python/notebook/training/digits/digit_model_training.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "('/lab/data/celldom/model/r0.7/digit_model',\n",
       " '/lab/data/celldom/dataset/training/digit/r0.6/validated/single',\n",
       " '/lab/data/celldom/dataset/training/digit/r0.6/manual/single')"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import os\n",
    "import re\n",
    "import glob\n",
    "import os.path as osp\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import tqdm\n",
    "import matplotlib.pyplot as plt\n",
    "from skimage import io as sk_io\n",
    "from skimage.transform import resize\n",
    "from skimage.color import gray2rgb, rgb2gray\n",
    "from skimage.exposure import rescale_intensity\n",
    "from cvutils.rectlabel import io as rectlabel_io\n",
    "from cvutils.mrcnn.session import init_keras_session\n",
    "from cvutils import visualize\n",
    "from celldom.dataset import utils as dataset_utils\n",
    "from celldom import seed\n",
    "\n",
    "import celldom\n",
    "import keras\n",
    "from keras.preprocessing import image\n",
    "import wandb\n",
    "from wandb.keras import WandbCallback\n",
    "\n",
    "init_keras_session()\n",
    "\n",
    "DATA_DIR_VALIDATED_EXTRACT = osp.join(celldom.get_dataset_dir(), 'training', 'digit', 'r0.6', 'validated', 'single')\n",
    "DATA_DIR_MANUAL_ANNOTATION = osp.join(celldom.get_dataset_dir(), 'training', 'digit', 'r0.6', 'manual', 'single')\n",
    "DATA_DIR_MANUAL_PARTITIONS = ['pt1', 'pt2', 'pt3', 'pt4']\n",
    "#MODEL_DIR = osp.join(os.environ['CELLDOM_DATA_DIR'], 'model', 'r0.6', 'digit_model')\n",
    "MODEL_DIR = osp.join(os.environ['CELLDOM_DATA_DIR'], 'model', 'r0.7', 'digit_model')\n",
    "\n",
    "MODEL_PATH_HEAD = osp.join(MODEL_DIR, 'single_digit_model_headonly.h5')\n",
    "MODEL_PATH = osp.join(MODEL_DIR, 'single_digit_model.h5')\n",
    "\n",
    "MODEL_DIR, DATA_DIR_VALIDATED_EXTRACT, DATA_DIR_MANUAL_ANNOTATION"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "BF_ST_000_APT_000Day0-APT-0-0.jpeg\n",
      "ls: write error: Broken pipe\n"
     ]
    }
   ],
   "source": [
    "!ls $DATA_DIR_VALIDATED_EXTRACT | head -n 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_actual_digit(file):\n",
    "    return osp.basename(file).replace('.jpeg', '').split('-')[-1]\n",
    "\n",
    "def load_extracted_images():\n",
    "    res = []\n",
    "    for f in glob.glob(osp.join(DATA_DIR_VALIDATED_EXTRACT, '*.jpeg')):\n",
    "        digit = get_actual_digit(f)\n",
    "        img = sk_io.imread(osp.join(DATA_DIR_VALIDATED_EXTRACT, f))\n",
    "        res.append((int(digit), img, f))\n",
    "    res = pd.DataFrame(res, columns=['digit', 'img', 'file'])\n",
    "    return res\n",
    "\n",
    "def load_annotated_images():\n",
    "    res = []\n",
    "    for pt in DATA_DIR_MANUAL_PARTITIONS:\n",
    "        df = dataset_utils.get_data_files(osp.join(DATA_DIR_MANUAL_ANNOTATION, pt))\n",
    "        for i, r in tqdm.tqdm(df.iterrows(), desc='partition=' + pt):\n",
    "            img_path, annot_path = r['image_path'], r['annot_path']\n",
    "            if r['annot_exists']:\n",
    "                try:\n",
    "                    shape, annotations = rectlabel_io.load_annotations(annot_path, assert_masks=False)\n",
    "                    assert len(annotations) == 1, 'Found multiple annotations in {}'.format(annot_path)\n",
    "                    digit = annotations[0].object_type\n",
    "                    if digit == 'NA':\n",
    "                        continue\n",
    "                    # Read RGB 8-bit image\n",
    "                    img = sk_io.imread(img_path)\n",
    "\n",
    "                    # Extract grayscale based on assumption of channel equality\n",
    "                    assert img.ndim == 3 and img.shape[-1] == 3\n",
    "                    assert np.allclose(img[..., 0], img[..., 1])\n",
    "                    assert np.allclose(img[..., 0], img[..., 2])\n",
    "                    img = img[..., 0]\n",
    "                    digit = int(digit)\n",
    "                except:\n",
    "                    print('Failure occurred for annotation file {}'.format(annot_path))\n",
    "                    raise\n",
    "                res.append((digit, img, img_path, pt))\n",
    "    res = pd.DataFrame(res, columns=['digit', 'img', 'file', 'partition'])\n",
    "    return res\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "partition=pt1: 2492it [00:02, 959.60it/s]\n",
      "partition=pt2: 456it [00:00, 1061.00it/s]\n",
      "partition=pt3: 2136it [00:02, 1032.98it/s]\n",
      "partition=pt4: 1308it [00:01, 1014.82it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 5.78 s, sys: 1.21 s, total: 6.99 s\n",
      "Wall time: 8.23 s\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "df = pd.concat([\n",
    "    load_extracted_images().assign(source='extract'),\n",
    "    load_annotated_images().assign(source='annotated')\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "9375"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# r0.2 num files = 2372, r0.6 = 8067, r0.7 = 9375\n",
    "len(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "annotated    6157\n",
       "extract      3218\n",
       "Name: source, dtype: int64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['source'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "source     partition\n",
       "annotated  pt1          2415\n",
       "           pt2           440\n",
       "           pt3          1994\n",
       "           pt4          1308\n",
       "dtype: int64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.groupby(['source', 'partition']).size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    2773\n",
       "1    2673\n",
       "3     810\n",
       "2     770\n",
       "4     605\n",
       "5     359\n",
       "9     351\n",
       "7     348\n",
       "6     345\n",
       "8     341\n",
       "Name: digit, dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['digit'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>digit</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>source</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>annotated</th>\n",
       "      <td>1285</td>\n",
       "      <td>2147</td>\n",
