examples/MNIST_easytorch_CNN.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "!pip install easytorch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "from easytorch import EasyTorch, ETRunner, ConfusionMatrix, ETMeter, AUCROCMetrics\n",
    "from torchvision import datasets, transforms\n",
    "from torch import nn\n",
    "import torch.nn.functional as F\n",
    "import torch\n",
    "from IPython.display import Image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "transform = transforms.Compose([\n",
    "    transforms.ToTensor(),\n",
    "    transforms.Normalize((0.1307,), (0.3081,))\n",
    "])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Define neural network. I just burrowed from here: https://github.com/pytorch/examples/blob/master/mnist/main.py**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "class MNISTNet(nn.Module):\n",
    "    def __init__(self):\n",
    "        super(MNISTNet, self).__init__()\n",
    "        self.conv1 = nn.Conv2d(1, 32, 3, 1)\n",
    "        self.conv2 = nn.Conv2d(32, 64, 3, 1)\n",
    "        self.dropout1 = nn.Dropout(0.25)\n",
    "        self.dropout2 = nn.Dropout(0.5)\n",
    "        self.fc1 = nn.Linear(9216, 128)\n",
    "        self.fc2 = nn.Linear(128, 10)\n",
    "\n",
    "    def forward(self, x):\n",
    "        x = self.conv1(x)\n",
    "        x = F.relu(x)\n",
    "        x = self.conv2(x)\n",
    "        x = F.relu(x)\n",
    "        x = F.max_pool2d(x, 2)\n",
    "        x = self.dropout1(x)\n",
    "        x = torch.flatten(x, 1)\n",
    "        x = self.fc1(x)\n",
    "        x = F.relu(x)\n",
    "        x = self.dropout2(x)\n",
    "        x = self.fc2(x)\n",
    "        output = F.log_softmax(x, dim=1)\n",
    "        return output"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "class MNISTTrainer(ETRunner):\n",
    "    def _init_nn_model(self):\n",
    "        self.nn['model'] = MNISTNet()\n",
    "\n",
    "    def iteration(self, batch):\n",
    "        inputs = batch[0].to(self.device['gpu']).float()\n",
    "        labels = batch[1].to(self.device['gpu']).long()\n",
    "\n",
    "        out = self.nn['model'](inputs)\n",
    "        loss = F.nll_loss(out, labels)\n",
    "\n",
    "        max_prob, pred = torch.max(out, 1)\n",
    "\n",
    "        meter = self.new_meter()\n",
    "        meter.averages.add(loss.item(), len(inputs))\n",
    "\n",
    "        # Add extra loss if needed as:\n",
    "        meter.averages.add(loss.item() * 0.3, len(inputs), index=1)\n",
    "\n",
    "        meter.metrics['cmf'].add(pred, labels.float())\n",
    "        meter.metrics['auc'].add(max_prob, labels.float())\n",
    "\n",
    "        return {'loss': loss, 'meter': meter, 'predictions': pred}\n",
    "\n",
    "    def init_cache(self):\n",
    "        self.cache['log_header'] = 'Loss1,Loss2|Accuracy,F1,Precision,Recall'\n",
    "        self.cache.update(monitor_metric='f1', metric_direction='maximize')\n",
    "\n",
    "    def new_meter(self):\n",
    "        return ETMeter(\n",
    "            num_averages=2,  # Since we are tracing two losses\n",
    "            cmf=ConfusionMatrix(num_classes=10),\n",
    "            auc=AUCROCMetrics(),\n",
    "            device=self.device['gpu']\n",
    "        )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_dataset = datasets.MNIST('../data', train=True, download=True,\n",
    "                               transform=transform)\n",
    "val_dataset = datasets.MNIST('../data', train=False,\n",
    "                             transform=transform)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[WARNING]! 1 GPU(s) requested but GPU(s) not detected. Using CPU(Much slower).\n",
      "Starting with the following parameters:\n",
      "{'batch_size': 128,\n",
      " 'dataset_dir': '',\n",
      " 'epochs': 5,\n",
      " 'force': False,\n",
      " 'gpus': [],\n",
      " 'grad_accum_iters': 1,\n",
      " 'is_master': True,\n",
      " 'learning_rate': 0.001,\n",
      " 'load_limit': None,\n",
      " 'load_sparse': False,\n",
      " 'log_dir': 'net_logs',\n",
      " 'num_folds': None,\n",
      " 'num_workers': 0,\n",
      " 'patience': None,\n",
      " 'phase': 'train',\n",
      " 'pin_memory': False,\n",
      " 'pretrained_path': None,\n",
      " 'seed': 1471585,\n",
      " 'seed_all': False,\n",
      " 'split_ratio': None,\n",
      " 'use_ddp': False,\n",
      " 'verbose': True}\n",
      "1 split(s) loaded from 'net_logs/experiment/splits' directory.\n",
      "[SUCCESS]! Total params in model: 1199882\n",
      "Training ...\n",
      "Ep:1/5,Itr:1/469, Averages:[4.73507], Metrics:[0.14062, 0.1272, 0.14497, 0.11331]\n",
      "Ep:1/5,Itr:2/469, Averages:[13.15077], Metrics:[0.07812, 0.04356, 0.10707, 0.02734]\n",
      "Ep:1/5,Itr:3/469, Averages:[7.89299], Metrics:[0.19531, 0.19356, 0.17931, 0.21027]\n",
      "Ep:1/5,Itr:4/469, Averages:[5.66375], Metrics:[0.21094, 0.13937, 0.19537, 0.10832]\n",
      "Ep:1/5,Itr:5/469, Averages:[3.47898], Metrics:[0.32031, 0.28857, 0.3144, 0.26667]\n",
      "Ep:1/5,Itr:6/469, Averages:[2.71436], Metrics:[0.25781, 0.23197, 0.25333, 0.21393]\n",
      "Ep:1/5,Itr:7/469, Averages:[2.33873], Metrics:[0.28125, 0.2499, 0.26376, 0.23742]\n",
      "Ep:1/5,Itr:8/469, Averages:[2.00967], Metrics:[0.27344, 0.27112, 0.28937, 0.25503]\n",
      "Ep:1/5,Itr:9/469, Averages:[1.73309], Metrics:[0.42188, 0.39196, 0.41856, 0.36853]\n",
      "Ep:1/5,Itr:10/469, Averages:[1.64717], Metrics:[0.41406, 0.31482, 0.33653, 0.29574]\n",
      "Ep:1/5,Itr:11/469, Averages:[1.69172], Metrics:[0.39062, 0.40167, 0.35297, 0.46595]\n",
      "Ep:1/5,Itr:12/469, Averages:[1.68626], Metrics:[0.41406, 0.412, 0.40735, 0.41676]\n",
      "Ep:1/5,Itr:13/469, Averages:[1.65012], Metrics:[0.42188, 0.35524, 0.4147, 0.31069]\n",
      "Ep:1/5,Itr:14/469, Averages:[1.54865], Metrics:[0.46094, 0.41839, 0.41977, 0.41701]\n",
      "Ep:1/5,Itr:15/469, Averages:[1.56252], Metrics:[0.42188, 0.4039, 0.43054, 0.38036]\n",
      "Ep:1/5,Itr:16/469, Averages:[1.53515], Metrics:[0.44531, 0.43564, 0.42224, 0.44992]\n",
      "Ep:1/5,Itr:17/469, Averages:[1.62652], Metrics:[0.41406, 0.40735, 0.42091, 0.39464]\n",
      "Ep:1/5,Itr:18/469, Averages:[1.52897], Metrics:[0.44531, 0.41448, 0.44065, 0.39123]\n",
      "Ep:1/5,Itr:19/469, Averages:[1.41165], Metrics:[0.49219, 0.48519, 0.47476, 0.49607]\n",
      "Ep:1/5,Itr:20/469, Averages:[1.40776], Metrics:[0.48438, 0.4444, 0.45033, 0.43863]\n",
      "Ep:1/5,Itr:21/469, Averages:[1.31515], Metrics:[0.51562, 0.55197, 0.50503, 0.60854]\n",
      "Ep:1/5,Itr:22/469, Averages:[1.46083], Metrics:[0.51562, 0.56102, 0.50971, 0.62383]\n",
      "Ep:1/5,Itr:23/469, Averages:[1.33929], Metrics:[0.49219, 0.54063, 0.51686, 0.56669]\n",
      "Ep:1/5,Itr:24/469, Averages:[1.16072], Metrics:[0.57031, 0.47684, 0.49958, 0.45609]\n",
      "Ep:1/5,Itr:25/469, Averages:[1.28123], Metrics:[0.54688, 0.47882, 0.49413, 0.46443]\n",
      "Ep:1/5,Itr:26/469, Averages:[1.17193], Metrics:[0.5625, 0.54727, 0.50271, 0.60051]\n",
      "Ep:1/5,Itr:27/469, Averages:[1.19173], Metrics:[0.57812, 0.62107, 0.5634, 0.6919]\n",
      "Ep:1/5,Itr:28/469, Averages:[1.31499], Metrics:[0.52344, 0.58256, 0.54981, 0.61947]\n",
      "Ep:1/5,Itr:29/469, Averages:[1.27121], Metrics:[0.5625, 0.6038, 0.57392, 0.63697]\n",
      "Ep:1/5,Itr:30/469, Averages:[1.14091], Metrics:[0.60156, 0.59518, 0.55917, 0.63616]\n",
      "Ep:1/5,Itr:31/469, Averages:[1.10366], Metrics:[0.57031, 0.59447, 0.58098, 0.60859]\n",
      "Ep:1/5,Itr:32/469, Averages:[1.25225], Metrics:[0.53125, 0.56486, 0.54126, 0.59061]\n",
      "Ep:1/5,Itr:33/469, Averages:[1.25089], Metrics:[0.55469, 0.55556, 0.53891, 0.57328]\n",
      "Ep:1/5,Itr:34/469, Averages:[1.0833], Metrics:[0.57812, 0.5613, 0.55792, 0.56472]\n",
      "Ep:1/5,Itr:35/469, Averages:[1.17292], Metrics:[0.59375, 0.59201, 0.57376, 0.61146]\n",
      "Ep:1/5,Itr:36/469, Averages:[0.90104], Metrics:[0.65625, 0.61291, 0.58608, 0.64232]\n",
      "Ep:1/5,Itr:37/469, Averages:[0.93805], Metrics:[0.65625, 0.67329, 0.65003, 0.69827]\n",
      "Ep:1/5,Itr:38/469, Averages:[0.81401], Metrics:[0.72656, 0.73596, 0.72391, 0.74842]\n",
      "Ep:1/5,Itr:39/469, Averages:[0.9907], Metrics:[0.63281, 0.63449, 0.63206, 0.63693]\n",
      "Ep:1/5,Itr:40/469, Averages:[0.92136], Metrics:[0.67188, 0.67602, 0.67028, 0.68187]\n",
      "Ep:1/5,Itr:41/469, Averages:[0.94668], Metrics:[0.67969, 0.68649, 0.66833, 0.70568]\n",
      "Ep:1/5,Itr:42/469, Averages:[0.89255], Metrics:[0.69531, 0.6827, 0.68734, 0.67812]\n",
      "Ep:1/5,Itr:43/469, Averages:[0.92525], Metrics:[0.70312, 0.73172, 0.72431, 0.73927]\n",
      "Ep:1/5,Itr:44/469, Averages:[0.90273], Metrics:[0.69531, 0.69148, 0.68923, 0.69374]\n",
      "Ep:1/5,Itr:45/469, Averages:[0.95622], Metrics:[0.67188, 0.67499, 0.67598, 0.674]\n",
      "Ep:1/5,Itr:46/469, Averages:[0.95823], Metrics:[0.67969, 0.68191, 0.67644, 0.68748]\n",
      "Ep:1/5,Itr:47/469, Averages:[0.85807], Metrics:[0.73438, 0.71911, 0.70676, 0.7319]\n",
      "Ep:1/5,Itr:48/469, Averages:[0.74279], Metrics:[0.75, 0.75365, 0.73716, 0.77089]\n",
      "Ep:1/5,Itr:49/469, Averages:[0.87294], Metrics:[0.70312, 0.73517, 0.71481, 0.75672]\n",
      "Ep:1/5,Itr:50/469, Averages:[0.77224], Metrics:[0.69531, 0.71672, 0.70977, 0.72381]\n",
      "Ep:1/5,Itr:51/469, Averages:[0.92081], Metrics:[0.67969, 0.68131, 0.6739, 0.68889]\n",
      "Ep:1/5,Itr:52/469, Averages:[0.63324], Metrics:[0.78125, 0.75537, 0.7559, 0.75483]\n",
      "Ep:1/5,Itr:53/469, Averages:[0.82454], Metrics:[0.69531, 0.70151, 0.68753, 0.71608]\n",
      "Ep:1/5,Itr:54/469, Averages:[0.60457], Metrics:[0.79688, 0.7739, 0.7687, 0.77917]\n",
      "Ep:1/5,Itr:55/469, Averages:[0.71905], Metrics:[0.73438, 0.74315, 0.72295, 0.76451]\n",
      "Ep:1/5,Itr:56/469, Averages:[0.8395], Metrics:[0.72656, 0.70245, 0.69439, 0.71069]\n",
      "Ep:1/5,Itr:57/469, Averages:[0.91676], Metrics:[0.67188, 0.69466, 0.68173, 0.7081]\n",
      "Ep:1/5,Itr:58/469, Averages:[0.65959], Metrics:[0.77344, 0.74961, 0.73846, 0.7611]\n",
      "Ep:1/5,Itr:59/469, Averages:[0.76602], Metrics:[0.74219, 0.73177, 0.74286, 0.72101]\n",
      "Ep:1/5,Itr:60/469, Averages:[0.71773], Metrics:[0.72656, 0.71746, 0.71192, 0.72309]\n",
      "Ep:1/5,Itr:61/469, Averages:[0.72561], Metrics:[0.82031, 0.82401, 0.82738, 0.82067]\n",
      "Ep:1/5,Itr:62/469, Averages:[0.65172], Metrics:[0.82812, 0.81077, 0.82268, 0.79919]\n",
      "Ep:1/5,Itr:63/469, Averages:[0.65969], Metrics:[0.78906, 0.79852, 0.78297, 0.81469]\n",
      "Ep:1/5,Itr:64/469, Averages:[0.65145], Metrics:[0.78906, 0.78388, 0.77301, 0.79507]\n",
      "Ep:1/5,Itr:65/469, Averages:[0.59384], Metrics:[0.77344, 0.78051, 0.76936, 0.79198]\n",
      "Ep:1/5,Itr:66/469, Averages:[0.70085], Metrics:[0.8125, 0.8003, 0.78615, 0.81496]\n",
      "Ep:1/5,Itr:67/469, Averages:[0.70226], Metrics:[0.77344, 0.77076, 0.76501, 0.7766]\n",
      "Ep:1/5,Itr:68/469, Averages:[0.70965], Metrics:[0.78125, 0.76164, 0.76752, 0.75585]\n",
      "Ep:1/5,Itr:69/469, Averages:[0.74764], Metrics:[0.78906, 0.79019, 0.78802, 0.79237]\n",
      "Ep:1/5,Itr:70/469, Averages:[0.73688], Metrics:[0.78906, 0.79041, 0.79056, 0.79025]\n",
      "Ep:1/5,Itr:71/469, Averages:[0.71308], Metrics:[0.78906, 0.7948, 0.79908, 0.79056]\n",
