MNIST Classification and the Tutorial for CNN.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's try to implement a convolutional neural netowrk for digits classification\n",
    "Some borrowed from: \n",
    "\n",
    "https://github.com/yunjey/pytorch-tutorial/blob/master/tutorials/02-intermediate/convolutional_neural_network/main.py\n",
    "Ke Wang 2019.4.9"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib notebook\n",
    "import torch \n",
    "import torch.nn as nn\n",
    "import torchvision\n",
    "import torchvision.transforms as transforms\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import sigpy.plot as pl"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Device Configuration"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [],
   "source": [
    "device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Hyper parameters\n",
    "num_epochs = 50\n",
    "num_classes = 10\n",
    "batch_size = 100\n",
    "learning_rate = 0.001"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading http://yann.lecun.com/exdb/mnist/train-images-idx3-ubyte.gz\n",
      "Downloading http://yann.lecun.com/exdb/mnist/train-labels-idx1-ubyte.gz\n",
      "Downloading http://yann.lecun.com/exdb/mnist/t10k-images-idx3-ubyte.gz\n",
      "Downloading http://yann.lecun.com/exdb/mnist/t10k-labels-idx1-ubyte.gz\n",
      "Processing...\n",
      "Done!\n"
     ]
    }
   ],
   "source": [
    "#MNIST dataset\n",
    "train_dataset = torchvision.datasets.MNIST(root='data/',\n",
    "                                           train=True, \n",
    "                                           transform=transforms.ToTensor(),\n",
    "                                           download=True)\n",
    "test_dataset = torchvision.datasets.MNIST(root='data/',\n",
    "                                          train=False, \n",
    "                                          transform=transforms.ToTensor())\n",
    "\n",
    "# Data loader\n",
    "train_loader = torch.utils.data.DataLoader(dataset=train_dataset,\n",
    "                                           batch_size=batch_size, \n",
    "                                           shuffle=True)\n",
    "\n",
    "test_loader = torch.utils.data.DataLoader(dataset=test_dataset,\n",
    "                                          batch_size=batch_size, \n",
    "                                          shuffle=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "/* Put everything inside the global mpl namespace */\n",
       "window.mpl = {};\n",
       "\n",
       "\n",
       "mpl.get_websocket_type = function() {\n",
       "    if (typeof(WebSocket) !== 'undefined') {\n",
       "        return WebSocket;\n",
       "    } else if (typeof(MozWebSocket) !== 'undefined') {\n",
       "        return MozWebSocket;\n",
       "    } else {\n",
       "        alert('Your browser does not have WebSocket support.' +\n",
       "              'Please try Chrome, Safari or Firefox ≥ 6. ' +\n",
       "              'Firefox 4 and 5 are also supported but you ' +\n",
       "              'have to enable WebSockets in about:config.');\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure = function(figure_id, websocket, ondownload, parent_element) {\n",
       "    this.id = figure_id;\n",
       "\n",
       "    this.ws = websocket;\n",
       "\n",
       "    this.supports_binary = (this.ws.binaryType != undefined);\n",
       "\n",
       "    if (!this.supports_binary) {\n",
       "        var warnings = document.getElementById(\"mpl-warnings\");\n",
       "        if (warnings) {\n",
       "            warnings.style.display = 'block';\n",
       "            warnings.textContent = (\n",
       "                \"This browser does not support binary websocket messages. \" +\n",
       "                    \"Performance may be slow.\");\n",
       "        }\n",
       "    }\n",
       "\n",
       "    this.imageObj = new Image();\n",
       "\n",
       "    this.context = undefined;\n",
       "    this.message = undefined;\n",
       "    this.canvas = undefined;\n",
       "    this.rubberband_canvas = undefined;\n",
       "    this.rubberband_context = undefined;\n",
       "    this.format_dropdown = undefined;\n",
       "\n",
       "    this.image_mode = 'full';\n",
       "\n",
       "    this.root = $('<div/>');\n",
       "    this._root_extra_style(this.root)\n",
       "    this.root.attr('style', 'display: inline-block');\n",
       "\n",
       "    $(parent_element).append(this.root);\n",
       "\n",
       "    this._init_header(this);\n",
       "    this._init_canvas(this);\n",
       "    this._init_toolbar(this);\n",
       "\n",
       "    var fig = this;\n",
       "\n",
       "    this.waiting = false;\n",
       "\n",
       "    this.ws.onopen =  function () {\n",
       "            fig.send_message(\"supports_binary\", {value: fig.supports_binary});\n",
       "            fig.send_message(\"send_image_mode\", {});\n",
       "            if (mpl.ratio != 1) {\n",
       "                fig.send_message(\"set_dpi_ratio\", {'dpi_ratio': mpl.ratio});\n",
       "            }\n",
       "            fig.send_message(\"refresh\", {});\n",
       "        }\n",
       "\n",
       "    this.imageObj.onload = function() {\n",
       "            if (fig.image_mode == 'full') {\n",
       "                // Full images could contain transparency (where diff images\n",
       "                // almost always do), so we need to clear the canvas so that\n",
       "                // there is no ghosting.\n",
       "                fig.context.clearRect(0, 0, fig.canvas.width, fig.canvas.height);\n",
       "            }\n",
       "            fig.context.drawImage(fig.imageObj, 0, 0);\n",
       "        };\n",
       "\n",
       "    this.imageObj.onunload = function() {\n",
       "        fig.ws.close();\n",
       "    }\n",
       "\n",
       "    this.ws.onmessage = this._make_on_message_function(this);\n",
       "\n",
       "    this.ondownload = ondownload;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_header = function() {\n",
       "    var titlebar = $(\n",
       "        '<div class=\"ui-dialog-titlebar ui-widget-header ui-corner-all ' +\n",
       "        'ui-helper-clearfix\"/>');\n",
       "    var titletext = $(\n",
       "        '<div class=\"ui-dialog-title\" style=\"width: 100%; ' +\n",
       "        'text-align: center; padding: 3px;\"/>');\n",
       "    titlebar.append(titletext)\n",
       "    this.root.append(titlebar);\n",
       "    this.header = titletext[0];\n",
       "}\n",
       "\n",
