ 11.30316494],\n",
       "       [ -7.41749867]])"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(Y, axis=1, keepdims=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:54:00.559892Z",
     "start_time": "2020-11-03T23:54:00.556869Z"
    }
   },
   "outputs": [],
   "source": [
    "def softmax(Y):\n",
    "    '''Y is n_samples X n_classes'''\n",
    "    expY = np.exp(Y)\n",
    "    P = expY / np.sum(expY, axis=1, keepdims=True)\n",
    "    return P"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:54:01.098925Z",
     "start_time": "2020-11-03T23:54:01.094907Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1.42864493e-06, 9.99903932e-01, 9.46392344e-05],\n",
       "       [9.76922165e-01, 5.15763805e-05, 2.30262585e-02],\n",
       "       [3.13581200e-05, 5.54798920e-03, 9.94420653e-01],\n",
       "       [5.67431529e-02, 9.42922284e-01, 3.34563274e-04],\n",
       "       [9.88571362e-01, 1.14272723e-02, 1.36566637e-06]])"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "P = softmax(Y)\n",
    "P"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:54:01.622936Z",
     "start_time": "2020-11-03T23:54:01.619058Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1., 1., 1., 1., 1.])"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "P.sum(axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Once we have class probabilities, how do we convert these to classes, or categories?\n",
    "\n",
    "We just created output  matrix `Y` with 5 samples and 3 classes representing, let's say, \"hot\", \"warm\", and \"cold\".  Now we want to know for each of the 5 samples, was that sample \"hot\", \"warm\", or \"cold\"?\n",
    "\n",
    "To do this, we just need to identify which of the three class probabilities was largest for each sample.  Hey, maybe `numpy` has a function for this?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:54:02.366776Z",
     "start_time": "2020-11-03T23:54:02.364097Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1, 0, 2, 1, 0])"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.argmax(P, axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If we have an `np.array` of class names, we can use these integers as indices into the class names."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:54:03.268786Z",
     "start_time": "2020-11-03T23:54:03.265887Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['warm', 'hot', 'cold', 'warm', 'hot'], dtype='<U4')"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "class_names = np.array(['hot', 'warm', 'cold'])\n",
    "class_names[np.argmax(P, axis=1)]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Whenever we are dividing, we have to watch out for divide-by-zero errors.  This could happen in our `softmax` function if all of the `Y` values are large negative values."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:54:04.458211Z",
     "start_time": "2020-11-03T23:54:04.454629Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-10000006.34200973, -10000008.96672827, -10000007.67589873],\n",
       "       [-10000009.81675825, -10000004.74712031, -10000008.14881062],\n",
       "       [-10000005.20780124, -10000005.33599189, -10000004.97343861],\n",
       "       [-10000001.5391905 , -10000007.74223229, -10000001.4729939 ],\n",
       "       [-10000006.17404579, -10000000.78783279, -10000006.22983135]])"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Y = np.random.uniform(-10000010, -10000000, size=(5, 3))\n",
    "Y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:54:04.972940Z",
     "start_time": "2020-11-03T23:54:04.968998Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0., 0., 0.],\n",
       "       [0., 0., 0.],\n",
       "       [0., 0., 0.],\n",
       "       [0., 0., 0.],\n",
       "       [0., 0., 0.]])"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.exp(Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:54:05.639013Z",
     "start_time": "2020-11-03T23:54:05.634479Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "<ipython-input-10-79349de2e72c>:4: RuntimeWarning: invalid value encountered in true_divide\n",