      "Ep:1/5,Itr:72/469, Averages:[0.55633], Metrics:[0.82812, 0.83291, 0.84154, 0.82444]\n",
      "Ep:1/5,Itr:73/469, Averages:[0.47353], Metrics:[0.84375, 0.84248, 0.84298, 0.84198]\n",
      "Ep:1/5,Itr:74/469, Averages:[0.58625], Metrics:[0.83594, 0.83328, 0.83857, 0.82807]\n",
      "Ep:1/5,Itr:75/469, Averages:[0.61947], Metrics:[0.8125, 0.81149, 0.81365, 0.80935]\n",
      "Ep:1/5,Itr:76/469, Averages:[0.66924], Metrics:[0.80469, 0.80547, 0.79915, 0.81189]\n",
      "Ep:1/5,Itr:77/469, Averages:[0.55634], Metrics:[0.78906, 0.78738, 0.78752, 0.78725]\n",
      "Ep:1/5,Itr:78/469, Averages:[0.62552], Metrics:[0.77344, 0.7817, 0.7824, 0.781]\n",
      "Ep:1/5,Itr:79/469, Averages:[0.54591], Metrics:[0.80469, 0.82146, 0.81966, 0.82327]\n",
      "Ep:1/5,Itr:80/469, Averages:[0.59582], Metrics:[0.78906, 0.7956, 0.79412, 0.79709]\n",
      "Ep:1/5,Itr:81/469, Averages:[0.51752], Metrics:[0.82031, 0.82987, 0.82801, 0.83173]\n",
      "Ep:1/5,Itr:82/469, Averages:[0.53942], Metrics:[0.82031, 0.82454, 0.81475, 0.83456]\n",
      "Ep:1/5,Itr:83/469, Averages:[0.54725], Metrics:[0.82812, 0.81984, 0.8256, 0.81415]\n",
      "Ep:1/5,Itr:84/469, Averages:[0.45533], Metrics:[0.85156, 0.84606, 0.83969, 0.85253]\n",
      "Ep:1/5,Itr:85/469, Averages:[0.56133], Metrics:[0.82031, 0.81967, 0.81299, 0.82647]\n",
      "Ep:1/5,Itr:86/469, Averages:[0.58958], Metrics:[0.8125, 0.81961, 0.8191, 0.82011]\n",
      "Ep:1/5,Itr:87/469, Averages:[0.55008], Metrics:[0.82031, 0.8235, 0.82011, 0.82693]\n",
      "Ep:1/5,Itr:88/469, Averages:[0.60589], Metrics:[0.78125, 0.79355, 0.78032, 0.80723]\n",
      "Ep:1/5,Itr:89/469, Averages:[0.55817], Metrics:[0.83594, 0.84158, 0.84336, 0.83982]\n",
      "Ep:1/5,Itr:90/469, Averages:[0.60851], Metrics:[0.78906, 0.80513, 0.80946, 0.80083]\n",
      "Ep:1/5,Itr:91/469, Averages:[0.42256], Metrics:[0.85938, 0.8653, 0.87529, 0.85552]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ep:1/5,Itr:92/469, Averages:[0.54911], Metrics:[0.78125, 0.78441, 0.77593, 0.79307]\n",
      "Ep:1/5,Itr:93/469, Averages:[0.58628], Metrics:[0.79688, 0.79691, 0.80056, 0.7933]\n",
      "Ep:1/5,Itr:94/469, Averages:[0.63839], Metrics:[0.82812, 0.82288, 0.8236, 0.82216]\n",
      "Ep:1/5,Itr:95/469, Averages:[0.58461], Metrics:[0.8125, 0.81357, 0.82185, 0.80546]\n",
      "Ep:1/5,Itr:96/469, Averages:[0.47353], Metrics:[0.84375, 0.84864, 0.84465, 0.85267]\n",
      "Ep:1/5,Itr:97/469, Averages:[0.6453], Metrics:[0.82812, 0.83009, 0.83477, 0.82546]\n",
      "Ep:1/5,Itr:98/469, Averages:[0.45219], Metrics:[0.83594, 0.83489, 0.83176, 0.83805]\n",
      "Ep:1/5,Itr:99/469, Averages:[0.51156], Metrics:[0.85156, 0.85481, 0.85339, 0.85623]\n",
      "Ep:1/5,Itr:100/469, Averages:[0.46599], Metrics:[0.85938, 0.86827, 0.87323, 0.86338]\n",
      "Ep:1/5,Itr:101/469, Averages:[0.55356], Metrics:[0.82031, 0.824, 0.82203, 0.82598]\n",
      "Ep:1/5,Itr:102/469, Averages:[0.58705], Metrics:[0.80469, 0.81142, 0.79561, 0.82788]\n",
      "Ep:1/5,Itr:103/469, Averages:[0.3618], Metrics:[0.90625, 0.9059, 0.91042, 0.90143]\n",
      "Ep:1/5,Itr:104/469, Averages:[0.54018], Metrics:[0.82031, 0.8186, 0.81881, 0.81839]\n",
      "Ep:1/5,Itr:105/469, Averages:[0.56479], Metrics:[0.8125, 0.81929, 0.81497, 0.82365]\n",
      "Ep:1/5,Itr:106/469, Averages:[0.62969], Metrics:[0.8125, 0.80924, 0.80923, 0.80925]\n",
      "Ep:1/5,Itr:107/469, Averages:[0.42777], Metrics:[0.85938, 0.86916, 0.86136, 0.8771]\n",
      "Ep:1/5,Itr:108/469, Averages:[0.52881], Metrics:[0.84375, 0.84508, 0.84013, 0.85009]\n",
      "Ep:1/5,Itr:109/469, Averages:[0.47935], Metrics:[0.83594, 0.84144, 0.8413, 0.84158]\n",
      "Ep:1/5,Itr:110/469, Averages:[0.59502], Metrics:[0.8125, 0.82186, 0.8138, 0.83008]\n",
      "Ep:1/5,Itr:111/469, Averages:[0.47652], Metrics:[0.86719, 0.87348, 0.86758, 0.87946]\n",
      "Ep:1/5,Itr:112/469, Averages:[0.53661], Metrics:[0.82031, 0.82562, 0.82693, 0.8243]\n",
      "Ep:1/5,Itr:113/469, Averages:[0.38906], Metrics:[0.85938, 0.85732, 0.85942, 0.85523]\n",
      "Ep:1/5,Itr:114/469, Averages:[0.48634], Metrics:[0.82031, 0.78883, 0.79078, 0.78689]\n",
      "Ep:1/5,Itr:115/469, Averages:[0.63435], Metrics:[0.80469, 0.80262, 0.79148, 0.81408]\n",
      "Ep:1/5,Itr:116/469, Averages:[0.34723], Metrics:[0.88281, 0.88082, 0.88153, 0.88012]\n",
      "Ep:1/5,Itr:117/469, Averages:[0.39642], Metrics:[0.85156, 0.84329, 0.83319, 0.85364]\n",
      "Ep:1/5,Itr:118/469, Averages:[0.38684], Metrics:[0.86719, 0.85575, 0.85308, 0.85844]\n",
      "Ep:1/5,Itr:119/469, Averages:[0.57428], Metrics:[0.82031, 0.82322, 0.81695, 0.82959]\n",
      "Ep:1/5,Itr:120/469, Averages:[0.54139], Metrics:[0.84375, 0.85273, 0.8513, 0.85417]\n",
      "Ep:1/5,Itr:121/469, Averages:[0.46927], Metrics:[0.85156, 0.86051, 0.86304, 0.858]\n",
      "Ep:1/5,Itr:122/469, Averages:[0.33598], Metrics:[0.89844, 0.89657, 0.88824, 0.90507]\n",
      "Ep:1/5,Itr:123/469, Averages:[0.51563], Metrics:[0.83594, 0.84391, 0.84852, 0.83935]\n",
      "Ep:1/5,Itr:124/469, Averages:[0.45027], Metrics:[0.91406, 0.9198, 0.92784, 0.9119]\n",
      "Ep:1/5,Itr:125/469, Averages:[0.40345], Metrics:[0.88281, 0.88982, 0.89378, 0.88589]\n",
      "Ep:1/5,Itr:126/469, Averages:[0.33942], Metrics:[0.91406, 0.90974, 0.90533, 0.91419]\n",
      "Ep:1/5,Itr:127/469, Averages:[0.53142], Metrics:[0.86719, 0.87102, 0.86497, 0.87715]\n",
      "Ep:1/5,Itr:128/469, Averages:[0.41672], Metrics:[0.88281, 0.88907, 0.8841, 0.8941]\n",
      "Ep:1/5,Itr:129/469, Averages:[0.50685], Metrics:[0.83594, 0.83454, 0.83351, 0.83556]\n",
      "Ep:1/5,Itr:130/469, Averages:[0.38768], Metrics:[0.875, 0.87546, 0.86199, 0.88936]\n",
      "Ep:1/5,Itr:131/469, Averages:[0.36574], Metrics:[0.90625, 0.90189, 0.90332, 0.90046]\n",
      "Ep:1/5,Itr:132/469, Averages:[0.4034], Metrics:[0.89062, 0.88585, 0.88132, 0.89043]\n",
      "Ep:1/5,Itr:133/469, Averages:[0.31252], Metrics:[0.90625, 0.90156, 0.90152, 0.9016]\n",
      "Ep:1/5,Itr:134/469, Averages:[0.38981], Metrics:[0.90625, 0.9151, 0.9153, 0.91489]\n",
      "Ep:1/5,Itr:135/469, Averages:[0.35141], Metrics:[0.90625, 0.90811, 0.9078, 0.90841]\n",
      "Ep:1/5,Itr:136/469, Averages:[0.36555], Metrics:[0.89062, 0.8854, 0.87654, 0.89444]\n",
      "Ep:1/5,Itr:137/469, Averages:[0.50769], Metrics:[0.82031, 0.82472, 0.82499, 0.82444]\n",
      "Ep:1/5,Itr:138/469, Averages:[0.3932], Metrics:[0.86719, 0.87959, 0.87702, 0.88218]\n",
      "Ep:1/5,Itr:139/469, Averages:[0.45745], Metrics:[0.84375, 0.8582, 0.86281, 0.85364]\n",
      "Ep:1/5,Itr:140/469, Averages:[0.40872], Metrics:[0.875, 0.87189, 0.87198, 0.8718]\n",
      "Ep:1/5,Itr:141/469, Averages:[0.52822], Metrics:[0.84375, 0.84107, 0.83965, 0.84249]\n",
      "Ep:1/5,Itr:142/469, Averages:[0.38821], Metrics:[0.875, 0.88154, 0.88421, 0.87888]\n",
      "Ep:1/5,Itr:143/469, Averages:[0.43514], Metrics:[0.875, 0.8725, 0.87187, 0.87312]\n",
      "Ep:1/5,Itr:144/469, Averages:[0.56118], Metrics:[0.83594, 0.83271, 0.82952, 0.83594]\n",
      "Ep:1/5,Itr:145/469, Averages:[0.38873], Metrics:[0.85938, 0.85712, 0.8617, 0.85259]\n",
      "Ep:1/5,Itr:146/469, Averages:[0.36704], Metrics:[0.90625, 0.9131, 0.91708, 0.90916]\n",
      "Ep:1/5,Itr:147/469, Averages:[0.28895], Metrics:[0.89844, 0.89766, 0.89798, 0.89734]\n",
      "Ep:1/5,Itr:148/469, Averages:[0.39068], Metrics:[0.89062, 0.8822, 0.89063, 0.87392]\n",
      "Ep:1/5,Itr:149/469, Averages:[0.28554], Metrics:[0.91406, 0.91586, 0.91028, 0.9215]\n",
      "Ep:1/5,Itr:150/469, Averages:[0.43775], Metrics:[0.89062, 0.88528, 0.8948, 0.87596]\n",
      "Ep:1/5,Itr:151/469, Averages:[0.22085], Metrics:[0.9375, 0.93886, 0.93353, 0.94424]\n",
      "Ep:1/5,Itr:152/469, Averages:[0.24232], Metrics:[0.90625, 0.90696, 0.91079, 0.90317]\n",
      "Ep:1/5,Itr:153/469, Averages:[0.58935], Metrics:[0.83594, 0.83609, 0.83428, 0.83789]\n",
      "Ep:1/5,Itr:154/469, Averages:[0.38858], Metrics:[0.89062, 0.89049, 0.89058, 0.89041]\n",
      "Ep:1/5,Itr:155/469, Averages:[0.4165], Metrics:[0.875, 0.87766, 0.87155, 0.88387]\n",
      "Ep:1/5,Itr:156/469, Averages:[0.35477], Metrics:[0.91406, 0.92321, 0.92812, 0.91835]\n",
      "Ep:1/5,Itr:157/469, Averages:[0.27996], Metrics:[0.91406, 0.90908, 0.90476, 0.91343]\n",
      "Ep:1/5,Itr:158/469, Averages:[0.37583], Metrics:[0.90625, 0.9116, 0.91334, 0.90987]\n",
      "Ep:1/5,Itr:159/469, Averages:[0.41214], Metrics:[0.85938, 0.8637, 0.86064, 0.86678]\n",
      "Ep:1/5,Itr:160/469, Averages:[0.29635], Metrics:[0.89062, 0.88877, 0.89035, 0.8872]\n",
      "Ep:1/5,Itr:161/469, Averages:[0.32578], Metrics:[0.90625, 0.89925, 0.89239, 0.90622]\n",
      "Ep:1/5,Itr:162/469, Averages:[0.28558], Metrics:[0.90625, 0.90535, 0.90137, 0.90936]\n",
      "Ep:1/5,Itr:163/469, Averages:[0.44556], Metrics:[0.875, 0.88786, 0.88176, 0.89404]\n",
      "Ep:1/5,Itr:164/469, Averages:[0.40916], Metrics:[0.89062, 0.8951, 0.88187, 0.90873]\n",
      "Ep:1/5,Itr:165/469, Averages:[0.34683], Metrics:[0.90625, 0.90408, 0.90447, 0.90369]\n",
      "Ep:1/5,Itr:166/469, Averages:[0.41079], Metrics:[0.875, 0.87864, 0.86998, 0.88748]\n",
      "Ep:1/5,Itr:167/469, Averages:[0.43942], Metrics:[0.86719, 0.87109, 0.87103, 0.87114]\n",
      "Ep:1/5,Itr:168/469, Averages:[0.26646], Metrics:[0.91406, 0.91651, 0.91109, 0.92199]\n",
      "Ep:1/5,Itr:169/469, Averages:[0.24377], Metrics:[0.92969, 0.93146, 0.93114, 0.93179]\n",
      "Ep:1/5,Itr:170/469, Averages:[0.3801], Metrics:[0.875, 0.88334, 0.88405, 0.88263]\n",
      "Ep:1/5,Itr:171/469, Averages:[0.39336], Metrics:[0.89062, 0.88938, 0.89068, 0.88809]\n",
      "Ep:1/5,Itr:172/469, Averages:[0.26082], Metrics:[0.90625, 0.9071, 0.90219, 0.91207]\n",
      "Ep:1/5,Itr:173/469, Averages:[0.44168], Metrics:[0.85938, 0.85399, 0.85353, 0.85444]\n",
      "Ep:1/5,Itr:174/469, Averages:[0.37041], Metrics:[0.86719, 0.86713, 0.87683, 0.85764]\n",
      "Ep:1/5,Itr:175/469, Averages:[0.32482], Metrics:[0.90625, 0.91027, 0.90875, 0.91179]\n",
      "Ep:1/5,Itr:176/469, Averages:[0.4624], Metrics:[0.84375, 0.85545, 0.84622, 0.86489]\n",
      "Ep:1/5,Itr:177/469, Averages:[0.41761], Metrics:[0.88281, 0.88047, 0.88434, 0.87663]\n",
      "Ep:1/5,Itr:178/469, Averages:[0.33652], Metrics:[0.88281, 0.8813, 0.87973, 0.88287]\n",
      "Ep:1/5,Itr:179/469, Averages:[0.32784], Metrics:[0.9375, 0.93502, 0.92878, 0.94134]\n",
      "Ep:1/5,Itr:180/469, Averages:[0.32179], Metrics:[0.89844, 0.90927, 0.90719, 0.91136]\n",
      "Ep:1/5,Itr:181/469, Averages:[0.41619], Metrics:[0.85156, 0.84609, 0.85079, 0.84144]\n",
      "Ep:1/5,Itr:182/469, Averages:[0.37125], Metrics:[0.86719, 0.85866, 0.85723, 0.8601]\n",
      "Ep:1/5,Itr:183/469, Averages:[0.34584], Metrics:[0.9375, 0.93092, 0.91673, 0.94554]\n",
      "Ep:1/5,Itr:184/469, Averages:[0.38048], Metrics:[0.91406, 0.9132, 0.9115, 0.9149]\n",
      "Ep:1/5,Itr:185/469, Averages:[0.29647], Metrics:[0.89844, 0.89271, 0.90217, 0.88346]\n",
      "Ep:1/5,Itr:186/469, Averages:[0.31391], Metrics:[0.92188, 0.92598, 0.92631, 0.92565]\n",
      "Ep:1/5,Itr:187/469, Averages:[0.15691], Metrics:[0.94531, 0.94471, 0.94597, 0.94345]\n",
      "Ep:1/5,Itr:188/469, Averages:[0.2329], Metrics:[0.96094, 0.9609, 0.96351, 0.9583]\n",
      "Ep:1/5,Itr:189/469, Averages:[0.18753], Metrics:[0.96094, 0.95478, 0.95636, 0.95322]\n",