       "\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_canvas = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var canvas_div = $('<div/>');\n",
       "\n",
       "    canvas_div.attr('style', 'position: relative; clear: both; outline: 0');\n",
       "\n",
       "    function canvas_keyboard_event(event) {\n",
       "        return fig.key_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    canvas_div.keydown('key_press', canvas_keyboard_event);\n",
       "    canvas_div.keyup('key_release', canvas_keyboard_event);\n",
       "    this.canvas_div = canvas_div\n",
       "    this._canvas_extra_style(canvas_div)\n",
       "    this.root.append(canvas_div);\n",
       "\n",
       "    var canvas = $('<canvas/>');\n",
       "    canvas.addClass('mpl-canvas');\n",
       "    canvas.attr('style', \"left: 0; top: 0; z-index: 0; outline: 0\")\n",
       "\n",
       "    this.canvas = canvas[0];\n",
       "    this.context = canvas[0].getContext(\"2d\");\n",
       "\n",
       "    var backingStore = this.context.backingStorePixelRatio ||\n",
       "\tthis.context.webkitBackingStorePixelRatio ||\n",
       "\tthis.context.mozBackingStorePixelRatio ||\n",
       "\tthis.context.msBackingStorePixelRatio ||\n",
       "\tthis.context.oBackingStorePixelRatio ||\n",
       "\tthis.context.backingStorePixelRatio || 1;\n",
       "\n",
       "    mpl.ratio = (window.devicePixelRatio || 1) / backingStore;\n",
       "\n",
       "    var rubberband = $('<canvas/>');\n",
       "    rubberband.attr('style', \"position: absolute; left: 0; top: 0; z-index: 1;\")\n",
       "\n",
       "    var pass_mouse_events = true;\n",
       "\n",
       "    canvas_div.resizable({\n",
       "        start: function(event, ui) {\n",
       "            pass_mouse_events = false;\n",
       "        },\n",
       "        resize: function(event, ui) {\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "        stop: function(event, ui) {\n",
       "            pass_mouse_events = true;\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "    });\n",
       "\n",
       "    function mouse_event_fn(event) {\n",
       "        if (pass_mouse_events)\n",
       "            return fig.mouse_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    rubberband.mousedown('button_press', mouse_event_fn);\n",
       "    rubberband.mouseup('button_release', mouse_event_fn);\n",
       "    // Throttle sequential mouse events to 1 every 20ms.\n",
       "    rubberband.mousemove('motion_notify', mouse_event_fn);\n",
       "\n",
       "    rubberband.mouseenter('figure_enter', mouse_event_fn);\n",
       "    rubberband.mouseleave('figure_leave', mouse_event_fn);\n",
       "\n",
       "    canvas_div.on(\"wheel\", function (event) {\n",
       "        event = event.originalEvent;\n",
       "        event['data'] = 'scroll'\n",
       "        if (event.deltaY < 0) {\n",
       "            event.step = 1;\n",
       "        } else {\n",
       "            event.step = -1;\n",
       "        }\n",
       "        mouse_event_fn(event);\n",
       "    });\n",
       "\n",
       "    canvas_div.append(canvas);\n",
       "    canvas_div.append(rubberband);\n",
       "\n",
       "    this.rubberband = rubberband;\n",
       "    this.rubberband_canvas = rubberband[0];\n",
       "    this.rubberband_context = rubberband[0].getContext(\"2d\");\n",
       "    this.rubberband_context.strokeStyle = \"#000000\";\n",
       "\n",
       "    this._resize_canvas = function(width, height) {\n",
       "        // Keep the size of the canvas, canvas container, and rubber band\n",
       "        // canvas in synch.\n",
       "        canvas_div.css('width', width)\n",
       "        canvas_div.css('height', height)\n",
       "\n",
       "        canvas.attr('width', width * mpl.ratio);\n",
       "        canvas.attr('height', height * mpl.ratio);\n",
       "        canvas.attr('style', 'width: ' + width + 'px; height: ' + height + 'px;');\n",
       "\n",
       "        rubberband.attr('width', width);\n",
       "        rubberband.attr('height', height);\n",
       "    }\n",
       "\n",
       "    // Set the figure to an initial 600x600px, this will subsequently be updated\n",
       "    // upon first draw.\n",
       "    this._resize_canvas(600, 600);\n",
       "\n",
       "    // Disable right mouse context menu.\n",
       "    $(this.rubberband_canvas).bind(\"contextmenu\",function(e){\n",
       "        return false;\n",
       "    });\n",
       "\n",
       "    function set_focus () {\n",
       "        canvas.focus();\n",
       "        canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    window.setTimeout(set_focus, 100);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>')\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items) {\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) {\n",
       "            // put a spacer in here.\n",
       "            continue;\n",
       "        }\n",
       "        var button = $('<button/>');\n",
       "        button.addClass('ui-button ui-widget ui-state-default ui-corner-all ' +\n",
       "                        'ui-button-icon-only');\n",
       "        button.attr('role', 'button');\n",
       "        button.attr('aria-disabled', 'false');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "\n",
       "        var icon_img = $('<span/>');\n",
       "        icon_img.addClass('ui-button-icon-primary ui-icon');\n",
       "        icon_img.addClass(image);\n",
       "        icon_img.addClass('ui-corner-all');\n",
       "\n",
       "        var tooltip_span = $('<span/>');\n",
       "        tooltip_span.addClass('ui-button-text');\n",
       "        tooltip_span.html(tooltip);\n",
       "\n",
       "        button.append(icon_img);\n",
       "        button.append(tooltip_span);\n",
       "\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    var fmt_picker_span = $('<span/>');\n",
       "\n",
       "    var fmt_picker = $('<select/>');\n",
       "    fmt_picker.addClass('mpl-toolbar-option ui-widget ui-widget-content');\n",
       "    fmt_picker_span.append(fmt_picker);\n",
       "    nav_element.append(fmt_picker_span);\n",
       "    this.format_dropdown = fmt_picker[0];\n",
       "\n",
       "    for (var ind in mpl.extensions) {\n",
       "        var fmt = mpl.extensions[ind];\n",
       "        var option = $(\n",
       "            '<option/>', {selected: fmt === mpl.default_extension}).html(fmt);\n",
       "        fmt_picker.append(option)\n",
       "    }\n",
       "\n",
       "    // Add hover states to the ui-buttons\n",
       "    $( \".ui-button\" ).hover(\n",
       "        function() { $(this).addClass(\"ui-state-hover\");},\n",