      "  P = expY / np.sum(expY, axis=1, keepdims=True)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[nan, nan, nan],\n",
       "       [nan, nan, nan],\n",
       "       [nan, nan, nan],\n",
       "       [nan, nan, nan],\n",
       "       [nan, nan, nan]])"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "softmax(Y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can deal with that by a simple division of the numerator and denominator by $e^{\\text{max}_j(y_{n, j})}$"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T18:58:27.551571Z",
     "start_time": "2020-11-03T18:58:27.548497Z"
    }
   },
   "source": [
    "$$ \\begin{align*}\n",
    "p_{n,i} = \\frac{e^{y_{n,i}}}{\\sum_{j=0}^K e^{y_{n,j}}} \\frac{e^{-\\text{max}_j(y_{n,j})}}{e^{-\\text{max}_j(y_{n,j})}}\n",
    "\\end{align*}$$\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:54:08.245032Z",
     "start_time": "2020-11-03T23:54:08.242605Z"
    }
   },
   "outputs": [],
   "source": [
    "def softmax(Y):\n",
    "    '''Y is n_samples X n_classes'''\n",
    "    maxY_by_row = np.max(Y, axis=1, keepdims=True)   \n",
    "    expY = np.exp(Y - maxY_by_row)\n",
    "    P = expY / np.sum(expY, axis=1, keepdims=True)\n",
    "    return P"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:54:08.741150Z",
     "start_time": "2020-11-03T23:54:08.736827Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[7.48552941e-01, 5.42402578e-02, 1.97206801e-01],\n",
       "       [6.04529191e-03, 9.61906944e-01, 3.20477636e-02],\n",
       "       [3.18087559e-01, 2.79817061e-01, 4.02095379e-01],\n",
       "       [4.82984454e-01, 9.77206464e-04, 5.16038339e-01],\n",
       "       [4.53884076e-03, 9.91168587e-01, 4.29257199e-03]])"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "softmax(Y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Loss Function"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For regression, we used the `torch.nn.MSELoss()` loss function, because we wanted to minimize the mean-squared-error between all training desired target values and the outputs produced by the neural network.\n",
    "\n",
    "For classification, we will instead want to maximize the data likelihood, which is the product of all of the correct class probabilities over all samples.  Remember that we are using gradient descent to optimize our loss functions.  Now, the gradient (or derivative) of a product of a bunch of things is a very computationally heavy calculation.  (Why?)  So, instead of optimizing this product of probabilities, we will optimize the log of that product, which converts it into a sum of logs of those probabilities.  We can do this because the weight values that optimize the product of probabilities  are the same weight values that optimize the log of that product!\n",
    "\n",
    "So we want to maximize the log-likelihood.  Recall that the `torch.optim` functions are designed to minimize a loss (hence the name \"loss\").  Since we want to maximize the log-likelihood, we must define the negative-log-likelihood to be used by `torch.optim`.\n",
    "\n",
    "To do this in our python code, we simply replace\n",
    "```python\n",
    "torch.nn.MSELoss()\n",
    "```\n",
    "with\n",
    "```python\n",
    "torch.nn.NLLLoss()\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Targets as Class Labels\n",
    "\n",
    "The final step to convert our code from regression to classification is to construct the correct target, `T`, matrix.  For regression, we would create for `T` an `n_samples` x `n_outputs` matrix of desired output values. For classification, we instead create a matrix `T` of `n_samples` values, regardless of how many outputs, or classes, we will have.  The values of `T` must be from the set $\\{0, 1, \\ldots, K-1\\}$ where $K$ is the number of different class labels we have.  We can get the number of different class labels using python like\n",
    "```python\n",
    "n_classes = len(np.unique(T))\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Classification Data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's start with a toy problem.  Say we have samples, each consisting of three integers.  Define the classes as \n",