      "Ep:1/5,Itr:190/469, Averages:[0.28037], Metrics:[0.92969, 0.93463, 0.93125, 0.93803]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ep:1/5,Itr:191/469, Averages:[0.24046], Metrics:[0.92969, 0.92149, 0.92497, 0.91804]\n",
      "Ep:1/5,Itr:192/469, Averages:[0.26319], Metrics:[0.89844, 0.91379, 0.92388, 0.90392]\n",
      "Ep:1/5,Itr:193/469, Averages:[0.23744], Metrics:[0.92969, 0.93222, 0.92466, 0.9399]\n",
      "Ep:1/5,Itr:194/469, Averages:[0.28686], Metrics:[0.89062, 0.87812, 0.88451, 0.87183]\n",
      "Ep:1/5,Itr:195/469, Averages:[0.40528], Metrics:[0.875, 0.8798, 0.87557, 0.88408]\n",
      "Ep:1/5,Itr:196/469, Averages:[0.2417], Metrics:[0.90625, 0.9065, 0.88913, 0.92458]\n",
      "Ep:1/5,Itr:197/469, Averages:[0.24935], Metrics:[0.94531, 0.93976, 0.9371, 0.94245]\n",
      "Ep:1/5,Itr:198/469, Averages:[0.52739], Metrics:[0.875, 0.88509, 0.8801, 0.89014]\n",
      "Ep:1/5,Itr:199/469, Averages:[0.29536], Metrics:[0.90625, 0.89515, 0.89422, 0.89609]\n",
      "Ep:1/5,Itr:200/469, Averages:[0.36324], Metrics:[0.88281, 0.89039, 0.89393, 0.88688]\n",
      "Ep:1/5,Itr:201/469, Averages:[0.29965], Metrics:[0.90625, 0.90286, 0.90767, 0.8981]\n",
      "Ep:1/5,Itr:202/469, Averages:[0.20859], Metrics:[0.92969, 0.92758, 0.92469, 0.93048]\n",
      "Ep:1/5,Itr:203/469, Averages:[0.32655], Metrics:[0.875, 0.86975, 0.86563, 0.8739]\n",
      "Ep:1/5,Itr:204/469, Averages:[0.35609], Metrics:[0.88281, 0.88203, 0.88449, 0.87959]\n",
      "Ep:1/5,Itr:205/469, Averages:[0.24532], Metrics:[0.92188, 0.92618, 0.9275, 0.92485]\n",
      "Ep:1/5,Itr:206/469, Averages:[0.2847], Metrics:[0.90625, 0.91009, 0.91356, 0.90665]\n",
      "Ep:1/5,Itr:207/469, Averages:[0.27238], Metrics:[0.91406, 0.91715, 0.92063, 0.9137]\n",
      "Ep:1/5,Itr:208/469, Averages:[0.30149], Metrics:[0.89062, 0.90334, 0.89433, 0.91253]\n",
      "Ep:1/5,Itr:209/469, Averages:[0.37691], Metrics:[0.86719, 0.86371, 0.8629, 0.86453]\n",
      "Ep:1/5,Itr:210/469, Averages:[0.34489], Metrics:[0.85938, 0.86369, 0.85769, 0.86978]\n",
      "Ep:1/5,Itr:211/469, Averages:[0.33522], Metrics:[0.88281, 0.89021, 0.89873, 0.88186]\n",
      "Ep:1/5,Itr:212/469, Averages:[0.28005], Metrics:[0.9375, 0.93714, 0.93756, 0.93673]\n",
      "Ep:1/5,Itr:213/469, Averages:[0.33333], Metrics:[0.90625, 0.89512, 0.89275, 0.8975]\n",
      "Ep:1/5,Itr:214/469, Averages:[0.26299], Metrics:[0.90625, 0.9115, 0.91139, 0.9116]\n",
      "Ep:1/5,Itr:215/469, Averages:[0.27761], Metrics:[0.90625, 0.90323, 0.90953, 0.89702]\n",
      "Ep:1/5,Itr:216/469, Averages:[0.15226], Metrics:[0.94531, 0.9509, 0.94562, 0.95623]\n",
      "Ep:1/5,Itr:217/469, Averages:[0.24943], Metrics:[0.9375, 0.94306, 0.94229, 0.94383]\n",
      "Ep:1/5,Itr:218/469, Averages:[0.35713], Metrics:[0.88281, 0.89174, 0.88614, 0.89742]\n",
      "Ep:1/5,Itr:219/469, Averages:[0.29641], Metrics:[0.88281, 0.89314, 0.89972, 0.88664]\n",
      "Ep:1/5,Itr:220/469, Averages:[0.27562], Metrics:[0.89844, 0.89422, 0.89829, 0.89019]\n",
      "Ep:1/5,Itr:221/469, Averages:[0.2845], Metrics:[0.92969, 0.93132, 0.94096, 0.92187]\n",
      "Ep:1/5,Itr:222/469, Averages:[0.2418], Metrics:[0.89844, 0.89965, 0.89727, 0.90203]\n",
      "Ep:1/5,Itr:223/469, Averages:[0.4539], Metrics:[0.875, 0.86975, 0.8736, 0.86593]\n",
      "Ep:1/5,Itr:224/469, Averages:[0.18752], Metrics:[0.91406, 0.91162, 0.90928, 0.91397]\n",
      "Ep:1/5,Itr:225/469, Averages:[0.21841], Metrics:[0.9375, 0.9427, 0.94024, 0.94516]\n",
      "Ep:1/5,Itr:226/469, Averages:[0.33974], Metrics:[0.89062, 0.8869, 0.89408, 0.87983]\n",
      "Ep:1/5,Itr:227/469, Averages:[0.28463], Metrics:[0.92969, 0.92788, 0.92628, 0.92948]\n",
      "Ep:1/5,Itr:228/469, Averages:[0.34704], Metrics:[0.89844, 0.89452, 0.89912, 0.88996]\n",
      "Ep:1/5,Itr:229/469, Averages:[0.28006], Metrics:[0.92969, 0.92639, 0.91924, 0.93365]\n",
      "Ep:1/5,Itr:230/469, Averages:[0.31249], Metrics:[0.90625, 0.8939, 0.90079, 0.88712]\n",
      "Ep:1/5,Itr:231/469, Averages:[0.25448], Metrics:[0.92969, 0.93125, 0.92258, 0.94009]\n",
      "Ep:1/5,Itr:232/469, Averages:[0.25156], Metrics:[0.92188, 0.92543, 0.91575, 0.93531]\n",
      "Ep:1/5,Itr:233/469, Averages:[0.34118], Metrics:[0.90625, 0.87526, 0.86593, 0.88479]\n",
      "Ep:1/5,Itr:234/469, Averages:[0.18681], Metrics:[0.94531, 0.94807, 0.94914, 0.947]\n",
      "Ep:1/5,Itr:235/469, Averages:[0.21233], Metrics:[0.9375, 0.93129, 0.925, 0.93767]\n",
      "Ep:1/5,Itr:236/469, Averages:[0.38345], Metrics:[0.91406, 0.90669, 0.90816, 0.90524]\n",
      "Ep:1/5,Itr:237/469, Averages:[0.31733], Metrics:[0.91406, 0.91557, 0.91593, 0.91521]\n",
      "Ep:1/5,Itr:238/469, Averages:[0.3539], Metrics:[0.86719, 0.86838, 0.86505, 0.87173]\n",
      "Ep:1/5,Itr:239/469, Averages:[0.31408], Metrics:[0.92188, 0.91906, 0.92938, 0.90897]\n",
      "Ep:1/5,Itr:240/469, Averages:[0.22763], Metrics:[0.91406, 0.91371, 0.90889, 0.91858]\n",
      "Ep:1/5,Itr:241/469, Averages:[0.32376], Metrics:[0.91406, 0.922, 0.91847, 0.92555]\n",
      "Ep:1/5,Itr:242/469, Averages:[0.24136], Metrics:[0.92969, 0.92086, 0.91818, 0.92356]\n",
      "Ep:1/5,Itr:243/469, Averages:[0.35259], Metrics:[0.89844, 0.89863, 0.90307, 0.89423]\n",
      "Ep:1/5,Itr:244/469, Averages:[0.29646], Metrics:[0.89062, 0.88292, 0.88688, 0.879]\n",
      "Ep:1/5,Itr:245/469, Averages:[0.28977], Metrics:[0.90625, 0.90929, 0.90991, 0.90868]\n",
      "Ep:1/5,Itr:246/469, Averages:[0.3336], Metrics:[0.92969, 0.92734, 0.93917, 0.91581]\n",
      "Ep:1/5,Itr:247/469, Averages:[0.23424], Metrics:[0.9375, 0.93991, 0.94924, 0.93076]\n",
      "Ep:1/5,Itr:248/469, Averages:[0.3016], Metrics:[0.90625, 0.90414, 0.90639, 0.90191]\n",
      "Ep:1/5,Itr:249/469, Averages:[0.27281], Metrics:[0.90625, 0.90573, 0.90127, 0.91023]\n",
      "Ep:1/5,Itr:250/469, Averages:[0.31771], Metrics:[0.92188, 0.92196, 0.92327, 0.92065]\n",
      "Ep:1/5,Itr:251/469, Averages:[0.21385], Metrics:[0.91406, 0.91157, 0.91014, 0.91301]\n",
      "Ep:1/5,Itr:252/469, Averages:[0.28548], Metrics:[0.92188, 0.93115, 0.93411, 0.92821]\n",
      "Ep:1/5,Itr:253/469, Averages:[0.1828], Metrics:[0.94531, 0.94347, 0.94469, 0.94225]\n",
      "Ep:1/5,Itr:254/469, Averages:[0.40558], Metrics:[0.88281, 0.88322, 0.8847, 0.88175]\n",
      "Ep:1/5,Itr:255/469, Averages:[0.16346], Metrics:[0.9375, 0.93528, 0.93667, 0.93389]\n",
      "Ep:1/5,Itr:256/469, Averages:[0.1943], Metrics:[0.92969, 0.92739, 0.92747, 0.92731]\n",
      "Ep:1/5,Itr:257/469, Averages:[0.25193], Metrics:[0.90625, 0.9078, 0.91378, 0.90189]\n",
      "Ep:1/5,Itr:258/469, Averages:[0.16099], Metrics:[0.96875, 0.96702, 0.96556, 0.9685]\n",
      "Ep:1/5,Itr:259/469, Averages:[0.38737], Metrics:[0.92188, 0.92699, 0.92391, 0.9301]\n",
      "Ep:1/5,Itr:260/469, Averages:[0.33456], Metrics:[0.91406, 0.9195, 0.9216, 0.91741]\n",
      "Ep:1/5,Itr:261/469, Averages:[0.20328], Metrics:[0.9375, 0.94302, 0.94327, 0.94277]\n",
      "Ep:1/5,Itr:262/469, Averages:[0.22784], Metrics:[0.9375, 0.93642, 0.92915, 0.94381]\n",
      "Ep:1/5,Itr:263/469, Averages:[0.21613], Metrics:[0.96094, 0.95733, 0.95817, 0.9565]\n",
      "Ep:1/5,Itr:264/469, Averages:[0.2229], Metrics:[0.9375, 0.93174, 0.9317, 0.93178]\n",
      "Ep:1/5,Itr:265/469, Averages:[0.14886], Metrics:[0.96094, 0.95615, 0.95188, 0.96045]\n",
      "Ep:1/5,Itr:266/469, Averages:[0.26834], Metrics:[0.90625, 0.90453, 0.90788, 0.90121]\n",
      "Ep:1/5,Itr:267/469, Averages:[0.20659], Metrics:[0.92188, 0.92187, 0.92938, 0.91448]\n",
      "Ep:1/5,Itr:268/469, Averages:[0.18751], Metrics:[0.95312, 0.9528, 0.95401, 0.95159]\n",
      "Ep:1/5,Itr:269/469, Averages:[0.14085], Metrics:[0.96094, 0.95499, 0.95556, 0.95443]\n",
      "Ep:1/5,Itr:270/469, Averages:[0.20755], Metrics:[0.92969, 0.93234, 0.93873, 0.92603]\n",
      "Ep:1/5,Itr:271/469, Averages:[0.22759], Metrics:[0.92969, 0.92257, 0.9189, 0.92628]\n",
      "Ep:1/5,Itr:272/469, Averages:[0.19079], Metrics:[0.92969, 0.93715, 0.93509, 0.93922]\n",
      "Ep:1/5,Itr:273/469, Averages:[0.32779], Metrics:[0.91406, 0.91536, 0.91464, 0.91609]\n",
      "Ep:1/5,Itr:274/469, Averages:[0.2884], Metrics:[0.92188, 0.93021, 0.92817, 0.93226]\n",
      "Ep:1/5,Itr:275/469, Averages:[0.25188], Metrics:[0.91406, 0.9228, 0.92294, 0.92265]\n",
      "Ep:1/5,Itr:276/469, Averages:[0.28363], Metrics:[0.89062, 0.88515, 0.88824, 0.88208]\n",
      "Ep:1/5,Itr:277/469, Averages:[0.28379], Metrics:[0.9375, 0.9369, 0.94483, 0.92911]\n",
      "Ep:1/5,Itr:278/469, Averages:[0.24337], Metrics:[0.96875, 0.9694, 0.96786, 0.97095]\n",
      "Ep:1/5,Itr:279/469, Averages:[0.4313], Metrics:[0.91406, 0.91199, 0.91321, 0.91078]\n",
      "Ep:1/5,Itr:280/469, Averages:[0.32098], Metrics:[0.90625, 0.9086, 0.90885, 0.90836]\n",
      "Ep:1/5,Itr:281/469, Averages:[0.12464], Metrics:[0.97656, 0.97783, 0.97788, 0.97778]\n",
      "Ep:1/5,Itr:282/469, Averages:[0.28526], Metrics:[0.89844, 0.90279, 0.90264, 0.90294]\n",
      "Ep:1/5,Itr:283/469, Averages:[0.17968], Metrics:[0.95312, 0.95357, 0.95643, 0.95074]\n",
      "Ep:1/5,Itr:284/469, Averages:[0.28453], Metrics:[0.91406, 0.9166, 0.90089, 0.93285]\n",
      "Ep:1/5,Itr:285/469, Averages:[0.33783], Metrics:[0.92188, 0.91884, 0.91795, 0.91973]\n",
      "Ep:1/5,Itr:286/469, Averages:[0.26035], Metrics:[0.92969, 0.92866, 0.93043, 0.9269]\n",
      "Ep:1/5,Itr:287/469, Averages:[0.1644], Metrics:[0.94531, 0.95192, 0.95678, 0.94711]\n",
      "Ep:1/5,Itr:288/469, Averages:[0.37169], Metrics:[0.89062, 0.88589, 0.88338, 0.88841]\n",
      "Ep:1/5,Itr:289/469, Averages:[0.16485], Metrics:[0.94531, 0.94308, 0.94116, 0.945]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ep:1/5,Itr:290/469, Averages:[0.17863], Metrics:[0.92188, 0.92007, 0.92005, 0.92009]\n",
      "Ep:1/5,Itr:291/469, Averages:[0.12605], Metrics:[0.95312, 0.951, 0.95314, 0.94886]\n",
      "Ep:1/5,Itr:292/469, Averages:[0.1376], Metrics:[0.96094, 0.96326, 0.96357, 0.96295]\n",
      "Ep:1/5,Itr:293/469, Averages:[0.20592], Metrics:[0.92188, 0.92611, 0.9271, 0.92512]\n",
      "Ep:1/5,Itr:294/469, Averages:[0.20144], Metrics:[0.92188, 0.90404, 0.90075, 0.90736]\n",
      "Ep:1/5,Itr:295/469, Averages:[0.19335], Metrics:[0.92188, 0.92258, 0.93026, 0.91502]\n",
      "Ep:1/5,Itr:296/469, Averages:[0.21399], Metrics:[0.90625, 0.91481, 0.91518, 0.91444]\n",
      "Ep:1/5,Itr:297/469, Averages:[0.26538], Metrics:[0.9375, 0.93954, 0.94308, 0.93602]\n",
      "Ep:1/5,Itr:298/469, Averages:[0.16865], Metrics:[0.95312, 0.95445, 0.95759, 0.95134]\n",
      "Ep:1/5,Itr:299/469, Averages:[0.18573], Metrics:[0.92969, 0.9261, 0.93398, 0.91836]\n",
      "Ep:1/5,Itr:300/469, Averages:[0.17901], Metrics:[0.94531, 0.94197, 0.93731, 0.94668]\n",
      "Ep:1/5,Itr:301/469, Averages:[0.373], Metrics:[0.89062, 0.89604, 0.89937, 0.89274]\n",
      "Ep:1/5,Itr:302/469, Averages:[0.24617], Metrics:[0.94531, 0.94469, 0.94302, 0.94638]\n",
      "Ep:1/5,Itr:303/469, Averages:[0.2575], Metrics:[0.92969, 0.92893, 0.92685, 0.93101]\n",
      "Ep:1/5,Itr:304/469, Averages:[0.25083], Metrics:[0.92969, 0.9309, 0.93287, 0.92893]\n",
      "Ep:1/5,Itr:305/469, Averages:[0.13744], Metrics:[0.95312, 0.94909, 0.94878, 0.94939]\n",
      "Ep:1/5,Itr:306/469, Averages:[0.25358], Metrics:[0.92188, 0.92389, 0.92275, 0.92503]\n",
      "Ep:1/5,Itr:307/469, Averages:[0.19056], Metrics:[0.92969, 0.93142, 0.93704, 0.92587]\n",
      "Ep:1/5,Itr:308/469, Averages:[0.1808], Metrics:[0.9375, 0.93291, 0.93071, 0.93511]\n",