       "        function() { $(this).removeClass(\"ui-state-hover\");}\n",
       "    );\n",
       "\n",
       "    var status_bar = $('<span class=\"mpl-message\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.request_resize = function(x_pixels, y_pixels) {\n",
       "    // Request matplotlib to resize the figure. Matplotlib will then trigger a resize in the client,\n",
       "    // which will in turn request a refresh of the image.\n",
       "    this.send_message('resize', {'width': x_pixels, 'height': y_pixels});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_message = function(type, properties) {\n",
       "    properties['type'] = type;\n",
       "    properties['figure_id'] = this.id;\n",
       "    this.ws.send(JSON.stringify(properties));\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_draw_message = function() {\n",
       "    if (!this.waiting) {\n",
       "        this.waiting = true;\n",
       "        this.ws.send(JSON.stringify({type: \"draw\", figure_id: this.id}));\n",
       "    }\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    var format_dropdown = fig.format_dropdown;\n",
       "    var format = format_dropdown.options[format_dropdown.selectedIndex].value;\n",
       "    fig.ondownload(fig, format);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_resize = function(fig, msg) {\n",
       "    var size = msg['size'];\n",
       "    if (size[0] != fig.canvas.width || size[1] != fig.canvas.height) {\n",
       "        fig._resize_canvas(size[0], size[1]);\n",
       "        fig.send_message(\"refresh\", {});\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_rubberband = function(fig, msg) {\n",
       "    var x0 = msg['x0'] / mpl.ratio;\n",
       "    var y0 = (fig.canvas.height - msg['y0']) / mpl.ratio;\n",
       "    var x1 = msg['x1'] / mpl.ratio;\n",
       "    var y1 = (fig.canvas.height - msg['y1']) / mpl.ratio;\n",
       "    x0 = Math.floor(x0) + 0.5;\n",
       "    y0 = Math.floor(y0) + 0.5;\n",
       "    x1 = Math.floor(x1) + 0.5;\n",
       "    y1 = Math.floor(y1) + 0.5;\n",
       "    var min_x = Math.min(x0, x1);\n",
       "    var min_y = Math.min(y0, y1);\n",
       "    var width = Math.abs(x1 - x0);\n",
       "    var height = Math.abs(y1 - y0);\n",
       "\n",
       "    fig.rubberband_context.clearRect(\n",
       "        0, 0, fig.canvas.width, fig.canvas.height);\n",
       "\n",
       "    fig.rubberband_context.strokeRect(min_x, min_y, width, height);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_figure_label = function(fig, msg) {\n",
       "    // Updates the figure title.\n",
       "    fig.header.textContent = msg['label'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_cursor = function(fig, msg) {\n",
       "    var cursor = msg['cursor'];\n",
       "    switch(cursor)\n",
       "    {\n",
       "    case 0:\n",
       "        cursor = 'pointer';\n",
       "        break;\n",
       "    case 1:\n",
       "        cursor = 'default';\n",
       "        break;\n",
       "    case 2:\n",
       "        cursor = 'crosshair';\n",
       "        break;\n",
       "    case 3:\n",
       "        cursor = 'move';\n",
       "        break;\n",
       "    }\n",
       "    fig.rubberband_canvas.style.cursor = cursor;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_message = function(fig, msg) {\n",
       "    fig.message.textContent = msg['message'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_draw = function(fig, msg) {\n",
       "    // Request the server to send over a new figure.\n",
       "    fig.send_draw_message();\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_image_mode = function(fig, msg) {\n",
       "    fig.image_mode = msg['mode'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Called whenever the canvas gets updated.\n",
       "    this.send_message(\"ack\", {});\n",
       "}\n",
       "\n",
       "// A function to construct a web socket function for onmessage handling.\n",
       "// Called in the figure constructor.\n",
       "mpl.figure.prototype._make_on_message_function = function(fig) {\n",
       "    return function socket_on_message(evt) {\n",
       "        if (evt.data instanceof Blob) {\n",
       "            /* FIXME: We get \"Resource interpreted as Image but\n",
       "             * transferred with MIME type text/plain:\" errors on\n",
       "             * Chrome.  But how to set the MIME type?  It doesn't seem\n",
       "             * to be part of the websocket stream */\n",
       "            evt.data.type = \"image/png\";\n",
       "\n",
       "            /* Free the memory for the previous frames */\n",
       "            if (fig.imageObj.src) {\n",
       "                (window.URL || window.webkitURL).revokeObjectURL(\n",
       "                    fig.imageObj.src);\n",
       "            }\n",
       "\n",
       "            fig.imageObj.src = (window.URL || window.webkitURL).createObjectURL(\n",
       "                evt.data);\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "        else if (typeof evt.data === 'string' && evt.data.slice(0, 21) == \"data:image/png;base64\") {\n",
       "            fig.imageObj.src = evt.data;\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        var msg = JSON.parse(evt.data);\n",
       "        var msg_type = msg['type'];\n",
       "\n",
       "        // Call the  \"handle_{type}\" callback, which takes\n",
       "        // the figure and JSON message as its only arguments.\n",
       "        try {\n",
       "            var callback = fig[\"handle_\" + msg_type];\n",
       "        } catch (e) {\n",
       "            console.log(\"No handler for the '\" + msg_type + \"' message type: \", msg);\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        if (callback) {\n",
       "            try {\n",
       "                // console.log(\"Handling '\" + msg_type + \"' message: \", msg);\n",
       "                callback(fig, msg);\n",
       "            } catch (e) {\n",
       "                console.log(\"Exception inside the 'handler_\" + msg_type + \"' callback:\", e, e.stack, msg);\n",
       "            }\n",
       "        }\n",
       "    };\n",
       "}\n",
       "\n",
       "// from http://stackoverflow.com/questions/1114465/getting-mouse-location-in-canvas\n",
       "mpl.findpos = function(e) {\n",
       "    //this section is from http://www.quirksmode.org/js/events_properties.html\n",
       "    var targ;\n",
       "    if (!e)\n",
       "        e = window.event;\n",
       "    if (e.target)\n",
       "        targ = e.target;\n",
       "    else if (e.srcElement)\n",
       "        targ = e.srcElement;\n",