    "* class 0: the first integer is greater than the sum of the other two,\n",
    "* class 1: the second integer is greater than the sum of the other two,\n",
    "* class 2: if not class 0 or 1."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Making different data to have one input, so can make plots in the animation below."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:55:26.051863Z",
     "start_time": "2020-11-03T23:55:26.045889Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([[0.74527494],\n",
       "        [5.0835637 ],\n",
       "        [6.78310398],\n",
       "        [4.61419519],\n",
       "        [3.72353747],\n",
       "        [6.03706642],\n",
       "        [6.08794542],\n",
       "        [1.65337227],\n",
       "        [8.52025903],\n",
       "        [4.14847294],\n",
       "        [9.12528426],\n",
       "        [4.60843185],\n",
       "        [6.85429187],\n",
       "        [9.72365158],\n",
       "        [4.75608714],\n",
       "        [7.50723516],\n",
       "        [2.38985605],\n",
       "        [8.96662919],\n",
       "        [1.78390658],\n",
       "        [3.67431871],\n",
       "        [1.25314593],\n",
       "        [5.39825349],\n",
       "        [7.88542309],\n",
       "        [2.38447133],\n",
       "        [9.33439215],\n",
       "        [0.4903726 ],\n",
       "        [1.10060789],\n",
       "        [2.96704847],\n",
       "        [0.6130978 ],\n",
       "        [1.06123726],\n",
       "        [3.038108  ],\n",
       "        [1.23809719],\n",
       "        [8.27166477],\n",
       "        [6.91655985],\n",
       "        [0.81580073],\n",
       "        [7.60606028],\n",
       "        [8.42448999],\n",
       "        [4.69311645],\n",
       "        [3.14885818],\n",
       "        [9.34106032],\n",
       "        [9.97937935],\n",
       "        [3.56643008],\n",
       "        [7.9738946 ],\n",
       "        [5.77928201],\n",
       "        [0.84270622],\n",
       "        [4.77074341],\n",
       "        [8.08889513],\n",
       "        [3.96722853],\n",
       "        [9.86619619],\n",
       "        [0.36044911],\n",
       "        [4.00387576],\n",
       "        [2.01615881],\n",
       "        [3.3891037 ],\n",
       "        [9.58775457],\n",
       "        [0.39485132],\n",
       "        [5.09102195],\n",
       "        [4.07048577],\n",
       "        [8.63362479],\n",
       "        [3.20364354],\n",
       "        [0.86357438],\n",
       "        [6.18334872],\n",
       "        [4.25643311],\n",
       "        [1.84496523],\n",
       "        [1.23901617],\n",
       "        [1.2242609 ],\n",
       "        [0.65864965],\n",
       "        [3.00952802],\n",
       "        [5.6775327 ],\n",
       "        [3.47742279],\n",
       "        [3.96001918],\n",
       "        [4.248112  ],\n",
       "        [0.66483946],\n",
       "        [3.79403982],\n",
       "        [3.77625988],\n",
       "        [0.20251507],\n",
       "        [9.22809818],\n",
       "        [1.67147073],\n",
       "        [5.82970945],\n",
       "        [3.74473851],\n",
       "        [3.03438061],\n",
       "        [4.01882699],\n",
       "        [5.91906059],\n",
       "        [2.22921631],\n",
       "        [5.87101102],\n",
       "        [5.83028119],\n",
       "        [4.8643471 ],\n",
       "        [4.49265557],\n",
       "        [4.92287404],\n",
       "        [8.82721534],\n",
       "        [1.11244688],\n",
       "        [7.4515501 ],\n",
       "        [1.30582537],\n",
       "        [0.25648592],\n",
       "        [1.79978752],\n",
       "        [2.13956547],\n",
       "        [2.12839789],\n",
       "        [7.7676748 ],\n",
       "        [7.32798922],\n",
       "        [7.22492687],\n",
       "        [4.1238338 ]]),\n",
       " array([[0],\n",
       "        [1],\n",
       "        [2],\n",
       "        [1],\n",
       "        [1],\n",
       "        [2],\n",