      "Ep:1/5,Itr:309/469, Averages:[0.20776], Metrics:[0.92188, 0.93235, 0.92314, 0.94174]\n",
      "Ep:1/5,Itr:310/469, Averages:[0.21303], Metrics:[0.90625, 0.90339, 0.89859, 0.90824]\n",
      "Ep:1/5,Itr:311/469, Averages:[0.24503], Metrics:[0.94531, 0.94214, 0.93779, 0.94653]\n",
      "Ep:1/5,Itr:312/469, Averages:[0.27461], Metrics:[0.92969, 0.93594, 0.93809, 0.9338]\n",
      "Ep:1/5,Itr:313/469, Averages:[0.23478], Metrics:[0.9375, 0.93613, 0.93583, 0.93643]\n",
      "Ep:1/5,Itr:314/469, Averages:[0.13676], Metrics:[0.97656, 0.97683, 0.97033, 0.98342]\n",
      "Ep:1/5,Itr:315/469, Averages:[0.16596], Metrics:[0.94531, 0.95042, 0.95329, 0.94758]\n",
      "Ep:1/5,Itr:316/469, Averages:[0.27174], Metrics:[0.94531, 0.94657, 0.94917, 0.94398]\n",
      "Ep:1/5,Itr:317/469, Averages:[0.2011], Metrics:[0.95312, 0.9564, 0.95401, 0.95881]\n",
      "Ep:1/5,Itr:318/469, Averages:[0.28596], Metrics:[0.92969, 0.93219, 0.9294, 0.93499]\n",
      "Ep:1/5,Itr:319/469, Averages:[0.25882], Metrics:[0.92188, 0.92835, 0.92715, 0.92955]\n",
      "Ep:1/5,Itr:320/469, Averages:[0.35722], Metrics:[0.86719, 0.86791, 0.87481, 0.86112]\n",
      "Ep:1/5,Itr:321/469, Averages:[0.11949], Metrics:[0.95312, 0.95413, 0.96026, 0.94807]\n",
      "Ep:1/5,Itr:322/469, Averages:[0.17708], Metrics:[0.92969, 0.93123, 0.92932, 0.93314]\n",
      "Ep:1/5,Itr:323/469, Averages:[0.27221], Metrics:[0.92969, 0.93084, 0.92695, 0.93476]\n",
      "Ep:1/5,Itr:324/469, Averages:[0.09772], Metrics:[0.96875, 0.96802, 0.96519, 0.97087]\n",
      "Ep:1/5,Itr:325/469, Averages:[0.21117], Metrics:[0.9375, 0.93463, 0.93515, 0.93412]\n",
      "Ep:1/5,Itr:326/469, Averages:[0.30354], Metrics:[0.89844, 0.89976, 0.90591, 0.8937]\n",
      "Ep:1/5,Itr:327/469, Averages:[0.23703], Metrics:[0.89062, 0.88064, 0.88583, 0.87551]\n",
      "Ep:1/5,Itr:328/469, Averages:[0.21578], Metrics:[0.90625, 0.90503, 0.90351, 0.90655]\n",
      "Ep:1/5,Itr:329/469, Averages:[0.13933], Metrics:[0.96094, 0.96399, 0.96234, 0.96565]\n",
      "Ep:1/5,Itr:330/469, Averages:[0.23312], Metrics:[0.92188, 0.92667, 0.92365, 0.9297]\n",
      "Ep:1/5,Itr:331/469, Averages:[0.22778], Metrics:[0.94531, 0.94405, 0.94402, 0.94408]\n",
      "Ep:1/5,Itr:332/469, Averages:[0.10606], Metrics:[0.97656, 0.97703, 0.97778, 0.97628]\n",
      "Ep:1/5,Itr:333/469, Averages:[0.17231], Metrics:[0.94531, 0.95041, 0.95066, 0.95016]\n",
      "Ep:1/5,Itr:334/469, Averages:[0.18534], Metrics:[0.94531, 0.94776, 0.95302, 0.94255]\n",
      "Ep:1/5,Itr:335/469, Averages:[0.2733], Metrics:[0.94531, 0.94042, 0.94554, 0.93535]\n",
      "Ep:1/5,Itr:336/469, Averages:[0.25024], Metrics:[0.91406, 0.90462, 0.88927, 0.92051]\n",
      "Ep:1/5,Itr:337/469, Averages:[0.22843], Metrics:[0.92969, 0.93736, 0.93906, 0.93568]\n",
      "Ep:1/5,Itr:338/469, Averages:[0.18568], Metrics:[0.92969, 0.92166, 0.92386, 0.91948]\n",
      "Ep:1/5,Itr:339/469, Averages:[0.32916], Metrics:[0.90625, 0.90296, 0.89798, 0.90799]\n",
      "Ep:1/5,Itr:340/469, Averages:[0.25411], Metrics:[0.94531, 0.94597, 0.9454, 0.94654]\n",
      "Ep:1/5,Itr:341/469, Averages:[0.09078], Metrics:[0.96094, 0.96256, 0.96807, 0.95711]\n",
      "Ep:1/5,Itr:342/469, Averages:[0.1576], Metrics:[0.98438, 0.97981, 0.97946, 0.98016]\n",
      "Ep:1/5,Itr:343/469, Averages:[0.19067], Metrics:[0.9375, 0.93446, 0.9355, 0.93342]\n",
      "Ep:1/5,Itr:344/469, Averages:[0.19793], Metrics:[0.9375, 0.93756, 0.94051, 0.93462]\n",
      "Ep:1/5,Itr:345/469, Averages:[0.1691], Metrics:[0.92969, 0.93542, 0.92711, 0.94389]\n",
      "Ep:1/5,Itr:346/469, Averages:[0.24627], Metrics:[0.92188, 0.92258, 0.91789, 0.92732]\n",
      "Ep:1/5,Itr:347/469, Averages:[0.34584], Metrics:[0.90625, 0.9083, 0.90481, 0.91183]\n",
      "Ep:1/5,Itr:348/469, Averages:[0.32815], Metrics:[0.91406, 0.91505, 0.91434, 0.91577]\n",
      "Ep:1/5,Itr:349/469, Averages:[0.25342], Metrics:[0.91406, 0.91225, 0.91992, 0.9047]\n",
      "Ep:1/5,Itr:350/469, Averages:[0.18414], Metrics:[0.94531, 0.95635, 0.95813, 0.95458]\n",
      "Ep:1/5,Itr:351/469, Averages:[0.1984], Metrics:[0.92969, 0.92877, 0.93178, 0.92579]\n",
      "Ep:1/5,Itr:352/469, Averages:[0.30444], Metrics:[0.9375, 0.93648, 0.93657, 0.93639]\n",
      "Ep:1/5,Itr:353/469, Averages:[0.27671], Metrics:[0.91406, 0.9199, 0.91798, 0.92182]\n",
      "Ep:1/5,Itr:354/469, Averages:[0.1106], Metrics:[0.96094, 0.94925, 0.95064, 0.94786]\n",
      "Ep:1/5,Itr:355/469, Averages:[0.18395], Metrics:[0.9375, 0.93743, 0.93973, 0.93514]\n",
      "Ep:1/5,Itr:356/469, Averages:[0.29055], Metrics:[0.90625, 0.9001, 0.88943, 0.91103]\n",
      "Ep:1/5,Itr:357/469, Averages:[0.25417], Metrics:[0.91406, 0.90275, 0.89041, 0.91544]\n",
      "Ep:1/5,Itr:358/469, Averages:[0.18854], Metrics:[0.94531, 0.94881, 0.95417, 0.9435]\n",
      "Ep:1/5,Itr:359/469, Averages:[0.12965], Metrics:[0.95312, 0.95111, 0.95335, 0.94889]\n",
      "Ep:1/5,Itr:360/469, Averages:[0.17374], Metrics:[0.9375, 0.92323, 0.91504, 0.93156]\n",
      "Ep:1/5,Itr:361/469, Averages:[0.16634], Metrics:[0.96094, 0.95879, 0.95265, 0.96502]\n",
      "Ep:1/5,Itr:362/469, Averages:[0.21605], Metrics:[0.9375, 0.93879, 0.93979, 0.9378]\n",
      "Ep:1/5,Itr:363/469, Averages:[0.12753], Metrics:[0.96875, 0.96054, 0.94902, 0.97235]\n",
      "Ep:1/5,Itr:364/469, Averages:[0.15485], Metrics:[0.95312, 0.9574, 0.9573, 0.9575]\n",
      "Ep:1/5,Itr:365/469, Averages:[0.28881], Metrics:[0.92969, 0.93154, 0.9375, 0.92565]\n",
      "Ep:1/5,Itr:366/469, Averages:[0.28481], Metrics:[0.92969, 0.94058, 0.94212, 0.93904]\n",
      "Ep:1/5,Itr:367/469, Averages:[0.30134], Metrics:[0.9375, 0.93405, 0.93187, 0.93623]\n",
      "Ep:1/5,Itr:368/469, Averages:[0.13651], Metrics:[0.97656, 0.9768, 0.97803, 0.97557]\n",
      "Ep:1/5,Itr:369/469, Averages:[0.11969], Metrics:[0.95312, 0.95406, 0.95807, 0.95008]\n",
      "Ep:1/5,Itr:370/469, Averages:[0.24983], Metrics:[0.9375, 0.94058, 0.93962, 0.94155]\n",
      "Ep:1/5,Itr:371/469, Averages:[0.16911], Metrics:[0.94531, 0.94957, 0.94302, 0.9562]\n",
      "Ep:1/5,Itr:372/469, Averages:[0.25743], Metrics:[0.92188, 0.93343, 0.92191, 0.94524]\n",
      "Ep:1/5,Itr:373/469, Averages:[0.29024], Metrics:[0.90625, 0.90858, 0.91145, 0.90573]\n",
      "Ep:1/5,Itr:374/469, Averages:[0.27522], Metrics:[0.95312, 0.95359, 0.95556, 0.95162]\n",
      "Ep:1/5,Itr:375/469, Averages:[0.18083], Metrics:[0.94531, 0.94636, 0.93946, 0.95337]\n",
      "Ep:1/5,Itr:376/469, Averages:[0.13566], Metrics:[0.92969, 0.91624, 0.91568, 0.9168]\n",
      "Ep:1/5,Itr:377/469, Averages:[0.1935], Metrics:[0.9375, 0.93602, 0.93585, 0.9362]\n",
      "Ep:1/5,Itr:378/469, Averages:[0.12337], Metrics:[0.95312, 0.95174, 0.94746, 0.95606]\n",
      "Ep:1/5,Itr:379/469, Averages:[0.21488], Metrics:[0.91406, 0.91344, 0.90601, 0.92099]\n",
      "Ep:1/5,Itr:380/469, Averages:[0.15137], Metrics:[0.9375, 0.93928, 0.94056, 0.93801]\n",
      "Ep:1/5,Itr:381/469, Averages:[0.17961], Metrics:[0.95312, 0.95284, 0.95508, 0.95061]\n",
      "Ep:1/5,Itr:382/469, Averages:[0.1866], Metrics:[0.92969, 0.93412, 0.94437, 0.92408]\n",
      "Ep:1/5,Itr:383/469, Averages:[0.29881], Metrics:[0.90625, 0.90391, 0.90051, 0.90733]\n",
      "Ep:1/5,Itr:384/469, Averages:[0.20452], Metrics:[0.94531, 0.94941, 0.9486, 0.95021]\n",
      "Ep:1/5,Itr:385/469, Averages:[0.09321], Metrics:[0.97656, 0.97529, 0.97264, 0.97795]\n",
      "Ep:1/5,Itr:386/469, Averages:[0.29926], Metrics:[0.90625, 0.90815, 0.90547, 0.91085]\n",
      "Ep:1/5,Itr:387/469, Averages:[0.18074], Metrics:[0.94531, 0.94969, 0.94851, 0.95088]\n",
      "Ep:1/5,Itr:388/469, Averages:[0.11092], Metrics:[0.96094, 0.95518, 0.94995, 0.96046]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ep:1/5,Itr:389/469, Averages:[0.39967], Metrics:[0.90625, 0.90403, 0.89969, 0.90841]\n",
      "Ep:1/5,Itr:390/469, Averages:[0.23674], Metrics:[0.9375, 0.93609, 0.93896, 0.93322]\n",
      "Ep:1/5,Itr:391/469, Averages:[0.15213], Metrics:[0.94531, 0.94755, 0.94988, 0.94524]\n",
      "Ep:1/5,Itr:392/469, Averages:[0.23582], Metrics:[0.92188, 0.92336, 0.92483, 0.9219]\n",
      "Ep:1/5,Itr:393/469, Averages:[0.17773], Metrics:[0.9375, 0.939, 0.92774, 0.95054]\n",
      "Ep:1/5,Itr:394/469, Averages:[0.0819], Metrics:[0.98438, 0.98513, 0.98462, 0.98565]\n",
      "Ep:1/5,Itr:395/469, Averages:[0.28701], Metrics:[0.91406, 0.91733, 0.91162, 0.92311]\n",
      "Ep:1/5,Itr:396/469, Averages:[0.18216], Metrics:[0.95312, 0.95113, 0.94917, 0.95309]\n",
      "Ep:1/5,Itr:397/469, Averages:[0.22939], Metrics:[0.92188, 0.92474, 0.92663, 0.92286]\n",
      "Ep:1/5,Itr:398/469, Averages:[0.13355], Metrics:[0.96094, 0.96, 0.9666, 0.95349]\n",
      "Ep:1/5,Itr:399/469, Averages:[0.13127], Metrics:[0.96094, 0.95675, 0.95274, 0.96079]\n",
      "Ep:1/5,Itr:400/469, Averages:[0.11954], Metrics:[0.96875, 0.96553, 0.95481, 0.97649]\n",
      "Ep:1/5,Itr:401/469, Averages:[0.21105], Metrics:[0.91406, 0.91473, 0.90909, 0.92044]\n",
      "Ep:1/5,Itr:402/469, Averages:[0.12931], Metrics:[0.98438, 0.98101, 0.98167, 0.98036]\n",
      "Ep:1/5,Itr:403/469, Averages:[0.17083], Metrics:[0.96875, 0.97025, 0.97222, 0.96828]\n",
      "Ep:1/5,Itr:404/469, Averages:[0.27776], Metrics:[0.95312, 0.95541, 0.96089, 0.94999]\n",
      "Ep:1/5,Itr:405/469, Averages:[0.22293], Metrics:[0.91406, 0.9164, 0.91035, 0.92253]\n",
      "Ep:1/5,Itr:406/469, Averages:[0.2235], Metrics:[0.95312, 0.95356, 0.95347, 0.95365]\n",
      "Ep:1/5,Itr:407/469, Averages:[0.19216], Metrics:[0.9375, 0.93888, 0.9407, 0.93708]\n",
      "Ep:1/5,Itr:408/469, Averages:[0.09853], Metrics:[0.96875, 0.96355, 0.96342, 0.96369]\n",
      "Ep:1/5,Itr:409/469, Averages:[0.2694], Metrics:[0.91406, 0.91407, 0.90093, 0.9276]\n",
      "Ep:1/5,Itr:410/469, Averages:[0.14085], Metrics:[0.96094, 0.9538, 0.95552, 0.95208]\n",
      "Ep:1/5,Itr:411/469, Averages:[0.17911], Metrics:[0.94531, 0.95009, 0.94824, 0.95195]\n",
      "Ep:1/5,Itr:412/469, Averages:[0.13858], Metrics:[0.94531, 0.94895, 0.95182, 0.94611]\n",
      "Ep:1/5,Itr:413/469, Averages:[0.17169], Metrics:[0.92188, 0.9208, 0.92281, 0.9188]\n",
      "Ep:1/5,Itr:414/469, Averages:[0.19843], Metrics:[0.9375, 0.93373, 0.93541, 0.93206]\n",
      "Ep:1/5,Itr:415/469, Averages:[0.09915], Metrics:[0.97656, 0.97831, 0.9771, 0.97952]\n",
      "Ep:1/5,Itr:416/469, Averages:[0.13159], Metrics:[0.94531, 0.94788, 0.95018, 0.9456]\n",
      "Ep:1/5,Itr:417/469, Averages:[0.10823], Metrics:[0.96875, 0.97229, 0.97426, 0.97033]\n",
      "Ep:1/5,Itr:418/469, Averages:[0.2793], Metrics:[0.92969, 0.92246, 0.92194, 0.92297]\n",
      "Ep:1/5,Itr:419/469, Averages:[0.20294], Metrics:[0.96094, 0.96131, 0.96398, 0.95865]\n",
      "Ep:1/5,Itr:420/469, Averages:[0.07285], Metrics:[0.97656, 0.97797, 0.97697, 0.97897]\n",
      "Ep:1/5,Itr:421/469, Averages:[0.10357], Metrics:[0.95312, 0.95792, 0.95578, 0.96008]\n",
      "Ep:1/5,Itr:422/469, Averages:[0.19233], Metrics:[0.94531, 0.92822, 0.93528, 0.92126]\n",
      "Ep:1/5,Itr:423/469, Averages:[0.14002], Metrics:[0.94531, 0.9505, 0.95158, 0.94943]\n",
      "Ep:1/5,Itr:424/469, Averages:[0.31215], Metrics:[0.92969, 0.93189, 0.93173, 0.93204]\n",
      "Ep:1/5,Itr:425/469, Averages:[0.1708], Metrics:[0.95312, 0.9512, 0.94927, 0.95313]\n",
      "Ep:1/5,Itr:426/469, Averages:[0.08832], Metrics:[0.98438, 0.98397, 0.98397, 0.98397]\n",