       "    if (targ.nodeType == 3) // defeat Safari bug\n",
       "        targ = targ.parentNode;\n",
       "\n",
       "    // jQuery normalizes the pageX and pageY\n",
       "    // pageX,Y are the mouse positions relative to the document\n",
       "    // offset() returns the position of the element relative to the document\n",
       "    var x = e.pageX - $(targ).offset().left;\n",
       "    var y = e.pageY - $(targ).offset().top;\n",
       "\n",
       "    return {\"x\": x, \"y\": y};\n",
       "};\n",
       "\n",
       "/*\n",
       " * return a copy of an object with only non-object keys\n",
       " * we need this to avoid circular references\n",
       " * http://stackoverflow.com/a/24161582/3208463\n",
       " */\n",
       "function simpleKeys (original) {\n",
       "  return Object.keys(original).reduce(function (obj, key) {\n",
       "    if (typeof original[key] !== 'object')\n",
       "        obj[key] = original[key]\n",
       "    return obj;\n",
       "  }, {});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.mouse_event = function(event, name) {\n",
       "    var canvas_pos = mpl.findpos(event)\n",
       "\n",
       "    if (name === 'button_press')\n",
       "    {\n",
       "        this.canvas.focus();\n",
       "        this.canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    var x = canvas_pos.x * mpl.ratio;\n",
       "    var y = canvas_pos.y * mpl.ratio;\n",
       "\n",
       "    this.send_message(name, {x: x, y: y, button: event.button,\n",
       "                             step: event.step,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "\n",
       "    /* This prevents the web browser from automatically changing to\n",
       "     * the text insertion cursor when the button is pressed.  We want\n",
       "     * to control all of the cursor setting manually through the\n",
       "     * 'cursor' event from matplotlib */\n",
       "    event.preventDefault();\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    // Handle any extra behaviour associated with a key event\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.key_event = function(event, name) {\n",
       "\n",
       "    // Prevent repeat events\n",
       "    if (name == 'key_press')\n",
       "    {\n",
       "        if (event.which === this._key)\n",
       "            return;\n",
       "        else\n",
       "            this._key = event.which;\n",
       "    }\n",
       "    if (name == 'key_release')\n",
       "        this._key = null;\n",
       "\n",
       "    var value = '';\n",
       "    if (event.ctrlKey && event.which != 17)\n",
       "        value += \"ctrl+\";\n",
       "    if (event.altKey && event.which != 18)\n",
       "        value += \"alt+\";\n",
       "    if (event.shiftKey && event.which != 16)\n",
       "        value += \"shift+\";\n",
       "\n",
       "    value += 'k';\n",
       "    value += event.which.toString();\n",
       "\n",
       "    this._key_event_extra(event, name);\n",
       "\n",
       "    this.send_message(name, {key: value,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onclick = function(name) {\n",
       "    if (name == 'download') {\n",
       "        this.handle_save(this, null);\n",
       "    } else {\n",
       "        this.send_message(\"toolbar_button\", {name: name});\n",
       "    }\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onmouseover = function(tooltip) {\n",
       "    this.message.textContent = tooltip;\n",
       "};\n",
       "mpl.toolbar_items = [[\"Home\", \"Reset original view\", \"fa fa-home icon-home\", \"home\"], [\"Back\", \"Back to  previous view\", \"fa fa-arrow-left icon-arrow-left\", \"back\"], [\"Forward\", \"Forward to next view\", \"fa fa-arrow-right icon-arrow-right\", \"forward\"], [\"\", \"\", \"\", \"\"], [\"Pan\", \"Pan axes with left mouse, zoom with right\", \"fa fa-arrows icon-move\", \"pan\"], [\"Zoom\", \"Zoom to rectangle\", \"fa fa-square-o icon-check-empty\", \"zoom\"], [\"\", \"\", \"\", \"\"], [\"Download\", \"Download plot\", \"fa fa-floppy-o icon-save\", \"download\"]];\n",
       "\n",
       "mpl.extensions = [\"eps\", \"jpeg\", \"pdf\", \"png\", \"ps\", \"raw\", \"svg\", \"tif\"];\n",
       "\n",
       "mpl.default_extension = \"png\";var comm_websocket_adapter = function(comm) {\n",
       "    // Create a \"websocket\"-like object which calls the given IPython comm\n",
       "    // object with the appropriate methods. Currently this is a non binary\n",
       "    // socket, so there is still some room for performance tuning.\n",
       "    var ws = {};\n",
       "\n",
       "    ws.close = function() {\n",
       "        comm.close()\n",
       "    };\n",
       "    ws.send = function(m) {\n",
       "        //console.log('sending', m);\n",
       "        comm.send(m);\n",
       "    };\n",
       "    // Register the callback with on_msg.\n",
       "    comm.on_msg(function(msg) {\n",
       "        //console.log('receiving', msg['content']['data'], msg);\n",
       "        // Pass the mpl event to the overridden (by mpl) onmessage function.\n",
       "        ws.onmessage(msg['content']['data'])\n",
       "    });\n",
       "    return ws;\n",
       "}\n",
       "\n",
       "mpl.mpl_figure_comm = function(comm, msg) {\n",
       "    // This is the function which gets called when the mpl process\n",
       "    // starts-up an IPython Comm through the \"matplotlib\" channel.\n",
       "\n",
       "    var id = msg.content.data.id;\n",
       "    // Get hold of the div created by the display call when the Comm\n",
       "    // socket was opened in Python.\n",
       "    var element = $(\"#\" + id);\n",
       "    var ws_proxy = comm_websocket_adapter(comm)\n",
       "\n",
       "    function ondownload(figure, format) {\n",
       "        window.open(figure.imageObj.src);\n",
       "    }\n",
       "\n",
       "    var fig = new mpl.figure(id, ws_proxy,\n",
       "                           ondownload,\n",
       "                           element.get(0));\n",
       "\n",
       "    // Call onopen now - mpl needs it, as it is assuming we've passed it a real\n",
       "    // web socket which is closed, not our websocket->open comm proxy.\n",
       "    ws_proxy.onopen();\n",
       "\n",
       "    fig.parent_element = element.get(0);\n",
       "    fig.cell_info = mpl.find_output_cell(\"<div id='\" + id + \"'></div>\");\n",
       "    if (!fig.cell_info) {\n",
       "        console.error(\"Failed to find cell for figure\", id, fig);\n",
       "        return;\n",
       "    }\n",
       "\n",
       "    var output_index = fig.cell_info[2]\n",
