train-model.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This notebook shows how to train a Transparency by Design network (TbD-net) on the CLEVR dataset. Training for CoGenT can be done in much the same manner, simply changing the training and validation loaders and updating the evaluation code to evaluate over Condition A and Condition B."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from pathlib import Path\n",
    "\n",
    "import torch\n",
    "import torch.optim as optim\n",
    "import numpy as np\n",
    "\n",
    "from utils.clevr import load_vocab, ClevrDataLoaderNumpy, ClevrDataLoaderH5\n",
    "from tbd.module_net import TbDNet"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We want to run on CUDA if we have CUDA support so that we can train our model in a reasonable amount of time. We'll define the device accordingly."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "device = 'cuda' if torch.cuda.is_available() else 'cpu'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The first step is to create our model.\n",
    "\n",
    "We first load the `vocab` file, which holds a list of all the modules we're going to need. We then pass the vocab into our `TbDNet`, which creates an appropriate neural module for each operation in the vocab."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "vocab = load_vocab(Path('data/vocab.json'))\n",
    "\n",
    "# to train on 14x14 feature maps, pass feature_dim=(1024, 14, 14) to TbDNet\n",
    "tbd_net = TbDNet(vocab).to(device)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next up, we need to load our data.\n",
    "\n",
    "The `use_numpy_format` flag indicates whether we're loading data using npy files or HDF5 files.\n",
    "\n",
    "We create the appropriate `DataLoader` objects depending on the choice."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reading features from  data/train_features_hres.npy\n",
      "Reading questions from  data/training/train_questions.npy\n",
      "Reading image indices from data/training/train_image_idxs.npy\n",
      "Reading programs from data/training/train_programs.npy\n",
      "Reading answers from data/training/train_answers.npy\n",
      "Reading features from  data/val_features_hres.npy\n",
      "Reading questions from  data/validation/val_questions.npy\n",
      "Reading image indices from data/validation/val_image_idxs.npy\n",
      "Reading programs from data/validation/val_programs.npy\n",
      "Reading answers from data/validation/val_answers.npy\n"
     ]
    }
   ],
   "source": [
    "use_numpy_format = True\n",
    "\n",
    "if use_numpy_format:\n",
    "    train_loader_kwargs = {\n",
    "        'question_np': Path('data/training/train_questions.npy'),\n",
    "        'image_idx_np': Path('data/training/train_image_idxs.npy'),\n",
    "        'program_np': Path('data/training/train_programs.npy'),\n",
    "        'feature_np': Path('data/train_features_hres.npy'),\n",
    "        'answer_np': Path('data/training/train_answers.npy'),\n",
    "        'batch_size': 128,\n",
    "        'num_workers': 2,\n",
    "        'shuffle': True\n",
    "    }\n",
    "\n",
    "    val_loader_kwargs = {\n",
    "        'question_np': Path('data/validation/val_questions.npy'),\n",
    "        'image_idx_np': Path('data/validation/val_image_idxs.npy'),\n",
    "        'program_np': Path('data/validation/val_programs.npy'),\n",
    "        'feature_np': Path('clevr-iep/data/val_features_hres.npy'),\n",
    "        'answer_np': Path('data/validation/val_answers.npy'),\n",
    "        'batch_size': 128,\n",
    "        'num_workers': 2,\n",
    "        'shuffle': False\n",
    "    }\n",
    "    \n",
    "    train_loader = ClevrDataLoaderNumpy(**train_loader_kwargs)\n",
    "    val_loader = ClevrDataLoaderNumpy(**val_loader_kwargs)\n",
    "else:\n",
    "    train_loader_kwargs = {\n",
    "        'question_h5': Path('data/training/train_questions.h5'),\n",
    "        'feature_h5': Path('data/training/train_features_hres.h5'),\n",
    "        'batch_size': 128,\n",
    "        'num_workers': 2,\n",
    "        'shuffle': True\n",
    "    }\n",
    "    val_loader_kwargs = {\n",
    "        'question_h5': Path('data/validation/val_questions.h5'),\n",
    "        'feature_h5': Path('data/validation/val_features_hres.h5'),\n",
    "        'batch_size': 128,\n",