       "        [2],\n",
       "        [0],\n",
       "        [2],\n",
       "        [1],\n",
       "        [2],\n",
       "        [1],\n",
       "        [2],\n",
       "        [2],\n",
       "        [1],\n",
       "        [2],\n",
       "        [0],\n",
       "        [2],\n",
       "        [0],\n",
       "        [1],\n",
       "        [0],\n",
       "        [1],\n",
       "        [2],\n",
       "        [0],\n",
       "        [2],\n",
       "        [0],\n",
       "        [0],\n",
       "        [0],\n",
       "        [0],\n",
       "        [0],\n",
       "        [1],\n",
       "        [0],\n",
       "        [2],\n",
       "        [2],\n",
       "        [0],\n",
       "        [2],\n",
       "        [2],\n",
       "        [1],\n",
       "        [1],\n",
       "        [2],\n",
       "        [2],\n",
       "        [1],\n",
       "        [2],\n",
       "        [1],\n",
       "        [0],\n",
       "        [1],\n",
       "        [2],\n",
       "        [1],\n",
       "        [2],\n",
       "        [0],\n",
       "        [1],\n",
       "        [0],\n",
       "        [1],\n",
       "        [2],\n",
       "        [0],\n",
       "        [1],\n",
       "        [1],\n",
       "        [2],\n",
       "        [1],\n",
       "        [0],\n",
       "        [2],\n",
       "        [1],\n",
       "        [0],\n",
       "        [0],\n",
       "        [0],\n",
       "        [0],\n",
       "        [1],\n",
       "        [1],\n",
       "        [1],\n",
       "        [1],\n",
       "        [1],\n",
       "        [0],\n",
       "        [1],\n",
       "        [1],\n",
       "        [0],\n",
       "        [2],\n",
       "        [0],\n",
       "        [1],\n",
       "        [1],\n",
       "        [1],\n",
       "        [1],\n",
       "        [1],\n",
       "        [0],\n",
       "        [1],\n",
       "        [1],\n",
       "        [1],\n",
       "        [1],\n",
       "        [1],\n",
       "        [2],\n",
       "        [0],\n",
       "        [2],\n",
       "        [0],\n",
       "        [0],\n",
       "        [0],\n",
       "        [0],\n",
       "        [0],\n",
       "        [2],\n",
       "        [2],\n",
       "        [2],\n",
       "        [1]]))"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "n_samples = 100\n",
    "X = np.random.uniform(0, 10, size=(n_samples, 1))\n",
    "T = np.array([0 if (s[0] < 3) else \n",
    "              1 if (s[0] < 6) else \n",
    "              2 \n",
    "              for s in X]).reshape(-1, 1)\n",
    "Xtest = np.random.uniform(0, 10, size=(n_samples, 1))\n",
    "Ttest = np.array([0 if (s[0] < 3) else \n",
    "              1 if (s[0] < 6) else \n",
    "              2 \n",
    "              for s in Xtest]).reshape(-1, 1)\n",
    "\n",
    "\n",
    "X, T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:55:26.751529Z",
     "start_time": "2020-11-03T23:55:26.745676Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.74527494, 0.        ],\n",
       "       [5.0835637 , 1.        ],\n",
       "       [6.78310398, 2.        ],\n",
       "       [4.61419519, 1.        ],\n",
       "       [3.72353747, 1.        ],\n",
       "       [6.03706642, 2.        ],\n",
       "       [6.08794542, 2.        ],\n",
       "       [1.65337227, 0.        ],\n",
       "       [8.52025903, 2.        ],\n",
       "       [4.14847294, 1.        ],\n",
       "       [9.12528426, 2.        ],\n",
       "       [4.60843185, 1.        ],\n",
       "       [6.85429187, 2.        ],\n",
       "       [9.72365158, 2.        ],\n",
       "       [4.75608714, 1.        ],\n",
       "       [7.50723516, 2.        ],\n",
       "       [2.38985605, 0.        ],\n",
       "       [8.96662919, 2.        ],\n",
       "       [1.78390658, 0.        ],\n",
       "       [3.67431871, 1.        ],\n",
       "       [1.25314593, 0.        ],\n",
       "       [5.39825349, 1.        ],\n",
       "       [7.88542309, 2.        ],\n",
       "       [2.38447133, 0.        ],\n",