       "    var cell = fig.cell_info[0];\n",
       "\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.handle_close = function(fig, msg) {\n",
       "    var width = fig.canvas.width/mpl.ratio\n",
       "    fig.root.unbind('remove')\n",
       "\n",
       "    // Update the output cell to use the data from the current canvas.\n",
       "    fig.push_to_output();\n",
       "    var dataURL = fig.canvas.toDataURL();\n",
       "    // Re-enable the keyboard manager in IPython - without this line, in FF,\n",
       "    // the notebook keyboard shortcuts fail.\n",
       "    IPython.keyboard_manager.enable()\n",
       "    $(fig.parent_element).html('<img src=\"' + dataURL + '\" width=\"' + width + '\">');\n",
       "    fig.close_ws(fig, msg);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.close_ws = function(fig, msg){\n",
       "    fig.send_message('closing', msg);\n",
       "    // fig.ws.close()\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.push_to_output = function(remove_interactive) {\n",
       "    // Turn the data on the canvas into data in the output cell.\n",
       "    var width = this.canvas.width/mpl.ratio\n",
       "    var dataURL = this.canvas.toDataURL();\n",
       "    this.cell_info[1]['text/html'] = '<img src=\"' + dataURL + '\" width=\"' + width + '\">';\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Tell IPython that the notebook contents must change.\n",
       "    IPython.notebook.set_dirty(true);\n",
       "    this.send_message(\"ack\", {});\n",
       "    var fig = this;\n",
       "    // Wait a second, then push the new image to the DOM so\n",
       "    // that it is saved nicely (might be nice to debounce this).\n",
       "    setTimeout(function () { fig.push_to_output() }, 1000);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>')\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items){\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) { continue; };\n",
       "\n",
       "        var button = $('<button class=\"btn btn-default\" href=\"#\" title=\"' + name + '\"><i class=\"fa ' + image + ' fa-lg\"></i></button>');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    // Add the status bar.\n",
       "    var status_bar = $('<span class=\"mpl-message\" style=\"text-align:right; float: right;\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "\n",
       "    // Add the close button to the window.\n",
       "    var buttongrp = $('<div class=\"btn-group inline pull-right\"></div>');\n",
       "    var button = $('<button class=\"btn btn-mini btn-primary\" href=\"#\" title=\"Stop Interaction\"><i class=\"fa fa-power-off icon-remove icon-large\"></i></button>');\n",
       "    button.click(function (evt) { fig.handle_close(fig, {}); } );\n",
       "    button.mouseover('Stop Interaction', toolbar_mouse_event);\n",
       "    buttongrp.append(button);\n",
       "    var titlebar = this.root.find($('.ui-dialog-titlebar'));\n",
       "    titlebar.prepend(buttongrp);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(el){\n",
       "    var fig = this\n",
       "    el.on(\"remove\", function(){\n",
       "\tfig.close_ws(fig, {});\n",
       "    });\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(el){\n",
       "    // this is important to make the div 'focusable\n",
       "    el.attr('tabindex', 0)\n",
       "    // reach out to IPython and tell the keyboard manager to turn it's self\n",
       "    // off when our div gets focus\n",
       "\n",
       "    // location in version 3\n",
       "    if (IPython.notebook.keyboard_manager) {\n",
       "        IPython.notebook.keyboard_manager.register_events(el);\n",
       "    }\n",
       "    else {\n",
       "        // location in version 2\n",
       "        IPython.keyboard_manager.register_events(el);\n",
       "    }\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    var manager = IPython.notebook.keyboard_manager;\n",
       "    if (!manager)\n",
       "        manager = IPython.keyboard_manager;\n",
       "\n",
       "    // Check for shift+enter\n",
       "    if (event.shiftKey && event.which == 13) {\n",
       "        this.canvas_div.blur();\n",
       "        event.shiftKey = false;\n",
       "        // Send a \"J\" for go to next cell\n",
       "        event.which = 74;\n",
       "        event.keyCode = 74;\n",
       "        manager.command_mode();\n",
       "        manager.handle_keydown(event);\n",
       "    }\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    fig.ondownload(fig, null);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.find_output_cell = function(html_output) {\n",
       "    // Return the cell and output element which can be found *uniquely* in the notebook.\n",
       "    // Note - this is a bit hacky, but it is done because the \"notebook_saving.Notebook\"\n",
       "    // IPython event is triggered only after the cells have been serialised, which for\n",
       "    // our purposes (turning an active figure into a static one), is too late.\n",
       "    var cells = IPython.notebook.get_cells();\n",
       "    var ncells = cells.length;\n",
       "    for (var i=0; i<ncells; i++) {\n",
       "        var cell = cells[i];\n",
       "        if (cell.cell_type === 'code'){\n",
       "            for (var j=0; j<cell.output_area.outputs.length; j++) {\n",
       "                var data = cell.output_area.outputs[j];\n",
       "                if (data.data) {\n",
       "                    // IPython >= 3 moved mimebundle to data attribute of output\n",
       "                    data = data.data;\n",
       "                }\n",
       "                if (data['text/html'] == html_output) {\n",
       "                    return [cell, data, j];\n",
       "                }\n",
       "            }\n",
       "        }\n",
       "    }\n",
       "}\n",
       "\n",
       "// Register the function which deals with the matplotlib target/channel.\n",
       "// The kernel may be null if the page has been refreshed.\n",
       "if (IPython.notebook.kernel != null) {\n",
       "    IPython.notebook.kernel.comm_manager.register_target('matplotlib', mpl.mpl_figure_comm);\n",
       "}\n"
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"data:image/png;base64,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\" width=\"639.999990463257\">"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<sigpy.plot.ImagePlot at 0x7f4068150f98>"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pl.ImagePlot(np.array(train_dataset.train_data))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create Training data\n",
    "train_dataset_im = train_dataset.train_data\n",