      "Ep:1/5,Itr:427/469, Averages:[0.16668], Metrics:[0.94531, 0.94027, 0.9421, 0.93845]\n",
      "Ep:1/5,Itr:428/469, Averages:[0.48988], Metrics:[0.92969, 0.93193, 0.93064, 0.93322]\n",
      "Ep:1/5,Itr:429/469, Averages:[0.22977], Metrics:[0.92188, 0.91788, 0.9214, 0.91438]\n",
      "Ep:1/5,Itr:430/469, Averages:[0.1118], Metrics:[0.95312, 0.95216, 0.95411, 0.95022]\n",
      "Ep:1/5,Itr:431/469, Averages:[0.12528], Metrics:[0.96094, 0.9591, 0.95606, 0.96216]\n",
      "Ep:1/5,Itr:432/469, Averages:[0.25606], Metrics:[0.9375, 0.93749, 0.93796, 0.93702]\n",
      "Ep:1/5,Itr:433/469, Averages:[0.09834], Metrics:[0.96875, 0.97043, 0.96722, 0.97365]\n",
      "Ep:1/5,Itr:434/469, Averages:[0.22229], Metrics:[0.95312, 0.94384, 0.94253, 0.94514]\n",
      "Ep:1/5,Itr:435/469, Averages:[0.13833], Metrics:[0.9375, 0.93223, 0.93575, 0.92874]\n",
      "Ep:1/5,Itr:436/469, Averages:[0.25253], Metrics:[0.91406, 0.91441, 0.91607, 0.91274]\n",
      "Ep:1/5,Itr:437/469, Averages:[0.16068], Metrics:[0.94531, 0.9484, 0.94356, 0.95329]\n",
      "Ep:1/5,Itr:438/469, Averages:[0.16187], Metrics:[0.92188, 0.91998, 0.9097, 0.93051]\n",
      "Ep:1/5,Itr:439/469, Averages:[0.1505], Metrics:[0.94531, 0.94366, 0.94489, 0.94244]\n",
      "Ep:1/5,Itr:440/469, Averages:[0.11053], Metrics:[0.98438, 0.98288, 0.98578, 0.98]\n",
      "Ep:1/5,Itr:441/469, Averages:[0.17539], Metrics:[0.9375, 0.94022, 0.94385, 0.93663]\n",
      "Ep:1/5,Itr:442/469, Averages:[0.16256], Metrics:[0.96094, 0.96263, 0.96641, 0.95889]\n",
      "Ep:1/5,Itr:443/469, Averages:[0.29356], Metrics:[0.88281, 0.8788, 0.89149, 0.86647]\n",
      "Ep:1/5,Itr:444/469, Averages:[0.19791], Metrics:[0.9375, 0.93529, 0.93406, 0.93652]\n",
      "Ep:1/5,Itr:445/469, Averages:[0.29585], Metrics:[0.89062, 0.88566, 0.88225, 0.88909]\n",
      "Ep:1/5,Itr:446/469, Averages:[0.14296], Metrics:[0.9375, 0.93783, 0.9411, 0.93457]\n",
      "Ep:1/5,Itr:447/469, Averages:[0.1868], Metrics:[0.9375, 0.93823, 0.93736, 0.9391]\n",
      "Ep:1/5,Itr:448/469, Averages:[0.14069], Metrics:[0.96094, 0.96044, 0.96024, 0.96063]\n",
      "Ep:1/5,Itr:449/469, Averages:[0.16688], Metrics:[0.94531, 0.9425, 0.93892, 0.9461]\n",
      "Ep:1/5,Itr:450/469, Averages:[0.09836], Metrics:[0.96094, 0.95832, 0.96184, 0.95482]\n",
      "Ep:1/5,Itr:451/469, Averages:[0.20517], Metrics:[0.95312, 0.95381, 0.95417, 0.95346]\n",
      "Ep:1/5,Itr:452/469, Averages:[0.14314], Metrics:[0.96094, 0.96216, 0.96476, 0.95957]\n",
      "Ep:1/5,Itr:453/469, Averages:[0.21027], Metrics:[0.9375, 0.94142, 0.95056, 0.93247]\n",
      "Ep:1/5,Itr:454/469, Averages:[0.26015], Metrics:[0.92188, 0.92264, 0.92422, 0.92107]\n",
      "Ep:1/5,Itr:455/469, Averages:[0.20725], Metrics:[0.9375, 0.93485, 0.92813, 0.94167]\n",
      "Ep:1/5,Itr:456/469, Averages:[0.10982], Metrics:[0.97656, 0.98024, 0.98119, 0.97928]\n",
      "Ep:1/5,Itr:457/469, Averages:[0.12982], Metrics:[0.96094, 0.96089, 0.96203, 0.95975]\n",
      "Ep:1/5,Itr:458/469, Averages:[0.22656], Metrics:[0.96094, 0.96049, 0.95886, 0.96214]\n",
      "Ep:1/5,Itr:459/469, Averages:[0.11655], Metrics:[0.96875, 0.97092, 0.97336, 0.9685]\n",
      "Ep:1/5,Itr:460/469, Averages:[0.18152], Metrics:[0.95312, 0.95466, 0.95618, 0.95314]\n",
      "Ep:1/5,Itr:461/469, Averages:[0.09612], Metrics:[0.97656, 0.97387, 0.96833, 0.97947]\n",
      "Ep:1/5,Itr:462/469, Averages:[0.12542], Metrics:[0.96094, 0.96187, 0.96021, 0.96353]\n",
      "Ep:1/5,Itr:463/469, Averages:[0.35213], Metrics:[0.89844, 0.88937, 0.88401, 0.8948]\n",
      "Ep:1/5,Itr:464/469, Averages:[0.22034], Metrics:[0.92969, 0.93323, 0.93444, 0.93202]\n",
      "Ep:1/5,Itr:465/469, Averages:[0.26308], Metrics:[0.95312, 0.94855, 0.93657, 0.96083]\n",
      "Ep:1/5,Itr:466/469, Averages:[0.10626], Metrics:[0.97656, 0.97614, 0.97747, 0.97482]\n",
      "Ep:1/5,Itr:467/469, Averages:[0.15192], Metrics:[0.94531, 0.93798, 0.93692, 0.93904]\n",
      "Ep:1/5,Itr:468/469, Averages:[0.16552], Metrics:[0.9375, 0.94122, 0.94215, 0.94028]\n",
      "Ep:1/5,Itr:469/469, Averages:[0.14894], Metrics:[0.94792, 0.95247, 0.95585, 0.94911]\n",
      "validation ...\n",
      " Itr:1/79, Averages:[0.0272], Metrics:[0.99219, 0.99468, 0.99412, 0.99524]\n",
      " Itr:2/79, Averages:[0.04229], Metrics:[0.99219, 0.99188, 0.99286, 0.99091]\n",
      " Itr:3/79, Averages:[0.07406], Metrics:[0.96875, 0.9662, 0.96558, 0.96682]\n",
      " Itr:4/79, Averages:[0.05312], Metrics:[0.98438, 0.98346, 0.98462, 0.98231]\n",
      " Itr:5/79, Averages:[0.07346], Metrics:[0.97656, 0.97547, 0.96836, 0.98268]\n",
      " Itr:6/79, Averages:[0.13475], Metrics:[0.95312, 0.95472, 0.95961, 0.94988]\n",
      " Itr:7/79, Averages:[0.03528], Metrics:[0.99219, 0.99028, 0.99167, 0.98889]\n",
      " Itr:8/79, Averages:[0.15987], Metrics:[0.95312, 0.95259, 0.95721, 0.94801]\n",
      " Itr:9/79, Averages:[0.07363], Metrics:[0.96875, 0.97199, 0.97314, 0.97083]\n",
      " Itr:10/79, Averages:[0.15375], Metrics:[0.95312, 0.9525, 0.95074, 0.95427]\n",
      " Itr:11/79, Averages:[0.09663], Metrics:[0.96094, 0.96074, 0.96179, 0.9597]\n",
      " Itr:12/79, Averages:[0.09435], Metrics:[0.96875, 0.96521, 0.9612, 0.96927]\n",
      " Itr:13/79, Averages:[0.07582], Metrics:[0.97656, 0.97733, 0.97799, 0.97667]\n",
      " Itr:14/79, Averages:[0.14549], Metrics:[0.96094, 0.96346, 0.96609, 0.96083]\n",
      " Itr:15/79, Averages:[0.08046], Metrics:[0.97656, 0.97474, 0.97683, 0.97266]\n",
      " Itr:16/79, Averages:[0.09261], Metrics:[0.96875, 0.96587, 0.96842, 0.96333]\n",
      " Itr:17/79, Averages:[0.16064], Metrics:[0.9375, 0.9333, 0.9388, 0.92788]\n",
      " Itr:18/79, Averages:[0.10371], Metrics:[0.96094, 0.96055, 0.96424, 0.95688]\n",
      " Itr:19/79, Averages:[0.09023], Metrics:[0.96875, 0.96763, 0.9619, 0.97342]\n",
      " Itr:20/79, Averages:[0.09657], Metrics:[0.97656, 0.97895, 0.97772, 0.98018]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " Itr:21/79, Averages:[0.09337], Metrics:[0.98438, 0.98337, 0.98444, 0.98231]\n",
      " Itr:22/79, Averages:[0.03755], Metrics:[0.99219, 0.9873, 0.98571, 0.98889]\n",
      " Itr:23/79, Averages:[0.11812], Metrics:[0.96875, 0.96803, 0.97104, 0.96503]\n",
      " Itr:24/79, Averages:[0.08618], Metrics:[0.95312, 0.95561, 0.95636, 0.95487]\n",
      " Itr:25/79, Averages:[0.02976], Metrics:[0.99219, 0.99199, 0.99167, 0.99231]\n",
      " Itr:26/79, Averages:[0.05422], Metrics:[0.97656, 0.97042, 0.97786, 0.9631]\n",
      " Itr:27/79, Averages:[0.02166], Metrics:[0.99219, 0.99236, 0.99474, 0.99]\n",
      " Itr:28/79, Averages:[0.14174], Metrics:[0.96875, 0.96839, 0.96836, 0.96841]\n",
      " Itr:29/79, Averages:[0.05855], Metrics:[0.99219, 0.98875, 0.99, 0.9875]\n",
      " Itr:30/79, Averages:[0.19542], Metrics:[0.94531, 0.94749, 0.94762, 0.94736]\n",
      " Itr:31/79, Averages:[0.1232], Metrics:[0.96094, 0.95736, 0.96099, 0.95375]\n",
      " Itr:32/79, Averages:[0.09487], Metrics:[0.96094, 0.9575, 0.96532, 0.94981]\n",
      " Itr:33/79, Averages:[0.05595], Metrics:[0.98438, 0.98527, 0.98755, 0.98301]\n",
      " Itr:34/79, Averages:[0.08406], Metrics:[0.95312, 0.94696, 0.95431, 0.93972]\n",
      " Itr:35/79, Averages:[0.04738], Metrics:[0.99219, 0.99233, 0.99091, 0.99375]\n",
      " Itr:36/79, Averages:[0.15868], Metrics:[0.94531, 0.94532, 0.9467, 0.94394]\n",
      " Itr:37/79, Averages:[0.05004], Metrics:[0.97656, 0.97437, 0.97708, 0.97167]\n",
      " Itr:38/79, Averages:[0.08499], Metrics:[0.96875, 0.97178, 0.97184, 0.97173]\n",
      " Itr:39/79, Averages:[0.08947], Metrics:[0.96875, 0.96484, 0.96286, 0.96683]\n",
      " Itr:40/79, Averages:[0.00901], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:41/79, Averages:[0.00654], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:42/79, Averages:[0.00925], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:43/79, Averages:[0.00087], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:44/79, Averages:[0.01776], Metrics:[0.99219, 0.99226, 0.99167, 0.99286]\n",
      " Itr:45/79, Averages:[0.08095], Metrics:[0.96875, 0.96941, 0.96914, 0.96969]\n",
      " Itr:46/79, Averages:[0.02627], Metrics:[0.99219, 0.99199, 0.99167, 0.99231]\n",
      " Itr:47/79, Averages:[0.10308], Metrics:[0.97656, 0.97429, 0.97231, 0.97628]\n",
      " Itr:48/79, Averages:[0.06796], Metrics:[0.99219, 0.99129, 0.99091, 0.99167]\n",
      " Itr:49/79, Averages:[0.06656], Metrics:[0.96875, 0.97472, 0.97333, 0.9761]\n",
      " Itr:50/79, Averages:[0.00184], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:51/79, Averages:[0.03942], Metrics:[0.99219, 0.99393, 0.99412, 0.99375]\n",
      " Itr:52/79, Averages:[0.14337], Metrics:[0.97656, 0.97775, 0.97802, 0.97747]\n",
      " Itr:53/79, Averages:[0.03037], Metrics:[0.98438, 0.98598, 0.98591, 0.98606]\n",
      " Itr:54/79, Averages:[0.00393], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:55/79, Averages:[0.00353], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:56/79, Averages:[0.01777], Metrics:[0.99219, 0.9931, 0.99286, 0.99333]\n",
      " Itr:57/79, Averages:[0.00594], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:58/79, Averages:[0.00234], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:59/79, Averages:[0.01256], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:60/79, Averages:[0.0065], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:61/79, Averages:[0.01595], Metrics:[0.99219, 0.99289, 0.99167, 0.99412]\n",
      " Itr:62/79, Averages:[0.04187], Metrics:[0.99219, 0.99143, 0.99, 0.99286]\n",
      " Itr:63/79, Averages:[0.00552], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:64/79, Averages:[0.03721], Metrics:[0.98438, 0.98566, 0.98667, 0.98466]\n",
      " Itr:65/79, Averages:[0.01572], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:66/79, Averages:[0.02696], Metrics:[0.99219, 0.99282, 0.99231, 0.99333]\n",
      " Itr:67/79, Averages:[0.00773], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:68/79, Averages:[0.00096], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:69/79, Averages:[0.00046], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:70/79, Averages:[0.00283], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:71/79, Averages:[0.28496], Metrics:[0.95312, 0.96007, 0.95902, 0.96111]\n",
      " Itr:72/79, Averages:[0.00116], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:73/79, Averages:[0.00144], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:74/79, Averages:[0.00139], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:75/79, Averages:[0.01809], Metrics:[0.99219, 0.9931, 0.99286, 0.99333]\n",
      " Itr:76/79, Averages:[0.08161], Metrics:[0.97656, 0.97762, 0.97633, 0.97891]\n",
      " Itr:77/79, Averages:[0.16592], Metrics:[0.96875, 0.96931, 0.96795, 0.97067]\n",
      " Itr:78/79, Averages:[0.07557], Metrics:[0.97656, 0.97233, 0.97088, 0.97378]\n",
      " Itr:79/79, Averages:[0.00161], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      "[SUCCESS]! experiment validation Averages:[0.06311], Metrics:[0.9812, 0.98112, 0.98127, 0.98097]\n",
      "[SUCCESS]!  *** Best Model Saved!!! *** : 0.9811182449522878\n",
      "Ep:2/5,Itr:1/469, Averages:[0.09634], Metrics:[0.97656, 0.97221, 0.96912, 0.97531]\n",
      "Ep:2/5,Itr:2/469, Averages:[0.09266], Metrics:[0.96875, 0.97122, 0.97509, 0.96738]\n",