    "        'num_workers': 2,\n",
    "        'shuffle': False\n",
    "    }\n",
    "    \n",
    "    train_loader = ClevrDataLoaderH5(**train_loader_kwargs)\n",
    "    val_loader = ClevrDataLoaderH5(**val_loader_kwargs)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The answer format we prefer is different from that of Justin Johson *et al.*\n",
    "\n",
    "Their answers are sorted according to ASCII value, while we group by answer type and sort these alphabetically. Note that we have fewer answers than them, because we do not list the special tokens `<NULL>`, `<START>`, `<END>`, and `<UNK>` as answers."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def map_ans(answers):\n",
    "    ''' Map the answers from the format Justin Johnson et al. use to our own format. '''\n",
    "    ans_tensor = torch.LongTensor(answers.size())\n",
    "    ans_tensor[answers == 15] = 0  # blue\n",
    "    ans_tensor[answers == 16] = 1  # brown\n",
    "    ans_tensor[answers == 18] = 2  # cyan\n",
    "    ans_tensor[answers == 20] = 3  # gray\n",
    "    ans_tensor[answers == 21] = 4  # green\n",
    "    ans_tensor[answers == 25] = 5  # purple\n",
    "    ans_tensor[answers == 26] = 6  # red\n",
    "    ans_tensor[answers == 30] = 7  # yellow\n",
    "    ans_tensor[answers == 17] = 8  # cube\n",
    "    ans_tensor[answers == 19] = 9  # cylinder\n",
    "    ans_tensor[answers == 29] = 10 # sphere\n",
    "    ans_tensor[answers == 22] = 11 # large\n",
    "    ans_tensor[answers == 28] = 12 # small\n",
    "    ans_tensor[answers == 23] = 13 # metal\n",
    "    ans_tensor[answers == 27] = 14 # rubber\n",
    "    ans_tensor[answers == 24] = 15 # no\n",
    "    ans_tensor[answers == 31] = 16 # yes\n",
    "    ans_tensor[answers == 4] = 17  # 0\n",
    "    ans_tensor[answers == 5] = 18  # 1\n",
    "    ans_tensor[answers == 6] = 19  # 10 <- originally sorted by ASCII value, not numerical\n",
    "    ans_tensor[answers == 7] = 20  # 2\n",
    "    ans_tensor[answers == 8] = 21  # 3\n",
    "    ans_tensor[answers == 9] = 22  # 4\n",
    "    ans_tensor[answers == 10] = 23 # 5\n",
    "    ans_tensor[answers == 11] = 24 # 6\n",
    "    ans_tensor[answers == 12] = 25 # 7\n",
    "    ans_tensor[answers == 13] = 26 # 8\n",
    "    ans_tensor[answers == 14] = 27 # 9\n",
    "    return ans_tensor"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "At each epoch, we'll save our model's state, the epoch, and the training state."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def save_checkpoint(epoch, filename):\n",
    "    ''' Save the training state. '''\n",
    "    state = {\n",
    "        'epoch': epoch,\n",
    "        'state_dict': tbd_net.state_dict(),\n",
    "        'optimizer': optimizer.state_dict()\n",
    "        }\n",
    "    torch.save(state, filename)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def check_accuracy(loader):\n",
    "    '''\n",
    "    Convenience function for checking the accuracy of the model \n",
    "    over all the data in a given `DataLoader`\n",
    "    '''\n",
    "    torch.set_grad_enabled(False)\n",
    "    num_correct, num_samples = 0, 0\n",
    "    for batch in loader:\n",
    "        _, _, feats, answers, programs = batch\n",
    "        feats = feats.to(device)\n",
    "        programs = programs.to(device)\n",
    "\n",
    "        outs = tbd_net(feats, programs)\n",
    "        _, preds = outs.max(1)\n",
    "        mapped_ans = map_ans(answers)\n",
    "        num_correct += (preds.to('cpu') == mapped_ans).sum()\n",
    "        num_samples += preds.size(0)\n",
    "\n",
    "    return (num_correct, num_samples)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def write_acc(f, string, correct, total):\n",
    "    ''' Convenience function to write the accuracy to a file '''\n",
    "    percent = correct / total * 100 if total != 0 else 100\n",
    "    f.write(string.format(correct, total, percent))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Create the optimizer and loss function we will use to train our model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "optimizer = torch.optim.Adam(tbd_net.parameters(), 1e-04)\n",
    "xent_loss = torch.nn.CrossEntropyLoss().to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def train_epoch():\n",
    "    ''' Convenience function to train the model for one epoch.\n",