    "train_dataset_label = train_dataset.train_labels\n",
    "n_train = train_dataset_im.shape[0]\n",
    "n_permute = np.random.permutation(n_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Define Neural network Structure\n",
    "class ConvNet(nn.Module):\n",
    "    def __init__(self, num_classes=10):\n",
    "        super(ConvNet, self).__init__()\n",
    "        self.layer1 = nn.Sequential(\n",
    "            nn.Conv2d(1, 16, kernel_size=5, stride=1, padding=2),\n",
    "            nn.BatchNorm2d(16),\n",
    "            nn.ReLU(),\n",
    "            nn.MaxPool2d(kernel_size=2, stride=2))\n",
    "        self.layer2 = nn.Sequential(\n",
    "            nn.Conv2d(16, 32, kernel_size=5, stride=1, padding=2),\n",
    "            nn.BatchNorm2d(32),\n",
    "            nn.ReLU(),\n",
    "            nn.MaxPool2d(kernel_size=2, stride=2))\n",
    "        self.fc = nn.Linear(7*7*32, num_classes)\n",
    "        \n",
    "    def forward(self, x):\n",
    "        out = self.layer1(x)\n",
    "        out = self.layer2(out)\n",
    "        out = out.reshape(out.size(0), -1)\n",
    "        out = self.fc(out)\n",
    "        return out"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [1/50], Step [100/600], Loss: 0.2366\n",
      "Epoch [1/50], Step [200/600], Loss: 0.0646\n",
      "Epoch [1/50], Step [300/600], Loss: 0.1827\n",
      "Epoch [1/50], Step [400/600], Loss: 0.0313\n",
      "Epoch [1/50], Step [500/600], Loss: 0.1192\n",
      "Epoch [1/50], Step [600/600], Loss: 0.0138\n",
      "Epoch [2/50], Step [100/600], Loss: 0.1097\n",
      "Epoch [2/50], Step [200/600], Loss: 0.0252\n",
      "Epoch [2/50], Step [300/600], Loss: 0.0739\n",
      "Epoch [2/50], Step [400/600], Loss: 0.0109\n",
      "Epoch [2/50], Step [500/600], Loss: 0.0943\n",
      "Epoch [2/50], Step [600/600], Loss: 0.0056\n",
      "Epoch [3/50], Step [100/600], Loss: 0.1040\n",
      "Epoch [3/50], Step [200/600], Loss: 0.0284\n",
      "Epoch [3/50], Step [300/600], Loss: 0.0432\n",
      "Epoch [3/50], Step [400/600], Loss: 0.0057\n",
      "Epoch [3/50], Step [500/600], Loss: 0.0849\n",
      "Epoch [3/50], Step [600/600], Loss: 0.0045\n",
      "Epoch [4/50], Step [100/600], Loss: 0.0789\n",
      "Epoch [4/50], Step [200/600], Loss: 0.0262\n",
      "Epoch [4/50], Step [300/600], Loss: 0.0283\n",
      "Epoch [4/50], Step [400/600], Loss: 0.0036\n",
      "Epoch [4/50], Step [500/600], Loss: 0.0624\n",
      "Epoch [4/50], Step [600/600], Loss: 0.0031\n",
      "Epoch [5/50], Step [100/600], Loss: 0.0480\n",
      "Epoch [5/50], Step [200/600], Loss: 0.0211\n",
      "Epoch [5/50], Step [300/600], Loss: 0.0246\n",
      "Epoch [5/50], Step [400/600], Loss: 0.0024\n",
      "Epoch [5/50], Step [500/600], Loss: 0.0462\n",
      "Epoch [5/50], Step [600/600], Loss: 0.0023\n",
      "Epoch [6/50], Step [100/600], Loss: 0.0295\n",
      "Epoch [6/50], Step [200/600], Loss: 0.0121\n",
      "Epoch [6/50], Step [300/600], Loss: 0.0180\n",
      "Epoch [6/50], Step [400/600], Loss: 0.0018\n",
      "Epoch [6/50], Step [500/600], Loss: 0.0354\n",
      "Epoch [6/50], Step [600/600], Loss: 0.0022\n",
      "Epoch [7/50], Step [100/600], Loss: 0.0119\n",
      "Epoch [7/50], Step [200/600], Loss: 0.0079\n",
      "Epoch [7/50], Step [300/600], Loss: 0.0119\n",
      "Epoch [7/50], Step [400/600], Loss: 0.0014\n",
      "Epoch [7/50], Step [500/600], Loss: 0.0186\n",
      "Epoch [7/50], Step [600/600], Loss: 0.0016\n",
      "Epoch [8/50], Step [100/600], Loss: 0.0048\n",
      "Epoch [8/50], Step [200/600], Loss: 0.0046\n",
      "Epoch [8/50], Step [300/600], Loss: 0.0111\n",
      "Epoch [8/50], Step [400/600], Loss: 0.0014\n",
      "Epoch [8/50], Step [500/600], Loss: 0.0089\n",
      "Epoch [8/50], Step [600/600], Loss: 0.0006\n",
      "Epoch [9/50], Step [100/600], Loss: 0.0058\n",
      "Epoch [9/50], Step [200/600], Loss: 0.0025\n",
      "Epoch [9/50], Step [300/600], Loss: 0.0057\n",
      "Epoch [9/50], Step [400/600], Loss: 0.0008\n",
      "Epoch [9/50], Step [500/600], Loss: 0.0050\n",
      "Epoch [9/50], Step [600/600], Loss: 0.0008\n",
      "Epoch [10/50], Step [100/600], Loss: 0.0038\n",
      "Epoch [10/50], Step [200/600], Loss: 0.0018\n",
      "Epoch [10/50], Step [300/600], Loss: 0.0048\n",
      "Epoch [10/50], Step [400/600], Loss: 0.0009\n",
      "Epoch [10/50], Step [500/600], Loss: 0.0091\n",
      "Epoch [10/50], Step [600/600], Loss: 0.0007\n",
      "Epoch [11/50], Step [100/600], Loss: 0.0081\n",
      "Epoch [11/50], Step [200/600], Loss: 0.0029\n",
      "Epoch [11/50], Step [300/600], Loss: 0.0192\n",
      "Epoch [11/50], Step [400/600], Loss: 0.0013\n",
      "Epoch [11/50], Step [500/600], Loss: 0.0022\n",
      "Epoch [11/50], Step [600/600], Loss: 0.0026\n",
      "Epoch [12/50], Step [100/600], Loss: 0.0023\n",
      "Epoch [12/50], Step [200/600], Loss: 0.0009\n",
      "Epoch [12/50], Step [300/600], Loss: 0.0029\n",
      "Epoch [12/50], Step [400/600], Loss: 0.0004\n",
      "Epoch [12/50], Step [500/600], Loss: 0.0030\n",
      "Epoch [12/50], Step [600/600], Loss: 0.0006\n",
      "Epoch [13/50], Step [100/600], Loss: 0.0043\n",
      "Epoch [13/50], Step [200/600], Loss: 0.0008\n",
      "Epoch [13/50], Step [300/600], Loss: 0.0251\n",
      "Epoch [13/50], Step [400/600], Loss: 0.0001\n",
      "Epoch [13/50], Step [500/600], Loss: 0.0109\n",
      "Epoch [13/50], Step [600/600], Loss: 0.0003\n",
      "Epoch [14/50], Step [100/600], Loss: 0.0016\n",
      "Epoch [14/50], Step [200/600], Loss: 0.0060\n",
      "Epoch [14/50], Step [300/600], Loss: 0.0108\n",
      "Epoch [14/50], Step [400/600], Loss: 0.0001\n",
      "Epoch [14/50], Step [500/600], Loss: 0.0020\n",
      "Epoch [14/50], Step [600/600], Loss: 0.0027\n",
      "Epoch [15/50], Step [100/600], Loss: 0.0014\n",
      "Epoch [15/50], Step [200/600], Loss: 0.0002\n",
      "Epoch [15/50], Step [300/600], Loss: 0.0024\n",
      "Epoch [15/50], Step [400/600], Loss: 0.0001\n",
      "Epoch [15/50], Step [500/600], Loss: 0.0005\n",
      "Epoch [15/50], Step [600/600], Loss: 0.0073\n",
      "Epoch [16/50], Step [100/600], Loss: 0.0014\n",
      "Epoch [16/50], Step [200/600], Loss: 0.0072\n",
      "Epoch [16/50], Step [300/600], Loss: 0.0078\n",
      "Epoch [16/50], Step [400/600], Loss: 0.0001\n",
      "Epoch [16/50], Step [500/600], Loss: 0.0014\n",
      "Epoch [16/50], Step [600/600], Loss: 0.0020\n",
      "Epoch [17/50], Step [100/600], Loss: 0.0170\n",
      "Epoch [17/50], Step [200/600], Loss: 0.0011\n",
      "Epoch [17/50], Step [300/600], Loss: 0.0202\n",
      "Epoch [17/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [17/50], Step [500/600], Loss: 0.0004\n",