       "      <td>585</td>\n",
       "      <td>636</td>\n",
       "      <td>425</td>\n",
       "      <td>204</td>\n",
       "      <td>212</td>\n",
       "      <td>225</td>\n",
       "      <td>209</td>\n",
       "      <td>229</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>extract</th>\n",
       "      <td>1488</td>\n",
       "      <td>526</td>\n",
       "      <td>185</td>\n",
       "      <td>174</td>\n",
       "      <td>180</td>\n",
       "      <td>155</td>\n",
       "      <td>133</td>\n",
       "      <td>123</td>\n",
       "      <td>132</td>\n",
       "      <td>122</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "digit         0     1    2    3    4    5    6    7    8    9\n",
       "source                                                       \n",
       "annotated  1285  2147  585  636  425  204  212  225  209  229\n",
       "extract    1488   526  185  174  180  155  133  123  132  122"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.groupby(['source', 'digit']).size().unstack()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "uint8    9375\n",
       "Name: img, dtype: int64"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dtypes = df['img'].apply(lambda v: v.dtype)\n",
    "assert np.all(dtypes == np.uint8)\n",
    "dtypes.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>9375.000000</td>\n",
       "      <td>9375.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>42.103787</td>\n",
       "      <td>29.493867</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>10.503142</td>\n",
       "      <td>7.688318</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>22.000000</td>\n",
       "      <td>19.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>31.000000</td>\n",
       "      <td>22.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>42.000000</td>\n",
       "      <td>27.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>53.000000</td>\n",
       "      <td>36.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>58.000000</td>\n",
       "      <td>46.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 0            1\n",
       "count  9375.000000  9375.000000\n",
       "mean     42.103787    29.493867\n",
       "std      10.503142     7.688318\n",
       "min      22.000000    19.000000\n",
       "25%      31.000000    22.000000\n",
       "50%      42.000000    27.000000\n",
       "75%      53.000000    36.000000\n",
       "max      58.000000    46.000000"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame([r['img'].shape for _, r in df.iterrows()]).describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(7500, 1875)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "idx = np.arange(len(df))\n",
    "\n",
    "# Vary test_size (starting large) to get a sense of accuracy with a bigger sample and then decrease for final model\n",
    "idx_train, idx_test = train_test_split(idx, test_size=.2, random_state=seed, stratify=df['digit'])\n",
    "df_train, df_test = df.iloc[idx_train], df.iloc[idx_test]\n",
    "\n",
    "len(df_train), len(df_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>digit</th>\n",
       "      <th>digit</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.295867</td>\n",
       "      <td>0.295467</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.285067</td>\n",
       "      <td>0.285333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.086400</td>\n",
       "      <td>0.086400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.082133</td>\n",
       "      <td>0.082133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.064533</td>\n",
       "      <td>0.064533</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.038267</td>\n",
       "      <td>0.038400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.037467</td>\n",
       "      <td>0.037333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.037067</td>\n",
       "      <td>0.037333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.036800</td>\n",
       "      <td>0.036800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.036400</td>\n",
       "      <td>0.036267</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      digit     digit\n",
       "0  0.295867  0.295467\n",
       "1  0.285067  0.285333\n",
       "3  0.086400  0.086400\n",
       "2  0.082133  0.082133\n",
       "4  0.064533  0.064533\n",
       "5  0.038267  0.038400\n",
       "9  0.037467  0.037333\n",
       "7  0.037067  0.037333\n",
       "6  0.036800  0.036800\n",
       "8  0.036400  0.036267"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.concat([df_train['digit'].value_counts(normalize=True), df_test['digit'].value_counts(normalize=True)], axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "%run utils.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "target_shape = (32, 32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "from imgaug import augmenters as iaa\n",
    "from skimage import img_as_ubyte, img_as_float32\n",
    "\n",
    "sometimes = lambda aug: iaa.Sometimes(.2, aug)\n",
    "seq = iaa.Sequential([\n",
    "    # Simulate out-of-focus\n",
    "    sometimes(iaa.GaussianBlur(sigma=.3)),\n",
    "    \n",
    "    # Randomly alter scaling and simulate zooming\n",
    "    sometimes(iaa.OneOf([\n",
    "        iaa.CropAndPad(px=(0, 12)),\n",
    "        iaa.Crop(px=(0, 12))\n",
    "    ])),\n",
    "    \n",
    "    # Rotate up to 3 degrees and randomly scale\n",
    "    sometimes(iaa.Affine(\n",
    "        rotate=3., \n",
    "        scale={\"x\": (.8, 1.4), \"y\": (.8, 1.4)}, \n",
    "        translate_px=(0, 5),\n",
    "        mode='constant'\n",
    "    )),    \n",
    "    \n",
    "    # Randomly alter distribution\n",
    "    sometimes(iaa.OneOf([\n",
    "        iaa.Multiply((.8, 1.4)), \n",
    "        iaa.Add((-25, 25)),\n",
    "        iaa.ContrastNormalization((0.8, 1.2))\n",
    "    ]))\n",
    "], random_order=True)\n",
    "\n",
    "def augment(img):\n",
    "    # img should be provided as float in [0, 1]\n",
    "    assert img.min() >= 0 and img.max() <= 1.\n",
    "    img = rescale_intensity(img, out_range=np.uint8).astype(np.uint8)\n",
    "    res = img_as_float32(seq.augment_image(img))\n",
    "    #print(res.shape, res.dtype, res.min(), res.max())\n",