      "Ep:2/5,Itr:3/469, Averages:[0.16874], Metrics:[0.96094, 0.96454, 0.96839, 0.96071]\n",
      "Ep:2/5,Itr:4/469, Averages:[0.13404], Metrics:[0.96094, 0.96077, 0.96129, 0.96025]\n",
      "Ep:2/5,Itr:6/469, Averages:[0.15588], Metrics:[0.96094, 0.96094, 0.96336, 0.95853]\n",
      "Ep:2/5,Itr:8/469, Averages:[0.11485], Metrics:[0.97656, 0.97516, 0.97711, 0.97321]\n",
      "Ep:2/5,Itr:10/469, Averages:[0.11776], Metrics:[0.95703, 0.95732, 0.95846, 0.95619]\n",
      "Ep:2/5,Itr:12/469, Averages:[0.08012], Metrics:[0.98047, 0.97857, 0.9794, 0.97773]\n",
      "Ep:2/5,Itr:14/469, Averages:[0.11197], Metrics:[0.96094, 0.95948, 0.95992, 0.95905]\n",
      "Ep:2/5,Itr:16/469, Averages:[0.17321], Metrics:[0.96484, 0.96595, 0.96286, 0.96906]\n",
      "Ep:2/5,Itr:18/469, Averages:[0.10067], Metrics:[0.96875, 0.96849, 0.9682, 0.96879]\n",
      "Ep:2/5,Itr:21/469, Averages:[0.19109], Metrics:[0.9349, 0.9357, 0.93603, 0.93536]\n",
      "Ep:2/5,Itr:24/469, Averages:[0.17229], Metrics:[0.94792, 0.94802, 0.94954, 0.94651]\n",
      "Ep:2/5,Itr:27/469, Averages:[0.21069], Metrics:[0.94792, 0.94882, 0.94672, 0.95093]\n",
      "Ep:2/5,Itr:30/469, Averages:[0.22858], Metrics:[0.94531, 0.94615, 0.9453, 0.94701]\n",
      "Ep:2/5,Itr:33/469, Averages:[0.15025], Metrics:[0.95052, 0.95192, 0.9535, 0.95035]\n",
      "Ep:2/5,Itr:36/469, Averages:[0.14646], Metrics:[0.94531, 0.94658, 0.946, 0.94716]\n",
      "Ep:2/5,Itr:39/469, Averages:[0.11492], Metrics:[0.95833, 0.95853, 0.9599, 0.95716]\n",
      "Ep:2/5,Itr:42/469, Averages:[0.15814], Metrics:[0.95312, 0.95136, 0.95155, 0.95117]\n",
      "Ep:2/5,Itr:45/469, Averages:[0.14763], Metrics:[0.95052, 0.94874, 0.94851, 0.94897]\n",
      "Ep:2/5,Itr:48/469, Averages:[0.17435], Metrics:[0.95312, 0.95147, 0.95185, 0.9511]\n",
      "Ep:2/5,Itr:51/469, Averages:[0.10953], Metrics:[0.96354, 0.96362, 0.96381, 0.96342]\n",
      "Ep:2/5,Itr:54/469, Averages:[0.06615], Metrics:[0.97396, 0.97398, 0.97323, 0.97473]\n",
      "Ep:2/5,Itr:57/469, Averages:[0.16032], Metrics:[0.96354, 0.96411, 0.96374, 0.96448]\n",
      "Ep:2/5,Itr:60/469, Averages:[0.13755], Metrics:[0.96094, 0.95858, 0.95868, 0.95849]\n",
      "Ep:2/5,Itr:63/469, Averages:[0.16086], Metrics:[0.95312, 0.95277, 0.95386, 0.95169]\n",
      "Ep:2/5,Itr:66/469, Averages:[0.1267], Metrics:[0.96354, 0.96288, 0.9637, 0.96205]\n",
      "Ep:2/5,Itr:69/469, Averages:[0.16063], Metrics:[0.95312, 0.95052, 0.94958, 0.95147]\n",
      "Ep:2/5,Itr:72/469, Averages:[0.15288], Metrics:[0.94271, 0.94071, 0.93729, 0.94416]\n",
      "Ep:2/5,Itr:75/469, Averages:[0.15965], Metrics:[0.95052, 0.94917, 0.95035, 0.948]\n",
      "Ep:2/5,Itr:76/469, Averages:[0.11042], Metrics:[0.95312, 0.94279, 0.94065, 0.94494]\n",
      "Ep:2/5,Itr:80/469, Averages:[0.11658], Metrics:[0.97656, 0.9746, 0.97497, 0.97424]\n",
      "Ep:2/5,Itr:84/469, Averages:[0.16429], Metrics:[0.95703, 0.95638, 0.95616, 0.9566]\n",
      "Ep:2/5,Itr:88/469, Averages:[0.15828], Metrics:[0.95703, 0.9576, 0.95713, 0.95807]\n",
      "Ep:2/5,Itr:92/469, Averages:[0.14103], Metrics:[0.95898, 0.95947, 0.95937, 0.95957]\n",
      "Ep:2/5,Itr:96/469, Averages:[0.14164], Metrics:[0.96094, 0.96049, 0.96006, 0.96093]\n",
      "Ep:2/5,Itr:100/469, Averages:[0.13081], Metrics:[0.95703, 0.95579, 0.95555, 0.95603]\n",
      "Ep:2/5,Itr:104/469, Averages:[0.1482], Metrics:[0.9707, 0.96948, 0.97061, 0.96836]\n",
      "Ep:2/5,Itr:108/469, Averages:[0.13401], Metrics:[0.95898, 0.95916, 0.95812, 0.9602]\n",
      "Ep:2/5,Itr:112/469, Averages:[0.18048], Metrics:[0.95312, 0.95054, 0.94907, 0.95201]\n",
      "Ep:2/5,Itr:116/469, Averages:[0.16174], Metrics:[0.95703, 0.95677, 0.95753, 0.95602]\n",
      "Ep:2/5,Itr:120/469, Averages:[0.13255], Metrics:[0.96484, 0.96493, 0.96513, 0.96473]\n",
      "Ep:2/5,Itr:124/469, Averages:[0.09479], Metrics:[0.97461, 0.97391, 0.97339, 0.97444]\n",
      "Ep:2/5,Itr:128/469, Averages:[0.14695], Metrics:[0.95703, 0.95507, 0.95502, 0.95512]\n",
      "Ep:2/5,Itr:132/469, Averages:[0.19103], Metrics:[0.94922, 0.95078, 0.95089, 0.95067]\n",
      "Ep:2/5,Itr:136/469, Averages:[0.09896], Metrics:[0.96289, 0.96278, 0.96278, 0.96278]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ep:2/5,Itr:140/469, Averages:[0.13673], Metrics:[0.96094, 0.96027, 0.96053, 0.96]\n",
      "Ep:2/5,Itr:144/469, Averages:[0.1361], Metrics:[0.96289, 0.96139, 0.9603, 0.96248]\n",
      "Ep:2/5,Itr:148/469, Averages:[0.11312], Metrics:[0.97266, 0.97291, 0.97282, 0.973]\n",
      "Ep:2/5,Itr:152/469, Averages:[0.15578], Metrics:[0.94922, 0.94988, 0.94969, 0.95006]\n",
      "Ep:2/5,Itr:156/469, Averages:[0.18078], Metrics:[0.94727, 0.94713, 0.9468, 0.94746]\n",
      "Ep:2/5,Itr:160/469, Averages:[0.16733], Metrics:[0.94922, 0.95014, 0.94984, 0.95045]\n",
      "Ep:2/5,Itr:164/469, Averages:[0.16046], Metrics:[0.94922, 0.94948, 0.94948, 0.94948]\n",
      "Ep:2/5,Itr:168/469, Averages:[0.13905], Metrics:[0.95312, 0.95154, 0.95164, 0.95144]\n",
      "Ep:2/5,Itr:172/469, Averages:[0.12592], Metrics:[0.96094, 0.96107, 0.96057, 0.96157]\n",
      "Ep:2/5,Itr:176/469, Averages:[0.13386], Metrics:[0.95508, 0.95527, 0.95488, 0.95567]\n",
      "Ep:2/5,Itr:180/469, Averages:[0.12832], Metrics:[0.96094, 0.96114, 0.96153, 0.96075]\n",
      "Ep:2/5,Itr:184/469, Averages:[0.16827], Metrics:[0.94922, 0.94991, 0.94964, 0.95018]\n",
      "Ep:2/5,Itr:188/469, Averages:[0.21785], Metrics:[0.94141, 0.94153, 0.94184, 0.94122]\n",
      "Ep:2/5,Itr:192/469, Averages:[0.07788], Metrics:[0.98242, 0.98283, 0.98304, 0.98262]\n",
      "Ep:2/5,Itr:196/469, Averages:[0.20635], Metrics:[0.93164, 0.93236, 0.93265, 0.93208]\n",
      "Ep:2/5,Itr:200/469, Averages:[0.12966], Metrics:[0.96875, 0.96891, 0.96996, 0.96786]\n",
      "Ep:2/5,Itr:204/469, Averages:[0.17748], Metrics:[0.95117, 0.95124, 0.95305, 0.94944]\n",
      "Ep:2/5,Itr:208/469, Averages:[0.22429], Metrics:[0.94727, 0.94663, 0.94666, 0.94659]\n",
      "Ep:2/5,Itr:212/469, Averages:[0.1417], Metrics:[0.96289, 0.96349, 0.96367, 0.96331]\n",
      "Ep:2/5,Itr:216/469, Averages:[0.15286], Metrics:[0.95117, 0.94945, 0.94883, 0.95007]\n",
      "Ep:2/5,Itr:220/469, Averages:[0.1446], Metrics:[0.95703, 0.95754, 0.95753, 0.95755]\n",
      "Ep:2/5,Itr:224/469, Averages:[0.14132], Metrics:[0.94141, 0.93909, 0.93671, 0.94149]\n",
      "Ep:2/5,Itr:228/469, Averages:[0.10983], Metrics:[0.95898, 0.95964, 0.95885, 0.96043]\n",
      "Ep:2/5,Itr:232/469, Averages:[0.13908], Metrics:[0.96484, 0.96418, 0.96407, 0.96428]\n",
      "Ep:2/5,Itr:236/469, Averages:[0.10962], Metrics:[0.96094, 0.95899, 0.95885, 0.95913]\n",
      "Ep:2/5,Itr:240/469, Averages:[0.11631], Metrics:[0.96484, 0.96392, 0.96556, 0.9623]\n",
      "Ep:2/5,Itr:244/469, Averages:[0.18812], Metrics:[0.93945, 0.9411, 0.9393, 0.94291]\n",
      "Ep:2/5,Itr:248/469, Averages:[0.12718], Metrics:[0.96094, 0.96002, 0.95918, 0.96087]\n",
      "Ep:2/5,Itr:252/469, Averages:[0.16012], Metrics:[0.94922, 0.94918, 0.94965, 0.94871]\n",
      "Ep:2/5,Itr:256/469, Averages:[0.19538], Metrics:[0.94141, 0.9417, 0.93939, 0.94403]\n",
      "Ep:2/5,Itr:260/469, Averages:[0.11008], Metrics:[0.96094, 0.96082, 0.96028, 0.96136]\n",
      "Ep:2/5,Itr:264/469, Averages:[0.18341], Metrics:[0.94336, 0.9432, 0.94399, 0.94241]\n",
      "Ep:2/5,Itr:268/469, Averages:[0.13665], Metrics:[0.95703, 0.95561, 0.95548, 0.95574]\n",
      "Ep:2/5,Itr:272/469, Averages:[0.08733], Metrics:[0.97461, 0.97427, 0.97388, 0.97466]\n",
      "Ep:2/5,Itr:276/469, Averages:[0.11791], Metrics:[0.9707, 0.96827, 0.9684, 0.96814]\n",
      "Ep:2/5,Itr:280/469, Averages:[0.13069], Metrics:[0.9668, 0.96469, 0.96274, 0.96665]\n",
      "Ep:2/5,Itr:284/469, Averages:[0.13153], Metrics:[0.96094, 0.95901, 0.95815, 0.95987]\n",
      "Ep:2/5,Itr:288/469, Averages:[0.14055], Metrics:[0.96875, 0.96868, 0.96982, 0.96754]\n",
      "Ep:2/5,Itr:292/469, Averages:[0.11341], Metrics:[0.96484, 0.96421, 0.96207, 0.96636]\n",
      "Ep:2/5,Itr:296/469, Averages:[0.10444], Metrics:[0.97461, 0.97392, 0.97387, 0.97397]\n",
      "Ep:2/5,Itr:300/469, Averages:[0.13889], Metrics:[0.95508, 0.95377, 0.95433, 0.95322]\n",
      "Ep:2/5,Itr:304/469, Averages:[0.07298], Metrics:[0.98047, 0.97976, 0.98064, 0.97888]\n",
      "Ep:2/5,Itr:308/469, Averages:[0.14334], Metrics:[0.95312, 0.95303, 0.95309, 0.95297]\n",
      "Ep:2/5,Itr:312/469, Averages:[0.12207], Metrics:[0.96484, 0.96593, 0.96493, 0.96692]\n",
      "Ep:2/5,Itr:316/469, Averages:[0.10368], Metrics:[0.97461, 0.97449, 0.97445, 0.97453]\n",
      "Ep:2/5,Itr:320/469, Averages:[0.18964], Metrics:[0.93359, 0.933, 0.93297, 0.93302]\n",
      "Ep:2/5,Itr:325/469, Averages:[0.08153], Metrics:[0.97813, 0.97781, 0.97772, 0.9779]\n",
      "Ep:2/5,Itr:330/469, Averages:[0.14924], Metrics:[0.95312, 0.95326, 0.95408, 0.95244]\n",
      "Ep:2/5,Itr:335/469, Averages:[0.10711], Metrics:[0.96094, 0.96071, 0.96227, 0.95916]\n",
      "Ep:2/5,Itr:340/469, Averages:[0.1464], Metrics:[0.96406, 0.96432, 0.96494, 0.96369]\n",
      "Ep:2/5,Itr:345/469, Averages:[0.1421], Metrics:[0.96094, 0.96051, 0.95937, 0.96166]\n",
      "Ep:2/5,Itr:350/469, Averages:[0.10245], Metrics:[0.96719, 0.96775, 0.96727, 0.96823]\n",
      "Ep:2/5,Itr:355/469, Averages:[0.12286], Metrics:[0.96406, 0.96451, 0.96401, 0.965]\n",
      "Ep:2/5,Itr:360/469, Averages:[0.15085], Metrics:[0.95781, 0.95854, 0.95861, 0.95847]\n",
      "Ep:2/5,Itr:365/469, Averages:[0.13317], Metrics:[0.96562, 0.96524, 0.96492, 0.96556]\n",
      "Ep:2/5,Itr:370/469, Averages:[0.08079], Metrics:[0.97344, 0.97414, 0.97367, 0.97462]\n",
      "Ep:2/5,Itr:375/469, Averages:[0.1312], Metrics:[0.95469, 0.95435, 0.95618, 0.95252]\n",
      "Ep:2/5,Itr:380/469, Averages:[0.12232], Metrics:[0.96406, 0.96444, 0.96422, 0.96465]\n",
      "Ep:2/5,Itr:385/469, Averages:[0.13126], Metrics:[0.96562, 0.96524, 0.96556, 0.96492]\n",
      "Ep:2/5,Itr:390/469, Averages:[0.13318], Metrics:[0.95312, 0.9541, 0.95275, 0.95545]\n",
      "Ep:2/5,Itr:395/469, Averages:[0.12392], Metrics:[0.96719, 0.9679, 0.96815, 0.96765]\n",
      "Ep:2/5,Itr:400/469, Averages:[0.08015], Metrics:[0.97656, 0.97642, 0.97589, 0.97695]\n",
      "Ep:2/5,Itr:405/469, Averages:[0.12033], Metrics:[0.97344, 0.97301, 0.97341, 0.97261]\n",
      "Ep:2/5,Itr:410/469, Averages:[0.1172], Metrics:[0.96406, 0.9643, 0.96399, 0.96461]\n",
      "Ep:2/5,Itr:415/469, Averages:[0.1117], Metrics:[0.97031, 0.97151, 0.97088, 0.97215]\n",
      "Ep:2/5,Itr:420/469, Averages:[0.08322], Metrics:[0.97188, 0.9721, 0.972, 0.9722]\n",
      "Ep:2/5,Itr:425/469, Averages:[0.08907], Metrics:[0.97813, 0.97794, 0.97787, 0.97802]\n",
      "Ep:2/5,Itr:430/469, Averages:[0.14684], Metrics:[0.95156, 0.95259, 0.95099, 0.95419]\n",
      "Ep:2/5,Itr:435/469, Averages:[0.11821], Metrics:[0.97031, 0.97083, 0.97028, 0.97139]\n",
      "Ep:2/5,Itr:440/469, Averages:[0.11909], Metrics:[0.95937, 0.95929, 0.95867, 0.9599]\n",
      "Ep:2/5,Itr:445/469, Averages:[0.12608], Metrics:[0.96406, 0.96319, 0.96279, 0.96359]\n",
      "Ep:2/5,Itr:450/469, Averages:[0.08481], Metrics:[0.97344, 0.97242, 0.97387, 0.97099]\n",
      "Ep:2/5,Itr:455/469, Averages:[0.12599], Metrics:[0.96562, 0.96568, 0.96599, 0.96537]\n",
      "Ep:2/5,Itr:460/469, Averages:[0.12331], Metrics:[0.96406, 0.96443, 0.96441, 0.96445]\n",
      "Ep:2/5,Itr:465/469, Averages:[0.13157], Metrics:[0.95937, 0.95913, 0.95887, 0.95939]\n",