      "Epoch [17/50], Step [600/600], Loss: 0.0001\n",
      "Epoch [18/50], Step [100/600], Loss: 0.0022\n",
      "Epoch [18/50], Step [200/600], Loss: 0.0003\n",
      "Epoch [18/50], Step [300/600], Loss: 0.0008\n",
      "Epoch [18/50], Step [400/600], Loss: 0.0005\n",
      "Epoch [18/50], Step [500/600], Loss: 0.0003\n",
      "Epoch [18/50], Step [600/600], Loss: 0.0002\n",
      "Epoch [19/50], Step [100/600], Loss: 0.0002\n",
      "Epoch [19/50], Step [200/600], Loss: 0.0004\n",
      "Epoch [19/50], Step [300/600], Loss: 0.0016\n",
      "Epoch [19/50], Step [400/600], Loss: 0.0001\n",
      "Epoch [19/50], Step [500/600], Loss: 0.0003\n",
      "Epoch [19/50], Step [600/600], Loss: 0.0016\n",
      "Epoch [20/50], Step [100/600], Loss: 0.0018\n",
      "Epoch [20/50], Step [200/600], Loss: 0.0002\n",
      "Epoch [20/50], Step [300/600], Loss: 0.0006\n",
      "Epoch [20/50], Step [400/600], Loss: 0.0005\n",
      "Epoch [20/50], Step [500/600], Loss: 0.0012\n",
      "Epoch [20/50], Step [600/600], Loss: 0.0038\n",
      "Epoch [21/50], Step [100/600], Loss: 0.0004\n",
      "Epoch [21/50], Step [200/600], Loss: 0.0025\n",
      "Epoch [21/50], Step [300/600], Loss: 0.0001\n",
      "Epoch [21/50], Step [400/600], Loss: 0.0003\n",
      "Epoch [21/50], Step [500/600], Loss: 0.0003\n",
      "Epoch [21/50], Step [600/600], Loss: 0.0005\n",
      "Epoch [22/50], Step [100/600], Loss: 0.0003\n",
      "Epoch [22/50], Step [200/600], Loss: 0.0039\n",
      "Epoch [22/50], Step [300/600], Loss: 0.0003\n",
      "Epoch [22/50], Step [400/600], Loss: 0.0114\n",
      "Epoch [22/50], Step [500/600], Loss: 0.0355\n",
      "Epoch [22/50], Step [600/600], Loss: 0.0002\n",
      "Epoch [23/50], Step [100/600], Loss: 0.0206\n",
      "Epoch [23/50], Step [200/600], Loss: 0.0035\n",
      "Epoch [23/50], Step [300/600], Loss: 0.0005\n",
      "Epoch [23/50], Step [400/600], Loss: 0.0002\n",
      "Epoch [23/50], Step [500/600], Loss: 0.0002\n",
      "Epoch [23/50], Step [600/600], Loss: 0.0001\n",
      "Epoch [24/50], Step [100/600], Loss: 0.0001\n",
      "Epoch [24/50], Step [200/600], Loss: 0.0004\n",
      "Epoch [24/50], Step [300/600], Loss: 0.0019\n",
      "Epoch [24/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [24/50], Step [500/600], Loss: 0.0001\n",
      "Epoch [24/50], Step [600/600], Loss: 0.0002\n",
      "Epoch [25/50], Step [100/600], Loss: 0.0000\n",
      "Epoch [25/50], Step [200/600], Loss: 0.0009\n",
      "Epoch [25/50], Step [300/600], Loss: 0.0005\n",
      "Epoch [25/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [25/50], Step [500/600], Loss: 0.0001\n",
      "Epoch [25/50], Step [600/600], Loss: 0.0000\n",
      "Epoch [26/50], Step [100/600], Loss: 0.0001\n",
      "Epoch [26/50], Step [200/600], Loss: 0.0001\n",
      "Epoch [26/50], Step [300/600], Loss: 0.0048\n",
      "Epoch [26/50], Step [400/600], Loss: 0.0001\n",
      "Epoch [26/50], Step [500/600], Loss: 0.0041\n",
      "Epoch [26/50], Step [600/600], Loss: 0.0003\n",
      "Epoch [27/50], Step [100/600], Loss: 0.0015\n",
      "Epoch [27/50], Step [200/600], Loss: 0.0001\n",
      "Epoch [27/50], Step [300/600], Loss: 0.0002\n",
      "Epoch [27/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [27/50], Step [500/600], Loss: 0.0094\n",
      "Epoch [27/50], Step [600/600], Loss: 0.0001\n",
      "Epoch [28/50], Step [100/600], Loss: 0.0001\n",
      "Epoch [28/50], Step [200/600], Loss: 0.0002\n",
      "Epoch [28/50], Step [300/600], Loss: 0.0014\n",
      "Epoch [28/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [28/50], Step [500/600], Loss: 0.0000\n",
      "Epoch [28/50], Step [600/600], Loss: 0.0000\n",
      "Epoch [29/50], Step [100/600], Loss: 0.0000\n",
      "Epoch [29/50], Step [200/600], Loss: 0.0028\n",
      "Epoch [29/50], Step [300/600], Loss: 0.0001\n",
      "Epoch [29/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [29/50], Step [500/600], Loss: 0.0001\n",
      "Epoch [29/50], Step [600/600], Loss: 0.0000\n",
      "Epoch [30/50], Step [100/600], Loss: 0.0006\n",
      "Epoch [30/50], Step [200/600], Loss: 0.0023\n",
      "Epoch [30/50], Step [300/600], Loss: 0.0241\n",
      "Epoch [30/50], Step [400/600], Loss: 0.0002\n",
      "Epoch [30/50], Step [500/600], Loss: 0.0000\n",
      "Epoch [30/50], Step [600/600], Loss: 0.0001\n",
      "Epoch [31/50], Step [100/600], Loss: 0.0003\n",
      "Epoch [31/50], Step [200/600], Loss: 0.0000\n",
      "Epoch [31/50], Step [300/600], Loss: 0.0033\n",
      "Epoch [31/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [31/50], Step [500/600], Loss: 0.0001\n",
      "Epoch [31/50], Step [600/600], Loss: 0.0001\n",
      "Epoch [32/50], Step [100/600], Loss: 0.0003\n",
      "Epoch [32/50], Step [200/600], Loss: 0.0000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [32/50], Step [300/600], Loss: 0.0007\n",
      "Epoch [32/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [32/50], Step [500/600], Loss: 0.0001\n",
      "Epoch [32/50], Step [600/600], Loss: 0.0000\n",
      "Epoch [33/50], Step [100/600], Loss: 0.0001\n",
      "Epoch [33/50], Step [200/600], Loss: 0.0000\n",
      "Epoch [33/50], Step [300/600], Loss: 0.0003\n",
      "Epoch [33/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [33/50], Step [500/600], Loss: 0.0002\n",
      "Epoch [33/50], Step [600/600], Loss: 0.0000\n",
      "Epoch [34/50], Step [100/600], Loss: 0.0016\n",
      "Epoch [34/50], Step [200/600], Loss: 0.0005\n",
      "Epoch [34/50], Step [300/600], Loss: 0.0109\n",
      "Epoch [34/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [34/50], Step [500/600], Loss: 0.0001\n",
      "Epoch [34/50], Step [600/600], Loss: 0.0001\n",
      "Epoch [35/50], Step [100/600], Loss: 0.0001\n",
      "Epoch [35/50], Step [200/600], Loss: 0.0005\n",
      "Epoch [35/50], Step [300/600], Loss: 0.0001\n",
      "Epoch [35/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [35/50], Step [500/600], Loss: 0.0000\n",
      "Epoch [35/50], Step [600/600], Loss: 0.0000\n",
      "Epoch [36/50], Step [100/600], Loss: 0.0000\n",
      "Epoch [36/50], Step [200/600], Loss: 0.0001\n",
      "Epoch [36/50], Step [300/600], Loss: 0.0000\n",
      "Epoch [36/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [36/50], Step [500/600], Loss: 0.0000\n",
      "Epoch [36/50], Step [600/600], Loss: 0.0000\n",
      "Epoch [37/50], Step [100/600], Loss: 0.0000\n",
      "Epoch [37/50], Step [200/600], Loss: 0.0001\n",
      "Epoch [37/50], Step [300/600], Loss: 0.0006\n",
      "Epoch [37/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [37/50], Step [500/600], Loss: 0.0013\n",
      "Epoch [37/50], Step [600/600], Loss: 0.0001\n",
      "Epoch [38/50], Step [100/600], Loss: 0.0002\n",