    "    return res"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(dtype('float32'), (7500, 32, 32, 3), (1875, 32, 32, 3))"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def prep_image(img):\n",
    "    # Note that gray2rgb will take care of converting from uint8 to float in 0-1\n",
    "    assert img.dtype == np.uint8\n",
    "    \n",
    "    # Convert to 2D with target height/width\n",
    "    img = gray2rgb(resize(img, target_shape, mode='constant', anti_aliasing=True)).astype(np.float32)\n",
    "    \n",
    "    # Rescale by min/max\n",
    "    img = rescale_intensity(img, out_range=(0, 1))\n",
    "    \n",
    "    assert np.all(img <= 1.) and np.all(img >= 0.)\n",
    "    return img\n",
    "\n",
    "train_x = np.stack([prep_image(v) for v in df_train['img']], 0)\n",
    "train_y = keras.utils.to_categorical(df_train['digit'].values, 10)\n",
    "\n",
    "test_x = np.stack([prep_image(v) for v in df_test['img']], 0)\n",
    "test_y = keras.utils.to_categorical(df_test['digit'].values, 10)\n",
    "\n",
    "train_generator = image.ImageDataGenerator(preprocessing_function=augment)\n",
    "train_batches = train_generator.flow(train_x, y=train_y, seed=seed)\n",
    "\n",
    "test_generator = image.ImageDataGenerator()\n",
    "test_batches = test_generator.flow(test_x, y=test_y, seed=seed)\n",
    "\n",
    "train_x.dtype, train_x.shape, test_x.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x1872 with 160 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualize augmentations\n",
    "aug_imgs = []\n",
    "for img in df_train['img'].sample(n=80):\n",
    "    aug_imgs.append(prep_image(img))\n",
    "    aug_imgs.append(augment(prep_image(img)))\n",
    "visualize.display_images(aug_imgs, cols=8, size=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DescribeResult(nobs=3072, minmax=(0.0, 1.0), mean=0.40714803, variance=0.020912908, skewness=-0.2574825882911682, kurtosis=2.4543907329458623)"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from scipy.stats import describe\n",
    "describe(aug_imgs[0].ravel())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Initialize Modeling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'/lab/data/celldom/model/r0.7/digit_model'"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "MODEL_DIR"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "single_digit_model.h5  single_digit_model_headonly.h5\n"
     ]
    }
   ],
   "source": [
    "!ls $MODEL_DIR"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Clear everything currently in the modeling directory\n",
    "!rm -rf $MODEL_DIR/*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "W&B Run: https://app.wandb.ai/hammerlab/celldom/runs/c8ywa03w\n",
      "Wrap your training loop with `with wandb.monitor():` to display live results.\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<iframe src=\"https://app.wandb.ai/hammerlab/celldom/runs/c8ywa03w/edit?jupyter=true\" style=\"border:none;width:100%;height:300px\">\n",
       "            </iframe>"
      ],
      "text/plain": [
       "W&B Run https://app.wandb.ai/hammerlab/celldom/runs/c8ywa03w"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "wandb.init()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Run Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'/lab/data/celldom/model/r0.7/digit_model'"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "MODEL_DIR"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "if not osp.exists(MODEL_DIR):\n",
    "    os.makedirs(MODEL_DIR)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "!ls $MODEL_DIR"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = get_digit_model(input_shape=target_shape + (3,), mode='head')\n",
    "    \n",
    "model.compile(\n",
    "    loss=keras.losses.categorical_crossentropy,\n",
    "    optimizer=keras.optimizers.RMSprop(lr=.001),\n",
    "    metrics=['accuracy']\n",
    ")\n",
    "\n",
    "# Initialize from pre-trained model\n",
    "model.load_weights('/lab/repos/svhn/weights.hdf5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Resuming run: https://app.wandb.ai/hammerlab/celldom/runs/c8ywa03w\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<iframe src=\"https://app.wandb.ai/hammerlab/celldom/runs/c8ywa03w?jupyter=true\" style=\"border:none;width:100%;height:420px\">\n",
       "        </iframe>"
      ],
      "text/plain": [
       "<wandb.jupyter.Run at 0x7f8298864c50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/250\n",
      "235/235 [==============================] - 11s 48ms/step - loss: 0.5663 - acc: 0.8475 - val_loss: 0.0773 - val_acc: 0.9776\n",
      "Epoch 2/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.2982 - acc: 0.9145 - val_loss: 0.0458 - val_acc: 0.9824\n",
      "Epoch 3/250\n",
      "235/235 [==============================] - 10s 44ms/step - loss: 0.2578 - acc: 0.9325 - val_loss: 0.0311 - val_acc: 0.9893\n",
      "Epoch 4/250\n",
      "235/235 [==============================] - 10s 43ms/step - loss: 0.2344 - acc: 0.9364 - val_loss: 0.0260 - val_acc: 0.9936\n",
      "Epoch 5/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.1992 - acc: 0.9465 - val_loss: 0.0219 - val_acc: 0.9952\n",
      "Epoch 6/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.1910 - acc: 0.9501 - val_loss: 0.0189 - val_acc: 0.9941\n",
      "Epoch 7/250\n",
      "235/235 [==============================] - 11s 45ms/step - loss: 0.1745 - acc: 0.9560 - val_loss: 0.0173 - val_acc: 0.9947\n",
      "Epoch 8/250\n",
      "235/235 [==============================] - 10s 41ms/step - loss: 0.1674 - acc: 0.9547 - val_loss: 0.0171 - val_acc: 0.9952\n",
      "Epoch 9/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.1501 - acc: 0.9564 - val_loss: 0.0180 - val_acc: 0.9936\n",
      "Epoch 10/250\n",
      "235/235 [==============================] - 11s 46ms/step - loss: 0.1555 - acc: 0.9576 - val_loss: 0.0186 - val_acc: 0.9936\n",
      "Epoch 11/250\n",
      "235/235 [==============================] - 9s 40ms/step - loss: 0.1717 - acc: 0.9547 - val_loss: 0.0141 - val_acc: 0.9963\n",
      "Epoch 12/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.1645 - acc: 0.9618 - val_loss: 0.0138 - val_acc: 0.9957\n",
      "Epoch 13/250\n",
      "235/235 [==============================] - 10s 41ms/step - loss: 0.1591 - acc: 0.9617 - val_loss: 0.0132 - val_acc: 0.9963\n",