       "       [9.33439215, 2.        ],\n",
       "       [0.4903726 , 0.        ],\n",
       "       [1.10060789, 0.        ],\n",
       "       [2.96704847, 0.        ],\n",
       "       [0.6130978 , 0.        ],\n",
       "       [1.06123726, 0.        ],\n",
       "       [3.038108  , 1.        ],\n",
       "       [1.23809719, 0.        ],\n",
       "       [8.27166477, 2.        ],\n",
       "       [6.91655985, 2.        ],\n",
       "       [0.81580073, 0.        ],\n",
       "       [7.60606028, 2.        ],\n",
       "       [8.42448999, 2.        ],\n",
       "       [4.69311645, 1.        ],\n",
       "       [3.14885818, 1.        ],\n",
       "       [9.34106032, 2.        ],\n",
       "       [9.97937935, 2.        ],\n",
       "       [3.56643008, 1.        ],\n",
       "       [7.9738946 , 2.        ],\n",
       "       [5.77928201, 1.        ],\n",
       "       [0.84270622, 0.        ],\n",
       "       [4.77074341, 1.        ],\n",
       "       [8.08889513, 2.        ],\n",
       "       [3.96722853, 1.        ],\n",
       "       [9.86619619, 2.        ],\n",
       "       [0.36044911, 0.        ],\n",
       "       [4.00387576, 1.        ],\n",
       "       [2.01615881, 0.        ],\n",
       "       [3.3891037 , 1.        ],\n",
       "       [9.58775457, 2.        ],\n",
       "       [0.39485132, 0.        ],\n",
       "       [5.09102195, 1.        ],\n",
       "       [4.07048577, 1.        ],\n",
       "       [8.63362479, 2.        ],\n",
       "       [3.20364354, 1.        ],\n",
       "       [0.86357438, 0.        ],\n",
       "       [6.18334872, 2.        ],\n",
       "       [4.25643311, 1.        ],\n",
       "       [1.84496523, 0.        ],\n",
       "       [1.23901617, 0.        ],\n",
       "       [1.2242609 , 0.        ],\n",
       "       [0.65864965, 0.        ],\n",
       "       [3.00952802, 1.        ],\n",
       "       [5.6775327 , 1.        ],\n",
       "       [3.47742279, 1.        ],\n",
       "       [3.96001918, 1.        ],\n",
       "       [4.248112  , 1.        ],\n",
       "       [0.66483946, 0.        ],\n",
       "       [3.79403982, 1.        ],\n",
       "       [3.77625988, 1.        ],\n",
       "       [0.20251507, 0.        ],\n",
       "       [9.22809818, 2.        ],\n",
       "       [1.67147073, 0.        ],\n",
       "       [5.82970945, 1.        ],\n",
       "       [3.74473851, 1.        ],\n",
       "       [3.03438061, 1.        ],\n",
       "       [4.01882699, 1.        ],\n",
       "       [5.91906059, 1.        ],\n",
       "       [2.22921631, 0.        ],\n",
       "       [5.87101102, 1.        ],\n",
       "       [5.83028119, 1.        ],\n",
       "       [4.8643471 , 1.        ],\n",
       "       [4.49265557, 1.        ],\n",
       "       [4.92287404, 1.        ],\n",
       "       [8.82721534, 2.        ],\n",
       "       [1.11244688, 0.        ],\n",
       "       [7.4515501 , 2.        ],\n",
       "       [1.30582537, 0.        ],\n",
       "       [0.25648592, 0.        ],\n",
       "       [1.79978752, 0.        ],\n",
       "       [2.13956547, 0.        ],\n",
       "       [2.12839789, 0.        ],\n",
       "       [7.7676748 , 2.        ],\n",
       "       [7.32798922, 2.        ],\n",
       "       [7.22492687, 2.        ],\n",
       "       [4.1238338 , 1.        ]])"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.hstack((X, T))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:55:27.384083Z",
     "start_time": "2020-11-03T23:55:27.381349Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(32, 39, 29)"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(T == 0), np.sum(T == 1), np.sum(T == 2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Create, Train, and Use our Classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:55:28.204900Z",
     "start_time": "2020-11-03T23:55:28.201304Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'1.4.0'"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import torch\n",