      "Ep:2/5,Itr:469/469, Averages:[0.11139], Metrics:[0.97083, 0.97058, 0.97057, 0.97059]\n",
      "validation ...\n",
      " Itr:1/79, Averages:[0.0106], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:2/79, Averages:[0.02962], Metrics:[0.99219, 0.99188, 0.99286, 0.99091]\n",
      " Itr:3/79, Averages:[0.05694], Metrics:[0.96875, 0.9662, 0.96558, 0.96682]\n",
      " Itr:4/79, Averages:[0.02906], Metrics:[0.97656, 0.97604, 0.97747, 0.97462]\n",
      " Itr:6/79, Averages:[0.05166], Metrics:[0.97656, 0.97859, 0.97542, 0.98178]\n",
      " Itr:8/79, Averages:[0.14741], Metrics:[0.96094, 0.96054, 0.96309, 0.95801]\n",
      " Itr:10/79, Averages:[0.15202], Metrics:[0.9375, 0.93965, 0.93567, 0.94366]\n",
      " Itr:12/79, Averages:[0.08393], Metrics:[0.96875, 0.96572, 0.9635, 0.96795]\n",
      " Itr:14/79, Averages:[0.18336], Metrics:[0.94531, 0.94348, 0.94309, 0.94387]\n",
      " Itr:16/79, Averages:[0.04717], Metrics:[0.97656, 0.97842, 0.9812, 0.97565]\n",
      " Itr:18/79, Averages:[0.08588], Metrics:[0.96875, 0.96877, 0.97091, 0.96664]\n",
      " Itr:21/79, Averages:[0.11555], Metrics:[0.98438, 0.98236, 0.98, 0.98474]\n",
      " Itr:24/79, Averages:[0.07285], Metrics:[0.97656, 0.97922, 0.97905, 0.97939]\n",
      " Itr:27/79, Averages:[0.04308], Metrics:[0.99219, 0.99295, 0.99091, 0.995]\n",
      " Itr:30/79, Averages:[0.12435], Metrics:[0.95312, 0.95345, 0.95345, 0.95345]\n",
      " Itr:33/79, Averages:[0.04209], Metrics:[0.97656, 0.98317, 0.98423, 0.98212]\n",
      " Itr:36/79, Averages:[0.14957], Metrics:[0.96094, 0.9635, 0.96098, 0.96602]\n",
      " Itr:39/79, Averages:[0.11158], Metrics:[0.96875, 0.96761, 0.965, 0.97024]\n",
      " Itr:42/79, Averages:[0.01852], Metrics:[0.98438, 0.98268, 0.98083, 0.98452]\n",
      " Itr:45/79, Averages:[0.02325], Metrics:[0.99219, 0.99286, 0.99286, 0.99286]\n",
      " Itr:48/79, Averages:[0.03287], Metrics:[0.99219, 0.99212, 0.99091, 0.99333]\n",
      " Itr:51/79, Averages:[0.02104], Metrics:[0.99219, 0.99393, 0.99412, 0.99375]\n",
      " Itr:54/79, Averages:[0.00335], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:57/79, Averages:[0.01455], Metrics:[0.99219, 0.99282, 0.99333, 0.99231]\n",
      " Itr:60/79, Averages:[0.00226], Metrics:[1.0, 1.0, 1.0, 1.0]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " Itr:63/79, Averages:[0.02539], Metrics:[0.99219, 0.99226, 0.99167, 0.99286]\n",
      " Itr:66/79, Averages:[0.04825], Metrics:[0.98438, 0.98327, 0.98231, 0.98424]\n",
      " Itr:69/79, Averages:[0.00011], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:72/79, Averages:[0.00032], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:75/79, Averages:[0.0104], Metrics:[0.99219, 0.99188, 0.99286, 0.99091]\n",
      " Itr:76/79, Averages:[0.11455], Metrics:[0.96094, 0.96311, 0.95814, 0.96814]\n",
      "[SUCCESS]! experiment validation Averages:[0.05177], Metrics:[0.9841, 0.98411, 0.98413, 0.9841]\n",
      "[SUCCESS]!  *** Best Model Saved!!! *** : 0.9841129687961727\n",
      "Ep:3/5,Itr:1/469, Averages:[0.09829], Metrics:[0.96875, 0.96841, 0.97378, 0.96309]\n",
      "Ep:3/5,Itr:2/469, Averages:[0.13452], Metrics:[0.96094, 0.96746, 0.96944, 0.96549]\n",
      "Ep:3/5,Itr:4/469, Averages:[0.13735], Metrics:[0.94531, 0.94342, 0.94272, 0.94412]\n",
      "Ep:3/5,Itr:6/469, Averages:[0.06422], Metrics:[0.97656, 0.97633, 0.97551, 0.97715]\n",
      "Ep:3/5,Itr:9/469, Averages:[0.09757], Metrics:[0.97396, 0.97244, 0.97109, 0.9738]\n",
      "Ep:3/5,Itr:12/469, Averages:[0.07256], Metrics:[0.97396, 0.97416, 0.9739, 0.97443]\n",
      "Ep:3/5,Itr:15/469, Averages:[0.14167], Metrics:[0.96615, 0.96547, 0.96346, 0.96749]\n",
      "Ep:3/5,Itr:16/469, Averages:[0.06893], Metrics:[0.98438, 0.98313, 0.98444, 0.98182]\n",
      "Ep:3/5,Itr:20/469, Averages:[0.14917], Metrics:[0.95508, 0.95499, 0.95362, 0.95638]\n",
      "Ep:3/5,Itr:24/469, Averages:[0.1269], Metrics:[0.95898, 0.95924, 0.9606, 0.95787]\n",
      "Ep:3/5,Itr:28/469, Averages:[0.11228], Metrics:[0.96875, 0.96933, 0.96957, 0.96909]\n",
      "Ep:3/5,Itr:32/469, Averages:[0.11521], Metrics:[0.9707, 0.97019, 0.9693, 0.97109]\n",
      "Ep:3/5,Itr:36/469, Averages:[0.08014], Metrics:[0.97656, 0.97711, 0.97782, 0.97639]\n",
      "Ep:3/5,Itr:40/469, Averages:[0.08765], Metrics:[0.96875, 0.96727, 0.96841, 0.96614]\n",
      "Ep:3/5,Itr:45/469, Averages:[0.12047], Metrics:[0.96562, 0.9654, 0.96522, 0.96559]\n",
      "Ep:3/5,Itr:50/469, Averages:[0.0971], Metrics:[0.97813, 0.9773, 0.97747, 0.97713]\n",
      "Ep:3/5,Itr:55/469, Averages:[0.08167], Metrics:[0.97188, 0.9707, 0.97105, 0.97036]\n",
      "Ep:3/5,Itr:60/469, Averages:[0.14803], Metrics:[0.95156, 0.95, 0.95009, 0.94991]\n",
      "Ep:3/5,Itr:65/469, Averages:[0.10981], Metrics:[0.97656, 0.97634, 0.97514, 0.97754]\n",
      "Ep:3/5,Itr:70/469, Averages:[0.08662], Metrics:[0.96719, 0.96892, 0.96837, 0.96947]\n",
      "Ep:3/5,Itr:75/469, Averages:[0.12858], Metrics:[0.96562, 0.96526, 0.96522, 0.96531]\n",
      "Ep:3/5,Itr:80/469, Averages:[0.10293], Metrics:[0.9625, 0.96309, 0.9631, 0.96309]\n",
      "Ep:3/5,Itr:85/469, Averages:[0.0904], Metrics:[0.97656, 0.97638, 0.97649, 0.97628]\n",
      "Ep:3/5,Itr:90/469, Averages:[0.06337], Metrics:[0.98281, 0.98283, 0.98281, 0.98285]\n",
      "Ep:3/5,Itr:96/469, Averages:[0.14385], Metrics:[0.96354, 0.96227, 0.96263, 0.96192]\n",
      "Ep:3/5,Itr:102/469, Averages:[0.10801], Metrics:[0.97005, 0.96849, 0.96876, 0.96822]\n",
      "Ep:3/5,Itr:108/469, Averages:[0.09433], Metrics:[0.97266, 0.97156, 0.97202, 0.97109]\n",
      "Ep:3/5,Itr:114/469, Averages:[0.09541], Metrics:[0.97135, 0.97088, 0.97044, 0.97131]\n",
      "Ep:3/5,Itr:120/469, Averages:[0.10947], Metrics:[0.97005, 0.96968, 0.96951, 0.96986]\n",
      "Ep:3/5,Itr:126/469, Averages:[0.11014], Metrics:[0.95833, 0.95936, 0.96022, 0.95851]\n",
      "Ep:3/5,Itr:132/469, Averages:[0.11318], Metrics:[0.96615, 0.96623, 0.96633, 0.96614]\n",
      "Ep:3/5,Itr:138/469, Averages:[0.07546], Metrics:[0.97526, 0.9758, 0.9762, 0.9754]\n",
      "Ep:3/5,Itr:144/469, Averages:[0.10988], Metrics:[0.97526, 0.97522, 0.97576, 0.97467]\n",
      "Ep:3/5,Itr:150/469, Averages:[0.11819], Metrics:[0.97135, 0.97161, 0.9712, 0.97203]\n",
      "Ep:3/5,Itr:156/469, Averages:[0.11329], Metrics:[0.96745, 0.96599, 0.96648, 0.96551]\n",
      "Ep:3/5,Itr:162/469, Averages:[0.10892], Metrics:[0.96875, 0.96933, 0.97011, 0.96855]\n",
      "Ep:3/5,Itr:168/469, Averages:[0.08506], Metrics:[0.97266, 0.97273, 0.97306, 0.97241]\n",
      "Ep:3/5,Itr:174/469, Averages:[0.06239], Metrics:[0.97396, 0.97411, 0.97326, 0.97496]\n",
      "Ep:3/5,Itr:180/469, Averages:[0.10712], Metrics:[0.96875, 0.96887, 0.96872, 0.96903]\n",
      "Ep:3/5,Itr:186/469, Averages:[0.11547], Metrics:[0.96094, 0.96084, 0.96079, 0.96089]\n",
      "Ep:3/5,Itr:192/469, Averages:[0.0845], Metrics:[0.97786, 0.97772, 0.97806, 0.97738]\n",
      "Ep:3/5,Itr:198/469, Averages:[0.08595], Metrics:[0.97656, 0.97575, 0.97541, 0.9761]\n",
      "Ep:3/5,Itr:204/469, Averages:[0.09427], Metrics:[0.97266, 0.97303, 0.97228, 0.97378]\n",
      "Ep:3/5,Itr:210/469, Averages:[0.12357], Metrics:[0.96224, 0.9619, 0.96207, 0.96173]\n",
      "Ep:3/5,Itr:216/469, Averages:[0.09375], Metrics:[0.96875, 0.96895, 0.96984, 0.96806]\n",
      "Ep:3/5,Itr:222/469, Averages:[0.06002], Metrics:[0.98307, 0.9822, 0.98156, 0.98284]\n",
      "Ep:3/5,Itr:228/469, Averages:[0.1034], Metrics:[0.96615, 0.96648, 0.96612, 0.96684]\n",
      "Ep:3/5,Itr:234/469, Averages:[0.05242], Metrics:[0.98177, 0.98138, 0.98056, 0.9822]\n",
      "Ep:3/5,Itr:238/469, Averages:[0.07952], Metrics:[0.97266, 0.97243, 0.97215, 0.9727]\n",
      "Ep:3/5,Itr:245/469, Averages:[0.09293], Metrics:[0.96987, 0.96945, 0.96981, 0.96909]\n",
      "Ep:3/5,Itr:252/469, Averages:[0.07789], Metrics:[0.97321, 0.97336, 0.97299, 0.97373]\n",
      "Ep:3/5,Itr:259/469, Averages:[0.08954], Metrics:[0.9721, 0.9724, 0.97186, 0.97294]\n",
      "Ep:3/5,Itr:266/469, Averages:[0.06855], Metrics:[0.97433, 0.97498, 0.97521, 0.97474]\n",
      "Ep:3/5,Itr:273/469, Averages:[0.07869], Metrics:[0.9721, 0.97202, 0.97161, 0.97243]\n",
      "Ep:3/5,Itr:280/469, Averages:[0.06366], Metrics:[0.97545, 0.9754, 0.97524, 0.97556]\n",
      "Ep:3/5,Itr:287/469, Averages:[0.12094], Metrics:[0.96652, 0.96601, 0.96608, 0.96594]\n",
      "Ep:3/5,Itr:294/469, Averages:[0.12603], Metrics:[0.96763, 0.968, 0.96784, 0.96816]\n",
      "Ep:3/5,Itr:301/469, Averages:[0.09931], Metrics:[0.9654, 0.96501, 0.9652, 0.96481]\n",
      "Ep:3/5,Itr:308/469, Averages:[0.12907], Metrics:[0.97433, 0.97378, 0.97395, 0.97361]\n",
      "Ep:3/5,Itr:315/469, Averages:[0.11294], Metrics:[0.96875, 0.96687, 0.96613, 0.96762]\n",
      "Ep:3/5,Itr:322/469, Averages:[0.08213], Metrics:[0.97433, 0.97358, 0.97341, 0.97376]\n",
      "Ep:3/5,Itr:329/469, Averages:[0.08167], Metrics:[0.97098, 0.97147, 0.97146, 0.97147]\n",
      "Ep:3/5,Itr:336/469, Averages:[0.1203], Metrics:[0.9721, 0.9717, 0.9718, 0.9716]\n",
      "Ep:3/5,Itr:343/469, Averages:[0.08509], Metrics:[0.97545, 0.97615, 0.9761, 0.9762]\n",
      "Ep:3/5,Itr:350/469, Averages:[0.07193], Metrics:[0.97656, 0.97605, 0.9759, 0.97621]\n",
      "Ep:3/5,Itr:357/469, Averages:[0.09063], Metrics:[0.97879, 0.97834, 0.97811, 0.97857]\n",
      "Ep:3/5,Itr:364/469, Averages:[0.09145], Metrics:[0.96763, 0.96824, 0.96845, 0.96803]\n",
      "Ep:3/5,Itr:371/469, Averages:[0.09488], Metrics:[0.97879, 0.97894, 0.97882, 0.97907]\n",
      "Ep:3/5,Itr:378/469, Averages:[0.1422], Metrics:[0.96875, 0.96854, 0.96906, 0.96803]\n",
      "Ep:3/5,Itr:385/469, Averages:[0.08674], Metrics:[0.96875, 0.96752, 0.96796, 0.96708]\n",
      "Ep:3/5,Itr:392/469, Averages:[0.09376], Metrics:[0.97433, 0.97373, 0.97365, 0.97381]\n",
      "Ep:3/5,Itr:399/469, Averages:[0.11581], Metrics:[0.96875, 0.9682, 0.96755, 0.96886]\n",
      "Ep:3/5,Itr:406/469, Averages:[0.08072], Metrics:[0.97545, 0.97498, 0.97531, 0.97465]\n",
      "Ep:3/5,Itr:413/469, Averages:[0.07379], Metrics:[0.97879, 0.97831, 0.97778, 0.97884]\n",
      "Ep:3/5,Itr:420/469, Averages:[0.08148], Metrics:[0.97433, 0.97398, 0.97412, 0.97383]\n",
      "Ep:3/5,Itr:427/469, Averages:[0.12224], Metrics:[0.96205, 0.96261, 0.96288, 0.96234]\n",
      "Ep:3/5,Itr:434/469, Averages:[0.08746], Metrics:[0.97656, 0.97679, 0.97636, 0.97721]\n",
      "Ep:3/5,Itr:441/469, Averages:[0.06488], Metrics:[0.97991, 0.97907, 0.97833, 0.97981]\n",
      "Ep:3/5,Itr:448/469, Averages:[0.10544], Metrics:[0.96875, 0.96809, 0.96728, 0.96891]\n",
      "Ep:3/5,Itr:455/469, Averages:[0.10228], Metrics:[0.97098, 0.97092, 0.97015, 0.97169]\n",
      "Ep:3/5,Itr:462/469, Averages:[0.12338], Metrics:[0.96429, 0.96377, 0.96402, 0.96352]\n",
      "Ep:3/5,Itr:469/469, Averages:[0.09034], Metrics:[0.97338, 0.97229, 0.97139, 0.97319]\n",
      "validation ...\n",
      " Itr:1/79, Averages:[0.01753], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:2/79, Averages:[0.01973], Metrics:[0.99219, 0.99188, 0.99286, 0.99091]\n",
      " Itr:4/79, Averages:[0.02266], Metrics:[0.98438, 0.98346, 0.98462, 0.98231]\n",
      " Itr:6/79, Averages:[0.06321], Metrics:[0.97656, 0.97741, 0.97842, 0.9764]\n",
      " Itr:9/79, Averages:[0.04328], Metrics:[0.97656, 0.97771, 0.98083, 0.97462]\n",