      "Epoch 14/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.1551 - acc: 0.9593 - val_loss: 0.0127 - val_acc: 0.9957\n",
      "Epoch 15/250\n",
      "235/235 [==============================] - 10s 41ms/step - loss: 0.1440 - acc: 0.9614 - val_loss: 0.0116 - val_acc: 0.9963\n",
      "Epoch 16/250\n",
      "235/235 [==============================] - 9s 40ms/step - loss: 0.1428 - acc: 0.9629 - val_loss: 0.0094 - val_acc: 0.9973\n",
      "Epoch 17/250\n",
      "235/235 [==============================] - 10s 41ms/step - loss: 0.1316 - acc: 0.9647 - val_loss: 0.0137 - val_acc: 0.9952\n",
      "Epoch 18/250\n",
      "235/235 [==============================] - 10s 41ms/step - loss: 0.1342 - acc: 0.9666 - val_loss: 0.0098 - val_acc: 0.9952\n",
      "Epoch 19/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.1342 - acc: 0.9644 - val_loss: 0.0092 - val_acc: 0.9963\n",
      "Epoch 20/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.1423 - acc: 0.9638 - val_loss: 0.0109 - val_acc: 0.9963\n",
      "Epoch 21/250\n",
      "235/235 [==============================] - 10s 41ms/step - loss: 0.1329 - acc: 0.9669 - val_loss: 0.0102 - val_acc: 0.9957\n",
      "Epoch 22/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.1344 - acc: 0.9669 - val_loss: 0.0096 - val_acc: 0.9963\n",
      "\n",
      "Epoch 00022: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
      "Epoch 23/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.1191 - acc: 0.9702 - val_loss: 0.0093 - val_acc: 0.9963\n",
      "Epoch 24/250\n",
      "235/235 [==============================] - 10s 43ms/step - loss: 0.1307 - acc: 0.9682 - val_loss: 0.0090 - val_acc: 0.9963\n",
      "Epoch 25/250\n",
      "235/235 [==============================] - 10s 41ms/step - loss: 0.1235 - acc: 0.9665 - val_loss: 0.0086 - val_acc: 0.9963\n",
      "Epoch 26/250\n",
      "235/235 [==============================] - 10s 43ms/step - loss: 0.1272 - acc: 0.9680 - val_loss: 0.0089 - val_acc: 0.9963\n",
      "Epoch 27/250\n",
      "235/235 [==============================] - 10s 41ms/step - loss: 0.1247 - acc: 0.9667 - val_loss: 0.0092 - val_acc: 0.9963\n",
      "Epoch 28/250\n",
      "235/235 [==============================] - 11s 45ms/step - loss: 0.1152 - acc: 0.9711 - val_loss: 0.0087 - val_acc: 0.9963\n",
      "\n",
      "Epoch 00028: ReduceLROnPlateau reducing learning rate to 1.0000000474974514e-05.\n",
      "Epoch 29/250\n",
      "235/235 [==============================] - 10s 44ms/step - loss: 0.1154 - acc: 0.9687 - val_loss: 0.0086 - val_acc: 0.9963\n",
      "Epoch 30/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.1219 - acc: 0.9677 - val_loss: 0.0086 - val_acc: 0.9963\n",
      "Epoch 31/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.1250 - acc: 0.9660 - val_loss: 0.0086 - val_acc: 0.9963\n",
      "\n",
      "Epoch 00031: ReduceLROnPlateau reducing learning rate to 1.0000000656873453e-06.\n",
      "Epoch 32/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.1142 - acc: 0.9711 - val_loss: 0.0086 - val_acc: 0.9963\n",
      "Epoch 33/250\n",
      "235/235 [==============================] - 10s 43ms/step - loss: 0.1557 - acc: 0.9628 - val_loss: 0.0086 - val_acc: 0.9963\n",
      "Epoch 34/250\n",
      "235/235 [==============================] - 10s 43ms/step - loss: 0.1252 - acc: 0.9677 - val_loss: 0.0086 - val_acc: 0.9963\n",
      "\n",
      "Epoch 00034: ReduceLROnPlateau reducing learning rate to 1.0000001111620805e-07.\n",
      "Epoch 35/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.1048 - acc: 0.9703 - val_loss: 0.0086 - val_acc: 0.9963\n",
      "Epoch 36/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.1204 - acc: 0.9687 - val_loss: 0.0086 - val_acc: 0.9963\n",
      "Epoch 37/250\n",
      "235/235 [==============================] - 10s 44ms/step - loss: 0.1118 - acc: 0.9705 - val_loss: 0.0086 - val_acc: 0.9963\n",
      "\n",
      "Epoch 00037: ReduceLROnPlateau reducing learning rate to 1.000000082740371e-08.\n",
      "Epoch 38/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.1232 - acc: 0.9677 - val_loss: 0.0086 - val_acc: 0.9963\n",
      "Epoch 39/250\n",
      "235/235 [==============================] - 10s 43ms/step - loss: 0.1211 - acc: 0.9682 - val_loss: 0.0086 - val_acc: 0.9963\n"
     ]
    }
   ],
   "source": [
    "save_model = keras.callbacks.ModelCheckpoint(MODEL_PATH_HEAD, monitor='val_loss', mode='min', verbose=0,\n",
    "                                             save_best_only=True, save_weights_only=False, period=1)\n",
    "early_stopping = keras.callbacks.EarlyStopping(monitor='val_loss', min_delta=0, patience=10, verbose=0, mode='min')\n",
    "learning_rate = keras.callbacks.ReduceLROnPlateau(patience=3, monitor='val_loss', mode='min', verbose=1)\n",
    "callbacks = [save_model, early_stopping, learning_rate]\n",
    "\n",
    "# tensorboard = keras.callbacks.TensorBoard(log_dir=MODEL_DIR)\n",
    "# callbacks = [save_model, early_stopping, learning_rate, tensorboard]\n",
    "\n",
    "with wandb.monitor(display=False):\n",
    "    model.fit_generator(\n",
    "        train_batches, \n",
    "        epochs=250,\n",
    "        callbacks=callbacks + [WandbCallback(save_model=False)],\n",
    "        validation_data=test_batches\n",
    "    )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Fine-Tuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "#model = get_digit_model(input_shape=target_shape + (3,), mode='tune')\n",
    "model = get_digit_model(input_shape=target_shape + (3,), mode='all')\n",
    "    \n",
    "model.compile(\n",
    "    loss=keras.losses.categorical_crossentropy,\n",
    "    optimizer=keras.optimizers.RMSprop(lr=.0001),\n",
    "    metrics=['accuracy']\n",
    ")\n",
    "\n",
    "# Initialize from pre-trained model\n",
    "model.load_weights(MODEL_PATH_HEAD)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Resuming run: https://app.wandb.ai/hammerlab/celldom/runs/c8ywa03w\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<iframe src=\"https://app.wandb.ai/hammerlab/celldom/runs/c8ywa03w?jupyter=true\" style=\"border:none;width:100%;height:420px\">\n",
       "        </iframe>"
      ],
      "text/plain": [
       "<wandb.jupyter.Run at 0x7f8298918ef0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/250\n",