      "Epoch [38/50], Step [200/600], Loss: 0.0102\n",
      "Epoch [38/50], Step [300/600], Loss: 0.0001\n",
      "Epoch [38/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [38/50], Step [500/600], Loss: 0.0002\n",
      "Epoch [38/50], Step [600/600], Loss: 0.0006\n",
      "Epoch [39/50], Step [100/600], Loss: 0.0000\n",
      "Epoch [39/50], Step [200/600], Loss: 0.0015\n",
      "Epoch [39/50], Step [300/600], Loss: 0.0000\n",
      "Epoch [39/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [39/50], Step [500/600], Loss: 0.0000\n",
      "Epoch [39/50], Step [600/600], Loss: 0.0000\n",
      "Epoch [40/50], Step [100/600], Loss: 0.0000\n",
      "Epoch [40/50], Step [200/600], Loss: 0.0001\n",
      "Epoch [40/50], Step [300/600], Loss: 0.0000\n",
      "Epoch [40/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [40/50], Step [500/600], Loss: 0.0000\n",
      "Epoch [40/50], Step [600/600], Loss: 0.0000\n",
      "Epoch [41/50], Step [100/600], Loss: 0.0000\n",
      "Epoch [41/50], Step [200/600], Loss: 0.0001\n",
      "Epoch [41/50], Step [300/600], Loss: 0.0000\n",
      "Epoch [41/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [41/50], Step [500/600], Loss: 0.0000\n",
      "Epoch [41/50], Step [600/600], Loss: 0.0000\n",
      "Epoch [42/50], Step [100/600], Loss: 0.0000\n",
      "Epoch [42/50], Step [200/600], Loss: 0.0001\n",
      "Epoch [42/50], Step [300/600], Loss: 0.0000\n",
      "Epoch [42/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [42/50], Step [500/600], Loss: 0.0000\n",
      "Epoch [42/50], Step [600/600], Loss: 0.0000\n",
      "Epoch [43/50], Step [100/600], Loss: 0.0000\n",
      "Epoch [43/50], Step [200/600], Loss: 0.0001\n",
      "Epoch [43/50], Step [300/600], Loss: 0.0000\n",
      "Epoch [43/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [43/50], Step [500/600], Loss: 0.0000\n",
      "Epoch [43/50], Step [600/600], Loss: 0.0000\n",
      "Epoch [44/50], Step [100/600], Loss: 0.0000\n",
      "Epoch [44/50], Step [200/600], Loss: 0.0001\n",
      "Epoch [44/50], Step [300/600], Loss: 0.0000\n",
      "Epoch [44/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [44/50], Step [500/600], Loss: 0.0000\n",
      "Epoch [44/50], Step [600/600], Loss: 0.0000\n",
      "Epoch [45/50], Step [100/600], Loss: 0.0000\n",
      "Epoch [45/50], Step [200/600], Loss: 0.0000\n",
      "Epoch [45/50], Step [300/600], Loss: 0.0000\n",
      "Epoch [45/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [45/50], Step [500/600], Loss: 0.0000\n",
      "Epoch [45/50], Step [600/600], Loss: 0.0000\n",
      "Epoch [46/50], Step [100/600], Loss: 0.0000\n",
      "Epoch [46/50], Step [200/600], Loss: 0.0000\n",
      "Epoch [46/50], Step [300/600], Loss: 0.0000\n",
      "Epoch [46/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [46/50], Step [500/600], Loss: 0.0000\n",
      "Epoch [46/50], Step [600/600], Loss: 0.0000\n",
      "Epoch [47/50], Step [100/600], Loss: 0.0000\n",
      "Epoch [47/50], Step [200/600], Loss: 0.0000\n",
      "Epoch [47/50], Step [300/600], Loss: 0.0000\n",
      "Epoch [47/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [47/50], Step [500/600], Loss: 0.0000\n",
      "Epoch [47/50], Step [600/600], Loss: 0.0000\n",
      "Epoch [48/50], Step [100/600], Loss: 0.0000\n",
      "Epoch [48/50], Step [200/600], Loss: 0.0000\n",
      "Epoch [48/50], Step [300/600], Loss: 0.0000\n",
      "Epoch [48/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [48/50], Step [500/600], Loss: 0.0000\n",
      "Epoch [48/50], Step [600/600], Loss: 0.0000\n",
      "Epoch [49/50], Step [100/600], Loss: 0.0000\n",
      "Epoch [49/50], Step [200/600], Loss: 0.0000\n",
      "Epoch [49/50], Step [300/600], Loss: 0.0000\n",
      "Epoch [49/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [49/50], Step [500/600], Loss: 0.0000\n",
      "Epoch [49/50], Step [600/600], Loss: 0.0000\n",
      "Epoch [50/50], Step [100/600], Loss: 0.0000\n",
      "Epoch [50/50], Step [200/600], Loss: 0.0000\n",
      "Epoch [50/50], Step [300/600], Loss: 0.0000\n",
      "Epoch [50/50], Step [400/600], Loss: 0.0000\n",
      "Epoch [50/50], Step [500/600], Loss: 0.0000\n",
      "Epoch [50/50], Step [600/600], Loss: 0.0000\n"
     ]
    }
   ],
   "source": [
    "# Define a network\n",
    "model = ConvNet(num_classes).to(device)\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate)\n",
    "total_step = int(n_train/batch_size)\n",
    "for epoch in range(num_epochs):\n",
    "    for i in range(total_step):\n",
    "        index_train = n_permute[i*batch_size:(i+1)*batch_size]\n",
    "        images = train_dataset_im[index_train,:,:].unsqueeze(1).to(device).float()/255.0\n",
    "        labels = train_dataset_label[index_train].to(device)\n",
    "        \n",
    "        # Forward pass\n",
    "        outputs = model(images)\n",
    "        loss = criterion(outputs, labels)\n",
    "        \n",
    "        # Backward and optimize\n",
    "        optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        \n",
    "        if (i+1) % 100 == 0:\n",
    "            print ('Epoch [{}/{}], Step [{}/{}], Loss: {:.4f}' \n",
    "                   .format(epoch+1, num_epochs, i+1, total_step, loss.item()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create Testing data\n",
    "test_dataset_im = test_dataset.test_data\n",
    "test_dataset_label = test_dataset.test_labels\n",
    "n_test = test_dataset_im.shape[0]\n",
    "n_permute_test = np.random.permutation(n_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test Accuracy of the model on the 10000 test images: 99.33 %\n"
     ]
    }
   ],
   "source": [
    "# Test the model\n",
    "model.eval()  # eval mode (batchnorm uses moving mean/variance instead of mini-batch mean/variance)\n",
    "total_step = int(n_test/batch_size)\n",
    "with torch.no_grad():\n",
    "    correct = 0\n",
    "    total = 0\n",
    "    for i in range(total_step):\n",
    "        index_test = n_permute_test[i*batch_size:(i+1)*batch_size]\n",
    "        images = test_dataset_im[index_test,:,:].unsqueeze(1).to(device).float()/255.0\n",
    "        labels = test_dataset_label[index_test].to(device)\n",
    "\n",
    "        outputs = model(images)\n",
    "        _, predicted = torch.max(outputs.data, 1)\n",
    "        total += labels.size(0)\n",
    "        correct += (predicted == labels).sum().item()\n",
    "\n",
    "    print('Test Accuracy of the model on the 10000 test images: {} %'.format(100 * correct / total))\n",
    "\n",
    "# Save the model checkpoint\n",
    "torch.save(model.state_dict(), 'model.ckpt')\n"
   ]
  }
 ],
 "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
}