    "torch.__version__"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Remember that `pytorch` requires inputs of single precision float.  The `NLLLoss` function requires target class labels to be one-dimensional and double ints."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:55:29.333939Z",
     "start_time": "2020-11-03T23:55:29.329653Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[0.7453],\n",
       "        [5.0836],\n",
       "        [6.7831],\n",
       "        [4.6142],\n",
       "        [3.7235],\n",
       "        [6.0371],\n",
       "        [6.0879],\n",
       "        [1.6534],\n",
       "        [8.5203],\n",
       "        [4.1485],\n",
       "        [9.1253],\n",
       "        [4.6084],\n",
       "        [6.8543],\n",
       "        [9.7237],\n",
       "        [4.7561],\n",
       "        [7.5072],\n",
       "        [2.3899],\n",
       "        [8.9666],\n",
       "        [1.7839],\n",
       "        [3.6743],\n",
       "        [1.2531],\n",
       "        [5.3983],\n",
       "        [7.8854],\n",
       "        [2.3845],\n",
       "        [9.3344],\n",
       "        [0.4904],\n",
       "        [1.1006],\n",
       "        [2.9670],\n",
       "        [0.6131],\n",
       "        [1.0612],\n",
       "        [3.0381],\n",
       "        [1.2381],\n",
       "        [8.2717],\n",
       "        [6.9166],\n",
       "        [0.8158],\n",
       "        [7.6061],\n",
       "        [8.4245],\n",
       "        [4.6931],\n",
       "        [3.1489],\n",
       "        [9.3411],\n",
       "        [9.9794],\n",
       "        [3.5664],\n",
       "        [7.9739],\n",
       "        [5.7793],\n",
       "        [0.8427],\n",
       "        [4.7707],\n",
       "        [8.0889],\n",
       "        [3.9672],\n",
       "        [9.8662],\n",
       "        [0.3604],\n",
       "        [4.0039],\n",
       "        [2.0162],\n",
       "        [3.3891],\n",
       "        [9.5878],\n",
       "        [0.3949],\n",
       "        [5.0910],\n",
       "        [4.0705],\n",
       "        [8.6336],\n",
       "        [3.2036],\n",
       "        [0.8636],\n",
       "        [6.1833],\n",
       "        [4.2564],\n",
       "        [1.8450],\n",
       "        [1.2390],\n",
       "        [1.2243],\n",
       "        [0.6586],\n",
       "        [3.0095],\n",
       "        [5.6775],\n",
       "        [3.4774],\n",
       "        [3.9600],\n",
       "        [4.2481],\n",
       "        [0.6648],\n",
       "        [3.7940],\n",
       "        [3.7763],\n",
       "        [0.2025],\n",
       "        [9.2281],\n",
       "        [1.6715],\n",
       "        [5.8297],\n",
       "        [3.7447],\n",
       "        [3.0344],\n",
       "        [4.0188],\n",
       "        [5.9191],\n",
       "        [2.2292],\n",
       "        [5.8710],\n",
       "        [5.8303],\n",
       "        [4.8643],\n",
       "        [4.4927],\n",
       "        [4.9229],\n",
       "        [8.8272],\n",
       "        [1.1124],\n",
       "        [7.4516],\n",
       "        [1.3058],\n",
       "        [0.2565],\n",
       "        [1.7998],\n",
       "        [2.1396],\n",
       "        [2.1284],\n",
       "        [7.7677],\n",
       "        [7.3280],\n",
       "        [7.2249],\n",
       "        [4.1238]])"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "torch.from_numpy(X).float()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:55:30.204671Z",
     "start_time": "2020-11-03T23:55:30.201052Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([0, 1, 2, 1, 1, 2, 2, 0, 2, 1, 2, 1, 2, 2, 1, 2, 0, 2, 0, 1, 0, 1, 2, 0,\n",
       "        2, 0, 0, 0, 0, 0, 1, 0, 2, 2, 0, 2, 2, 1, 1, 2, 2, 1, 2, 1, 0, 1, 2, 1,\n",
       "        2, 0, 1, 0, 1, 2, 0, 1, 1, 2, 1, 0, 2, 1, 0, 0, 0, 0, 1, 1, 1, 1, 1, 0,\n",
       "        1, 1, 0, 2, 0, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 2, 0, 2, 0, 0, 0, 0, 0,\n",
       "        2, 2, 2, 1])"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "torch.from_numpy(T).reshape(-1).long()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:55:31.931782Z",