      "235/235 [==============================] - 11s 46ms/step - loss: 0.1367 - acc: 0.9661 - val_loss: 0.0012 - val_acc: 1.0000\n",
      "Epoch 2/250\n",
      "235/235 [==============================] - 10s 41ms/step - loss: 0.0973 - acc: 0.9730 - val_loss: 8.5180e-04 - val_acc: 1.0000\n",
      "Epoch 3/250\n",
      "235/235 [==============================] - 11s 45ms/step - loss: 0.1033 - acc: 0.9738 - val_loss: 7.9742e-04 - val_acc: 1.0000\n",
      "Epoch 4/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.0935 - acc: 0.9730 - val_loss: 0.0012 - val_acc: 1.0000\n",
      "Epoch 5/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.1235 - acc: 0.9688 - val_loss: 0.0014 - val_acc: 0.9989\n",
      "\n",
      "Epoch 00005: ReduceLROnPlateau reducing learning rate to 9.999999747378752e-06.\n",
      "Epoch 6/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.0893 - acc: 0.9769 - val_loss: 0.0012 - val_acc: 0.9995\n",
      "Epoch 7/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.0918 - acc: 0.9754 - val_loss: 0.0013 - val_acc: 0.9995\n",
      "Epoch 8/250\n",
      "235/235 [==============================] - 10s 43ms/step - loss: 0.0792 - acc: 0.9783 - val_loss: 0.0011 - val_acc: 0.9995\n",
      "\n",
      "Epoch 00008: ReduceLROnPlateau reducing learning rate to 9.999999747378752e-07.\n",
      "Epoch 9/250\n",
      "235/235 [==============================] - 10s 41ms/step - loss: 0.1157 - acc: 0.9734 - val_loss: 0.0012 - val_acc: 0.9995\n",
      "Epoch 10/250\n",
      "235/235 [==============================] - 10s 43ms/step - loss: 0.0881 - acc: 0.9775 - val_loss: 0.0012 - val_acc: 0.9995\n",
      "Epoch 11/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.0936 - acc: 0.9757 - val_loss: 0.0011 - val_acc: 0.9995\n",
      "\n",
      "Epoch 00011: ReduceLROnPlateau reducing learning rate to 9.999999974752428e-08.\n",
      "Epoch 12/250\n",
      "235/235 [==============================] - 10s 42ms/step - loss: 0.0976 - acc: 0.9730 - val_loss: 0.0012 - val_acc: 0.9995\n",
      "Epoch 13/250\n",
      "235/235 [==============================] - 10s 44ms/step - loss: 0.0925 - acc: 0.9771 - val_loss: 0.0012 - val_acc: 0.9995\n"
     ]
    }
   ],
   "source": [
    "save_model = keras.callbacks.ModelCheckpoint(MODEL_PATH, monitor='val_loss', mode='min', verbose=0,\n",
    "                                             save_best_only=True, save_weights_only=False, period=1)\n",
    "early_stopping = keras.callbacks.EarlyStopping(monitor='val_loss', min_delta=0, patience=10, mode='min')\n",
    "learning_rate = keras.callbacks.ReduceLROnPlateau(patience=3, monitor='val_loss', mode='min', verbose=1)\n",
    "callbacks = [save_model, early_stopping, learning_rate]\n",
    "\n",
    "# tensorboard = keras.callbacks.TensorBoard(log_dir=MODEL_DIR)\n",
    "# callbacks = [save_model, early_stopping, learning_rate, tensorboard]\n",
    "\n",
    "with wandb.monitor(display=False):\n",
    "    model.fit_generator(\n",
    "        train_batches, \n",
    "        epochs=250,\n",
    "        callbacks=callbacks + [WandbCallback(save_model=False)],\n",
    "        validation_data=test_batches\n",
    "    )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Inference "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = keras.models.load_model(MODEL_PATH)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((7500, 32, 32, 3), (7500,))"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#inf_x, inf_y = test_x, test_y\n",
    "inf_x, inf_y = train_x, train_y\n",
    "inf_y = np.argmax(inf_y, axis=1)\n",
    "inf_x.shape, inf_y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(7500,)"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "inf_y_pred = model.predict_classes(inf_x)\n",
    "inf_y_pred.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9998666666666667"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Overall accuracy\n",
    "(inf_y_pred == inf_y).mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row0_col0 {\n",
       "            background-color:  #4897c4;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row0_col1 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row0_col2 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row0_col3 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row0_col4 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row0_col5 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row0_col6 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row0_col7 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row0_col8 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row0_col9 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row1_col0 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row1_col1 {\n",
       "            background-color:  #4897c4;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row1_col2 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row1_col3 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row1_col4 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row1_col5 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row1_col6 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row1_col7 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row1_col8 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row1_col9 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row2_col0 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row2_col1 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row2_col2 {\n",
       "            background-color:  #4897c4;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row2_col3 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row2_col4 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row2_col5 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row2_col6 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row2_col7 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row2_col8 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row2_col9 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row3_col0 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row3_col1 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row3_col2 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row3_col3 {\n",