    "    Note that the reported training accuracy is not a true measure of accuracy after\n",
    "    training for a full epoch, as it is computed during training. However, it does give a\n",
    "    decent measure of progress.\n",
    "    '''\n",
    "    torch.set_grad_enabled(True)\n",
    "    num_correct, num_samples = 0, 0\n",
    "    for batch in train_loader:\n",
    "        _, _, feats, answers, programs = batch\n",
    "        feats = feats.to(device)\n",
    "        programs = programs.to(device)\n",
    "\n",
    "        optimizer.zero_grad()\n",
    "        outs = tbd_net(feats, programs)\n",
    "        _, preds = outs.max(1)\n",
    "        mapped_ans = map_ans(answers)\n",
    "        num_correct += (preds.to('cpu') == mapped_ans).sum()\n",
    "        num_samples += preds.size(0)\n",
    "        loss = xent_loss(outs, mapped_ans.to(device))\n",
    "    \n",
    "        # The following line applies an L1 penalty to the intermediate attention masks.\n",
    "        # This gives a signal to the model to minimize its attention outputs, which we \n",
    "        # find (nearly) removes spurious activations in background regions. Our thought\n",
    "        # process is that background regions normally get no signal to push them to zero\n",
    "        # because the model is able to effectively ignore them, so a somewhat noisy\n",
    "        # attention output is adequate. By applying this penalty, we incentivize minimizing\n",
    "        # the spurious activations, leading to better-looking outputs. For a comparison, see\n",
    "        # the visualize-outputs.ipynb notebook. If you wish to replicate the *original*\n",
    "        # results, comment the following line or change the multiplier to zero. In all \n",
    "        # other cases, we recommend keeping this factor and tuning it to your use-case if\n",
    "        # this default is not adequate.\n",
    "        loss += tbd_net.attention_sum * 2.5e-07\n",
    "        \n",
    "        loss_file.write('Loss: {}\\n'.format(loss.data[0]))\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "\n",
    "    val_correct, val_samples = check_accuracy(val_loader)\n",
    "    \n",
    "    write_acc(acc_file, 'Train Accuracy: {} / {} ({:.2f}%)\\n', num_correct, num_samples)\n",
    "    write_acc(acc_file, 'Val Accuracy: {} / {} ({:.2f}%)\\n', val_correct, val_samples)\n",
    "    acc_file.flush()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Open a loss log and an accuracy log for writing, then train for 30 epochs, saving our model after each epoch."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "starting epoch 1\n",
      "starting epoch 2\n",
      "starting epoch 3\n",
      "starting epoch 4\n",
      "starting epoch 5\n",
      "starting epoch 6\n",
      "starting epoch 7\n",
      "starting epoch 8\n",
      "starting epoch 9\n",
      "starting epoch 10\n",
      "starting epoch 11\n",
      "starting epoch 12\n",
      "starting epoch 13\n",
      "starting epoch 14\n",
      "starting epoch 15\n",
      "starting epoch 16\n",
      "starting epoch 17\n",
      "starting epoch 18\n",
      "starting epoch 19\n",
      "starting epoch 20\n",
      "starting epoch 21\n",
      "starting epoch 22\n",
      "starting epoch 23\n",
      "starting epoch 24\n",
      "starting epoch 25\n",
      "starting epoch 26\n",
      "starting epoch 27\n",
      "starting epoch 28\n",
      "starting epoch 29\n",
      "starting epoch 30\n"
     ]
    }
   ],
   "source": [
    "loss_file = open(Path('results/example-loss.txt'), 'a')\n",
    "acc_file = open(Path('results/example-acc.txt'), 'a')\n",
    "epoch = 0\n",
    "while epoch < 30:\n",
    "    save_checkpoint(epoch, 'example-{:02d}.pt'.format(epoch))\n",
    "    epoch += 1\n",
    "    print('starting epoch', epoch)\n",
    "    train_epoch()\n",
    "\n",
    "save_checkpoint(epoch, 'example-{:02d}.pt'.format(epoch))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "Now let's see how our model did!\n",
    "\n",
    "You can certainly use the following in a separate notebook to monitor progress during training, as well."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib notebook"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def plot(data, title):\n",
    "    ''' Scatterplot with a smoothed average line '''\n",