     "start_time": "2020-11-03T23:55:31.929316Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 2])"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.unique(T)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-11-03T23:55:47.136670Z",
     "start_time": "2020-11-03T23:55:34.997239Z"
    },
    "lines_to_next_cell": 0
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x864 with 5 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "n_inputs = X.shape[1]\n",
    "n_classes = len(np.unique(T))\n",
    "\n",
    "Xt = torch.from_numpy(X).float()\n",
    "Tt = torch.from_numpy(T).reshape(-1).long()\n",
    "\n",
    "Xtestt = torch.from_numpy(Xtest).float()\n",
    "Ttestt = torch.from_numpy(Ttest).reshape(-1).long()\n",
    "\n",
    "nnet = torch.nn.Sequential(torch.nn.Linear(n_inputs, 10), torch.nn.Tanh(), \n",
    "                           torch.nn.Linear(10, 20), torch.nn.Tanh(), \n",
    "                           torch.nn.Linear(20, 10), torch.nn.Tanh(), \n",
    "                           torch.nn.Linear(10, n_classes), torch.nn.LogSoftmax(dim=1))\n",
    "\n",
    "learning_rate = 0.01\n",
    "n_epochs = 10000\n",
    "\n",
    "optimizer = torch.optim.SGD(nnet.parameters(), lr=learning_rate)\n",
    "nll_f = torch.nn.NLLLoss()\n",
    "\n",
    "likelihood_trace = []\n",
    "likelihood_test_trace = []\n",
    "\n",
    "fig = plt.figure(figsize=(10, 12))\n",
    "\n",
    "def forward_all_layers(X):\n",
    "    Ys = [X]\n",
    "    for layer in nnet:\n",
    "        Ys.append(layer(Ys[-1]))\n",
    "    return Ys[1:]\n",
    "        \n",
    "for epoch in range(n_epochs):\n",
    "\n",
    "    logP = nnet(Xt)\n",
    "\n",
    "    nll = nll_f(logP, Tt)\n",
    "    # mse = mse_f(Y, Tt)\n",
    "    \n",
    "    optimizer.zero_grad()\n",
    "    nll.backward()\n",
    "    optimizer.step()\n",
    "    \n",
    "    # error traces for plotting\n",
    "    likelihood_trace.append((-nll).exp())\n",
    "    \n",
    "    logPtest = nnet(Xtestt)\n",
    "    likelihood_test_trace.append((-nll_f(logPtest, Ttestt)).exp())\n",
    "\n",
    "    if epoch % 1000 == 0 or epoch == n_epochs-1:\n",
    "        plt.clf()\n",
    "        \n",
    "        n_hidden_layers = (len(nnet) - 1) //2\n",
    "        nplots = 2 + n_hidden_layers\n",
    "\n",
    "        plt.subplot(nplots, 1, 1)\n",
    "        plt.plot(likelihood_trace[:epoch])\n",
    "        plt.plot(likelihood_test_trace[:epoch])\n",
    "        # plt.ylim(0, 0.7)\n",
    "        plt.xlabel('Epochs')\n",
    "        plt.ylabel('Likelihood')\n",
    "        plt.legend(('Train','Test'), loc='upper left')\n",
    "        \n",
    "        plt.subplot(nplots, 1, 2)\n",
    "        classes = logPtest.argmax(axis=1)\n",
    "        order = np.argsort(X, axis=0).reshape(-1)\n",
    "        plt.plot(X[order,:], T[order,:], 'o-', label='Train')\n",
    "        order = np.argsort(Xtest, axis=0).reshape(-1)\n",
    "        plt.plot(Xtest[order, :], Ttest[order, :], 'o-', label='Test')\n",
    "        plt.plot(Xtest[order, :], classes[order], 'o-')\n",
    "        plt.legend(('Training','Testing','Model'), loc='upper left')\n",
    "        plt.xlabel('$x$')\n",
    "        plt.ylabel('Actual and Predicted Class')\n",
    "        \n",
    "        Ys = forward_all_layers(Xt)\n",
    "        Z = Ys[:-1]\n",
    "        ploti = 2\n",
    "        for layeri in range(n_hidden_layers, 0, -1):\n",
    "            ploti += 1\n",
    "            plt.subplot(nplots, 1, ploti)\n",
    "            order = np.argsort(X, axis=0).reshape(-1)\n",
    "            plt.plot(X[order,:], Z[layeri * 2 - 1][order,:].detach())\n",
    "            plt.xlabel('$x$')\n",
    "            plt.ylabel(f'Hidden Layer {layeri}');\n",
    "        \n",
    "        ipd.clear_output(wait=True)\n",
    "        ipd.display(fig)\n",
    "ipd.clear_output(wait=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "jupytext": {
   "formats": "ipynb,py:light"
  },
  "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": "3.8.3"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}