       "            background-color:  #4897c4;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row3_col4 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row3_col5 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row3_col6 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row3_col7 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row3_col8 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row3_col9 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row4_col0 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row4_col1 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row4_col2 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row4_col3 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row4_col4 {\n",
       "            background-color:  #4897c4;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row4_col5 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row4_col6 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row4_col7 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row4_col8 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row4_col9 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row5_col0 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row5_col1 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row5_col2 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row5_col3 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row5_col4 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row5_col5 {\n",
       "            background-color:  #4897c4;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row5_col6 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row5_col7 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row5_col8 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row5_col9 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row6_col0 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row6_col1 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row6_col2 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row6_col3 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row6_col4 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row6_col5 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row6_col6 {\n",
       "            background-color:  #4897c4;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row6_col7 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row6_col8 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row6_col9 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row7_col0 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row7_col1 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row7_col2 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row7_col3 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row7_col4 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row7_col5 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row7_col6 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row7_col7 {\n",
       "            background-color:  #4897c4;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row7_col8 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row7_col9 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row8_col0 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row8_col1 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row8_col2 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row8_col3 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row8_col4 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row8_col5 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row8_col6 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row8_col7 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row8_col8 {\n",
       "            background-color:  #4897c4;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row8_col9 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row9_col0 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row9_col1 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row9_col2 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row9_col3 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row9_col4 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row9_col5 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row9_col6 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row9_col7 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row9_col8 {\n",
       "            background-color:  #fff7fb;\n",
       "        }    #T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row9_col9 {\n",
       "            background-color:  #4897c4;\n",
       "        }</style>  \n",
       "<table id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002\" > \n",
       "<thead>    <tr> \n",
       "        <th class=\"index_name level0\" >true</th> \n",
       "        <th class=\"col_heading level0 col0\" >0</th> \n",
       "        <th class=\"col_heading level0 col1\" >1</th> \n",
       "        <th class=\"col_heading level0 col2\" >2</th> \n",
       "        <th class=\"col_heading level0 col3\" >3</th> \n",
       "        <th class=\"col_heading level0 col4\" >4</th> \n",
       "        <th class=\"col_heading level0 col5\" >5</th> \n",
       "        <th class=\"col_heading level0 col6\" >6</th> \n",
       "        <th class=\"col_heading level0 col7\" >7</th> \n",
       "        <th class=\"col_heading level0 col8\" >8</th> \n",
       "        <th class=\"col_heading level0 col9\" >9</th> \n",
       "    </tr>    <tr> \n",
       "        <th class=\"index_name level0\" >pred</th> \n",
       "        <th class=\"blank\" ></th> \n",
       "        <th class=\"blank\" ></th> \n",
       "        <th class=\"blank\" ></th> \n",
       "        <th class=\"blank\" ></th> \n",
       "        <th class=\"blank\" ></th> \n",
       "        <th class=\"blank\" ></th> \n",
       "        <th class=\"blank\" ></th> \n",
       "        <th class=\"blank\" ></th> \n",
       "        <th class=\"blank\" ></th> \n",
       "        <th class=\"blank\" ></th> \n",
       "    </tr></thead> \n",
       "<tbody>    <tr> \n",