    "    fig, ax = plt.subplots()\n",
    "    x = range(len(data))\n",
    "    ax.scatter(x, data, s=5)\n",
    "    c = np.cumsum(np.insert(data, 0, 0))\n",
    "    smoothed = (c[50:] - c[:-50]) / 50\n",
    "    ax.plot(smoothed, color='r')\n",
    "    ax.set_title(title)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Read in our batch losses and plot them"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "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",
       "        this.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 overriden (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",
       "        // select the cell after this one\n",
       "        var index = IPython.notebook.find_cell_index(this.cell_info[0]);\n",
       "        IPython.notebook.select(index + 1);\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": []
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"data:image/png;base64,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\" width=\"640\">"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "loss = []\n",
    "with open(Path('results/example-loss.txt')) as f:\n",
    "    for line in f:\n",
    "        loss.append(float(line[6:]))\n",
    "loss = np.array(loss)\n",
    "\n",
    "plot(loss, 'Loss')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The above loss curve looks pretty typical for our model.\n",
    "\n",
    "We get a very quick drop in the loss as our model learns some answers are ridiculous. For example, the `query_color` module should never output an answer that isn't a color and `count` should never output an answer that's not a number.\n",
    "\n",
    "We then begin learning useful attentions, which lead to a rapid decrease in the loss, following which we see a smooth decrease toward zero."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's read in our accuracy next"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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",
       "        this.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 overriden (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",
       "        // select the cell after this one\n",
       "        var index = IPython.notebook.find_cell_index(this.cell_info[0]);\n",
       "        IPython.notebook.select(index + 1);\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": []
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"data:image/png;base64,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\" width=\"640\">"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 47.78  48.61  50.09  79.28  93.39  96.29  97.57  98.22  98.7   98.91\n",
      "  99.07  99.2   99.31  99.42  99.43  99.48  99.53  99.58  99.6   99.66\n",
      "  99.7   99.7   99.69  99.74  99.75  99.74  99.73  99.7   99.76]\n",
      "[ 48.29  49.07  52.15  91.38  94.98  97.11  97.21  97.88  98.6   98.64\n",
      "  98.53  98.79  98.86  98.81  98.97  98.92  98.82  99.05  99.02  98.88\n",
      "  98.86  98.85  98.75  98.95  98.99  98.96  98.92  99.01  99.03]\n",
      "Highest validation accuracy: 99.05\n"
     ]
    }
   ],
   "source": [
    "train_acc = []\n",
    "val_acc = []\n",
    "with open(Path('results/example-acc.txt')) as f:\n",
    "    for line in f:\n",
    "        train_acc.append(float(line[-8:-3]))\n",
    "        val_acc.append(float(f.readline()[-8:-3]))\n",
    "    \n",
    "train_acc = np.array(train_acc)\n",
    "val_acc = np.array(val_acc)\n",
    "\n",
    "x = range(len(train_acc))\n",
    "fig, ax = plt.subplots()\n",
    "ax.plot(x, train_acc, label='Train', c='g')\n",
    "ax.plot(x, val_acc, label='Val', c='g', ls=':')\n",
    "ax.legend()\n",
    "\n",
    "print(train_acc)\n",
    "print(val_acc)\n",
    "print('Highest validation accuracy: {}'.format(val_acc.max()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We see our model learns remarkably quickly. After the first couple epochs, the model is refining attentions, having learned generally what they should be doing. Again, note that the training accuracy is not quite a correct measure of training performance since it is computed during training. This is why training accuracy is slightly below validation accuracy for a while."
   ]
  }
 ],
 "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": "3.6.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}