       "        <th id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002level0_row0\" class=\"row_heading level0 row0\" >0</th> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row0_col0\" class=\"data row0 col0\" >2219</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row0_col1\" class=\"data row0 col1\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row0_col2\" class=\"data row0 col2\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row0_col3\" class=\"data row0 col3\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row0_col4\" class=\"data row0 col4\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row0_col5\" class=\"data row0 col5\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row0_col6\" class=\"data row0 col6\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row0_col7\" class=\"data row0 col7\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row0_col8\" class=\"data row0 col8\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row0_col9\" class=\"data row0 col9\" >0</td> \n",
       "    </tr>    <tr> \n",
       "        <th id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002level0_row1\" class=\"row_heading level0 row1\" >1</th> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row1_col0\" class=\"data row1 col0\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row1_col1\" class=\"data row1 col1\" >2137</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row1_col2\" class=\"data row1 col2\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row1_col3\" class=\"data row1 col3\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row1_col4\" class=\"data row1 col4\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row1_col5\" class=\"data row1 col5\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row1_col6\" class=\"data row1 col6\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row1_col7\" class=\"data row1 col7\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row1_col8\" class=\"data row1 col8\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row1_col9\" class=\"data row1 col9\" >0</td> \n",
       "    </tr>    <tr> \n",
       "        <th id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002level0_row2\" class=\"row_heading level0 row2\" >2</th> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row2_col0\" class=\"data row2 col0\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row2_col1\" class=\"data row2 col1\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row2_col2\" class=\"data row2 col2\" >616</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row2_col3\" class=\"data row2 col3\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row2_col4\" class=\"data row2 col4\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row2_col5\" class=\"data row2 col5\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row2_col6\" class=\"data row2 col6\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row2_col7\" class=\"data row2 col7\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row2_col8\" class=\"data row2 col8\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row2_col9\" class=\"data row2 col9\" >0</td> \n",
       "    </tr>    <tr> \n",
       "        <th id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002level0_row3\" class=\"row_heading level0 row3\" >3</th> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row3_col0\" class=\"data row3 col0\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row3_col1\" class=\"data row3 col1\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row3_col2\" class=\"data row3 col2\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row3_col3\" class=\"data row3 col3\" >648</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row3_col4\" class=\"data row3 col4\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row3_col5\" class=\"data row3 col5\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row3_col6\" class=\"data row3 col6\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row3_col7\" class=\"data row3 col7\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row3_col8\" class=\"data row3 col8\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row3_col9\" class=\"data row3 col9\" >0</td> \n",
       "    </tr>    <tr> \n",
       "        <th id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002level0_row4\" class=\"row_heading level0 row4\" >4</th> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row4_col0\" class=\"data row4 col0\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row4_col1\" class=\"data row4 col1\" >1</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row4_col2\" class=\"data row4 col2\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row4_col3\" class=\"data row4 col3\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row4_col4\" class=\"data row4 col4\" >484</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row4_col5\" class=\"data row4 col5\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row4_col6\" class=\"data row4 col6\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row4_col7\" class=\"data row4 col7\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row4_col8\" class=\"data row4 col8\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row4_col9\" class=\"data row4 col9\" >0</td> \n",
       "    </tr>    <tr> \n",
       "        <th id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002level0_row5\" class=\"row_heading level0 row5\" >5</th> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row5_col0\" class=\"data row5 col0\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row5_col1\" class=\"data row5 col1\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row5_col2\" class=\"data row5 col2\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row5_col3\" class=\"data row5 col3\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row5_col4\" class=\"data row5 col4\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row5_col5\" class=\"data row5 col5\" >287</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row5_col6\" class=\"data row5 col6\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row5_col7\" class=\"data row5 col7\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row5_col8\" class=\"data row5 col8\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row5_col9\" class=\"data row5 col9\" >0</td> \n",
       "    </tr>    <tr> \n",
       "        <th id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002level0_row6\" class=\"row_heading level0 row6\" >6</th> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row6_col0\" class=\"data row6 col0\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row6_col1\" class=\"data row6 col1\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row6_col2\" class=\"data row6 col2\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row6_col3\" class=\"data row6 col3\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row6_col4\" class=\"data row6 col4\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row6_col5\" class=\"data row6 col5\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row6_col6\" class=\"data row6 col6\" >276</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row6_col7\" class=\"data row6 col7\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row6_col8\" class=\"data row6 col8\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row6_col9\" class=\"data row6 col9\" >0</td> \n",
       "    </tr>    <tr> \n",
       "        <th id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002level0_row7\" class=\"row_heading level0 row7\" >7</th> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row7_col0\" class=\"data row7 col0\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row7_col1\" class=\"data row7 col1\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row7_col2\" class=\"data row7 col2\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row7_col3\" class=\"data row7 col3\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row7_col4\" class=\"data row7 col4\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row7_col5\" class=\"data row7 col5\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row7_col6\" class=\"data row7 col6\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row7_col7\" class=\"data row7 col7\" >278</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row7_col8\" class=\"data row7 col8\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row7_col9\" class=\"data row7 col9\" >0</td> \n",
       "    </tr>    <tr> \n",
       "        <th id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002level0_row8\" class=\"row_heading level0 row8\" >8</th> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row8_col0\" class=\"data row8 col0\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row8_col1\" class=\"data row8 col1\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row8_col2\" class=\"data row8 col2\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row8_col3\" class=\"data row8 col3\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row8_col4\" class=\"data row8 col4\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row8_col5\" class=\"data row8 col5\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row8_col6\" class=\"data row8 col6\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row8_col7\" class=\"data row8 col7\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row8_col8\" class=\"data row8 col8\" >273</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row8_col9\" class=\"data row8 col9\" >0</td> \n",
       "    </tr>    <tr> \n",
       "        <th id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002level0_row9\" class=\"row_heading level0 row9\" >9</th> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row9_col0\" class=\"data row9 col0\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row9_col1\" class=\"data row9 col1\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row9_col2\" class=\"data row9 col2\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row9_col3\" class=\"data row9 col3\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row9_col4\" class=\"data row9 col4\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row9_col5\" class=\"data row9 col5\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row9_col6\" class=\"data row9 col6\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row9_col7\" class=\"data row9 col7\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row9_col8\" class=\"data row9 col8\" >0</td> \n",
       "        <td id=\"T_bd3e6c34_c7ed_11e8_b83a_0242ac110002row9_col9\" class=\"data row9 col9\" >281</td> \n",
       "    </tr></tbody> \n",
       "</table> "
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x7f8233236ba8>"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(dict(pred=inf_y_pred, true=inf_y)).groupby(['pred', 'true']).size().unstack().fillna(0)\\\n",
    "    .style.background_gradient(high=.7)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Show Misclassified Instances (if any)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "misclassified = np.argwhere(inf_y_pred != inf_y).ravel()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(misclassified)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([6988])"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "misclassified"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#n_samp = 10\n",
    "#indexes = pd.Series(list(range(len(pred_y_class)))).sample(n_samp)\n",
    "\n",
    "n_samp = len(misclassified)\n",
    "indexes = misclassified\n",
    "\n",
    "fig, ax = plt.subplots(n_samp, 1)\n",
    "if n_samp == 1:\n",
    "    ax = [ax]\n",
    "fig.set_size_inches(4, n_samp * 4)\n",
    "for i, idx in enumerate(indexes):\n",
    "    ax[i].imshow(inf_x[idx])\n",
    "    ax[i].set_title('Pred: {} - True: {} (idx {})'.format(inf_y_pred[idx], inf_y[idx], idx))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'/lab/data/celldom/dataset/training/digit/r0.6/manual/single/pt3/chip_01-9afff6da4531832f02d8bd6b3e33a9d1-4-st-0-0.png'"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Use the \"idx\" above to get the file the example came from\n",
    "df.iloc[idx_train[5931]]['file']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'/lab/data/celldom/model/r0.7/digit_model/single_digit_model.h5'"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "MODEL_PATH"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Overwrite model in cache to avoid needing to redownload from google storage\n",
    "!cp /lab/data/celldom/model/r0.6/digit_model/single_digit_model.h5 /lab/data/celldom/.celldom/cache/models/single_digit_model.h5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Upload to GS\n",
    "# gsutil cp /data/disk2/celldom/model/r0.7/digit_model/single_digit_model.h5 gs://celldom/models/r0.7/single_digit_model.h5"
   ]
  }
 ],
 "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}