      " Itr:12/79, Averages:[0.05171], Metrics:[0.99219, 0.99161, 0.99231, 0.99091]\n",
      " Itr:15/79, Averages:[0.05298], Metrics:[0.97656, 0.97474, 0.97683, 0.97266]\n",
      " Itr:16/79, Averages:[0.02595], Metrics:[0.99219, 0.99205, 0.99412, 0.99]\n",
      " Itr:20/79, Averages:[0.06287], Metrics:[0.96875, 0.97216, 0.96939, 0.97495]\n",
      " Itr:24/79, Averages:[0.05277], Metrics:[0.97656, 0.97771, 0.9785, 0.97692]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " Itr:28/79, Averages:[0.09753], Metrics:[0.96875, 0.96902, 0.97067, 0.96738]\n",
      " Itr:32/79, Averages:[0.07047], Metrics:[0.96875, 0.96449, 0.9712, 0.95787]\n",
      " Itr:36/79, Averages:[0.09887], Metrics:[0.97656, 0.9759, 0.97348, 0.97833]\n",
      " Itr:40/79, Averages:[0.00585], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:45/79, Averages:[0.00977], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:50/79, Averages:[6.e-05], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:55/79, Averages:[0.00228], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:60/79, Averages:[0.00073], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:65/79, Averages:[0.02355], Metrics:[0.99219, 0.99017, 0.99286, 0.9875]\n",
      " Itr:70/79, Averages:[0.00032], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:75/79, Averages:[0.00984], Metrics:[0.99219, 0.99188, 0.99286, 0.99091]\n",
      "[SUCCESS]! experiment validation Averages:[0.03836], Metrics:[0.9872, 0.98715, 0.98714, 0.98716]\n",
      "[SUCCESS]!  *** Best Model Saved!!! *** : 0.9871485232042972\n",
      "Ep:4/5,Itr:1/469, Averages:[0.10441], Metrics:[0.96875, 0.96988, 0.9671, 0.97268]\n",
      "Ep:4/5,Itr:2/469, Averages:[0.19055], Metrics:[0.9375, 0.93633, 0.94021, 0.93248]\n",
      "Ep:4/5,Itr:4/469, Averages:[0.06761], Metrics:[0.98047, 0.98015, 0.98044, 0.97985]\n",
      "Ep:4/5,Itr:6/469, Averages:[0.05524], Metrics:[0.98438, 0.98413, 0.98428, 0.98399]\n",
      "Ep:4/5,Itr:12/469, Averages:[0.08191], Metrics:[0.97526, 0.9749, 0.9743, 0.9755]\n",
      "Ep:4/5,Itr:16/469, Averages:[0.07615], Metrics:[0.96875, 0.96924, 0.96826, 0.97023]\n",
      "Ep:4/5,Itr:20/469, Averages:[0.07725], Metrics:[0.97852, 0.97874, 0.97789, 0.97959]\n",
      "Ep:4/5,Itr:25/469, Averages:[0.06837], Metrics:[0.98125, 0.98171, 0.98236, 0.98107]\n",
      "Ep:4/5,Itr:30/469, Averages:[0.05456], Metrics:[0.98594, 0.98534, 0.98527, 0.98541]\n",
      "Ep:4/5,Itr:35/469, Averages:[0.06299], Metrics:[0.98281, 0.98292, 0.98235, 0.98349]\n",
      "Ep:4/5,Itr:42/469, Averages:[0.09244], Metrics:[0.97768, 0.97798, 0.97775, 0.97821]\n",
      "Ep:4/5,Itr:48/469, Averages:[0.04782], Metrics:[0.98958, 0.98984, 0.98938, 0.99029]\n",
      "Ep:4/5,Itr:54/469, Averages:[0.09488], Metrics:[0.97396, 0.97326, 0.97382, 0.97269]\n",
      "Ep:4/5,Itr:60/469, Averages:[0.07602], Metrics:[0.97917, 0.97812, 0.97875, 0.97749]\n",
      "Ep:4/5,Itr:66/469, Averages:[0.08291], Metrics:[0.97656, 0.9773, 0.97644, 0.97815]\n",
      "Ep:4/5,Itr:72/469, Averages:[0.09723], Metrics:[0.96875, 0.96757, 0.96755, 0.96758]\n",
      "Ep:4/5,Itr:77/469, Averages:[0.09913], Metrics:[0.97188, 0.9711, 0.97066, 0.97153]\n",
      "Ep:4/5,Itr:84/469, Averages:[0.07719], Metrics:[0.97545, 0.97471, 0.97525, 0.97417]\n",
      "Ep:4/5,Itr:91/469, Averages:[0.10583], Metrics:[0.96875, 0.96826, 0.96741, 0.9691]\n",
      "Ep:4/5,Itr:98/469, Averages:[0.08242], Metrics:[0.97321, 0.97333, 0.97391, 0.97276]\n",
      "Ep:4/5,Itr:105/469, Averages:[0.07456], Metrics:[0.97879, 0.97828, 0.97824, 0.97833]\n",
      "Ep:4/5,Itr:112/469, Averages:[0.06442], Metrics:[0.97879, 0.97865, 0.97902, 0.97827]\n",
      "Ep:4/5,Itr:119/469, Averages:[0.09186], Metrics:[0.97656, 0.97618, 0.9768, 0.97556]\n",
      "Ep:4/5,Itr:126/469, Averages:[0.06677], Metrics:[0.97656, 0.97585, 0.97593, 0.97576]\n",
      "Ep:4/5,Itr:133/469, Averages:[0.08463], Metrics:[0.96987, 0.96996, 0.97019, 0.96972]\n",
      "Ep:4/5,Itr:140/469, Averages:[0.05674], Metrics:[0.98103, 0.98067, 0.98056, 0.98077]\n",
      "Ep:4/5,Itr:147/469, Averages:[0.08125], Metrics:[0.97768, 0.97642, 0.97571, 0.97714]\n",
      "Ep:4/5,Itr:154/469, Averages:[0.06769], Metrics:[0.97991, 0.97948, 0.9794, 0.97956]\n",
      "Ep:4/5,Itr:160/469, Averages:[0.06663], Metrics:[0.98177, 0.98251, 0.98266, 0.98235]\n",
      "Ep:4/5,Itr:168/469, Averages:[0.06819], Metrics:[0.97949, 0.97949, 0.97914, 0.97985]\n",
      "Ep:4/5,Itr:176/469, Averages:[0.06888], Metrics:[0.97852, 0.97852, 0.97837, 0.97867]\n",
      "Ep:4/5,Itr:184/469, Averages:[0.07649], Metrics:[0.97754, 0.97764, 0.97839, 0.97689]\n",
      "Ep:4/5,Itr:192/469, Averages:[0.06657], Metrics:[0.97461, 0.97468, 0.97501, 0.97434]\n",
      "Ep:4/5,Itr:200/469, Averages:[0.08125], Metrics:[0.97754, 0.97796, 0.97793, 0.97799]\n",
      "Ep:4/5,Itr:208/469, Averages:[0.09798], Metrics:[0.96875, 0.9683, 0.96784, 0.96876]\n",
      "Ep:4/5,Itr:216/469, Averages:[0.08554], Metrics:[0.97461, 0.97453, 0.97444, 0.97463]\n",
      "Ep:4/5,Itr:224/469, Averages:[0.0691], Metrics:[0.98145, 0.98147, 0.98118, 0.98175]\n",
      "Ep:4/5,Itr:232/469, Averages:[0.08685], Metrics:[0.97266, 0.97251, 0.97237, 0.97266]\n",
      "Ep:4/5,Itr:240/469, Averages:[0.07202], Metrics:[0.97754, 0.97738, 0.97721, 0.97755]\n",
      "Ep:4/5,Itr:248/469, Averages:[0.06694], Metrics:[0.9834, 0.98356, 0.98324, 0.98388]\n",
      "Ep:4/5,Itr:256/469, Averages:[0.07659], Metrics:[0.97949, 0.97982, 0.97973, 0.9799]\n",
      "Ep:4/5,Itr:264/469, Averages:[0.06226], Metrics:[0.97852, 0.97828, 0.97827, 0.97829]\n",
      "Ep:4/5,Itr:272/469, Averages:[0.08015], Metrics:[0.97559, 0.97518, 0.97509, 0.97527]\n",
      "Ep:4/5,Itr:280/469, Averages:[0.0872], Metrics:[0.97559, 0.97494, 0.97565, 0.97423]\n",
      "Ep:4/5,Itr:288/469, Averages:[0.07351], Metrics:[0.97852, 0.97829, 0.97796, 0.97862]\n",
      "Ep:4/5,Itr:296/469, Averages:[0.0659], Metrics:[0.97461, 0.97495, 0.97518, 0.97472]\n",
      "Ep:4/5,Itr:304/469, Averages:[0.06676], Metrics:[0.9834, 0.98332, 0.98386, 0.98278]\n",
      "Ep:4/5,Itr:312/469, Averages:[0.05585], Metrics:[0.9834, 0.98309, 0.9829, 0.98327]\n",
      "Ep:4/5,Itr:320/469, Averages:[0.05712], Metrics:[0.98242, 0.9822, 0.98147, 0.98293]\n",
      "Ep:4/5,Itr:324/469, Averages:[0.07885], Metrics:[0.97656, 0.97866, 0.97847, 0.97884]\n",
      "Ep:4/5,Itr:333/469, Averages:[0.05881], Metrics:[0.98351, 0.98325, 0.98351, 0.98299]\n",
      "Ep:4/5,Itr:342/469, Averages:[0.07346], Metrics:[0.97049, 0.9699, 0.97004, 0.96977]\n",
      "Ep:4/5,Itr:351/469, Averages:[0.06828], Metrics:[0.97569, 0.97553, 0.97574, 0.97533]\n",
      "Ep:4/5,Itr:360/469, Averages:[0.07343], Metrics:[0.97656, 0.97673, 0.97633, 0.97713]\n",
      "Ep:4/5,Itr:369/469, Averages:[0.05821], Metrics:[0.97917, 0.97918, 0.97931, 0.97906]\n",
      "Ep:4/5,Itr:378/469, Averages:[0.05665], Metrics:[0.97917, 0.97873, 0.9785, 0.97897]\n",
      "Ep:4/5,Itr:387/469, Averages:[0.07679], Metrics:[0.97569, 0.9757, 0.9767, 0.97471]\n",
      "Ep:4/5,Itr:396/469, Averages:[0.07844], Metrics:[0.98264, 0.98215, 0.98208, 0.98221]\n",
      "Ep:4/5,Itr:405/469, Averages:[0.06866], Metrics:[0.97743, 0.97695, 0.97648, 0.97742]\n",
      "Ep:4/5,Itr:414/469, Averages:[0.09973], Metrics:[0.97222, 0.97182, 0.97215, 0.97148]\n",
      "Ep:4/5,Itr:423/469, Averages:[0.09955], Metrics:[0.97222, 0.97385, 0.97418, 0.97353]\n",
      "Ep:4/5,Itr:432/469, Averages:[0.07125], Metrics:[0.97569, 0.97554, 0.97589, 0.97519]\n",
      "Ep:4/5,Itr:441/469, Averages:[0.07843], Metrics:[0.97743, 0.97717, 0.97692, 0.97742]\n",
      "Ep:4/5,Itr:450/469, Averages:[0.0678], Metrics:[0.9783, 0.97788, 0.97808, 0.97768]\n",
      "Ep:4/5,Itr:459/469, Averages:[0.09407], Metrics:[0.97309, 0.97349, 0.97332, 0.97366]\n",
      "Ep:4/5,Itr:468/469, Averages:[0.06921], Metrics:[0.97483, 0.97477, 0.97479, 0.97475]\n",
      "Ep:4/5,Itr:469/469, Averages:[0.09629], Metrics:[0.96875, 0.96919, 0.96806, 0.97033]\n",
      "validation ...\n",
      " Itr:1/79, Averages:[0.01341], Metrics:[0.99219, 0.99468, 0.99412, 0.99524]\n",
      " Itr:2/79, Averages:[0.02418], Metrics:[0.99219, 0.99188, 0.99286, 0.99091]\n",
      " Itr:4/79, Averages:[0.00916], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:6/79, Averages:[0.0387], Metrics:[0.97656, 0.97811, 0.9798, 0.97643]\n",
      " Itr:12/79, Averages:[0.06094], Metrics:[0.99219, 0.99161, 0.99231, 0.99091]\n",
      " Itr:16/79, Averages:[0.0577], Metrics:[0.96875, 0.97196, 0.97453, 0.9694]\n",
      " Itr:20/79, Averages:[0.02306], Metrics:[0.98438, 0.98574, 0.98542, 0.98606]\n",
      " Itr:25/79, Averages:[0.02386], Metrics:[0.99219, 0.99199, 0.99167, 0.99231]\n",
      " Itr:30/79, Averages:[0.10946], Metrics:[0.98438, 0.98484, 0.98452, 0.98516]\n",
      " Itr:35/79, Averages:[0.01105], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:42/79, Averages:[0.00228], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:48/79, Averages:[0.0282], Metrics:[0.98438, 0.98502, 0.98182, 0.98824]\n",
      " Itr:54/79, Averages:[0.00274], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:60/79, Averages:[0.00104], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:66/79, Averages:[0.02766], Metrics:[0.98438, 0.9849, 0.98231, 0.9875]\n",
      " Itr:72/79, Averages:[0.00022], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:77/79, Averages:[0.1157], Metrics:[0.97656, 0.977, 0.97564, 0.97837]\n",
      "[SUCCESS]! experiment validation Averages:[0.03841], Metrics:[0.9886, 0.98853, 0.98843, 0.98864]\n",
      "[SUCCESS]!  *** Best Model Saved!!! *** : 0.9885338411380431\n",
      "Ep:5/5,Itr:1/469, Averages:[0.06429], Metrics:[0.98438, 0.98875, 0.98875, 0.98875]\n",
      "Ep:5/5,Itr:2/469, Averages:[0.076], Metrics:[0.98438, 0.98392, 0.9812, 0.98667]\n",
      "Ep:5/5,Itr:6/469, Averages:[0.07497], Metrics:[0.97656, 0.97629, 0.97601, 0.97656]\n",
      "Ep:5/5,Itr:8/469, Averages:[0.0836], Metrics:[0.96094, 0.96027, 0.962, 0.95854]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ep:5/5,Itr:12/469, Averages:[0.08535], Metrics:[0.97852, 0.9773, 0.97684, 0.97776]\n",
      "Ep:5/5,Itr:15/469, Averages:[0.07203], Metrics:[0.97396, 0.97458, 0.97696, 0.97221]\n",
      "Ep:5/5,Itr:20/469, Averages:[0.04695], Metrics:[0.98906, 0.98932, 0.9893, 0.98933]\n",
      "Ep:5/5,Itr:24/469, Averages:[0.07766], Metrics:[0.98438, 0.98462, 0.98446, 0.98479]\n",
      "Ep:5/5,Itr:30/469, Averages:[0.05388], Metrics:[0.98568, 0.98569, 0.98594, 0.98544]\n",
      "Ep:5/5,Itr:36/469, Averages:[0.05982], Metrics:[0.98047, 0.98004, 0.97974, 0.98034]\n",
      "Ep:5/5,Itr:42/469, Averages:[0.05551], Metrics:[0.98307, 0.98233, 0.98206, 0.9826]\n",
      "Ep:5/5,Itr:49/469, Averages:[0.05329], Metrics:[0.98549, 0.98519, 0.98512, 0.98525]\n",
      "Ep:5/5,Itr:56/469, Averages:[0.08934], Metrics:[0.97321, 0.9734, 0.97417, 0.97263]\n",
      "Ep:5/5,Itr:63/469, Averages:[0.07748], Metrics:[0.98103, 0.98074, 0.97983, 0.98166]\n",
      "Ep:5/5,Itr:70/469, Averages:[0.08272], Metrics:[0.97768, 0.97765, 0.97823, 0.97707]\n",
      "Ep:5/5,Itr:77/469, Averages:[0.07239], Metrics:[0.97656, 0.97656, 0.97645, 0.97666]\n",
      "Ep:5/5,Itr:80/469, Averages:[0.05178], Metrics:[0.98177, 0.98033, 0.98014, 0.98052]\n",
      "Ep:5/5,Itr:88/469, Averages:[0.0407], Metrics:[0.98633, 0.98652, 0.98619, 0.98685]\n",
      "Ep:5/5,Itr:96/469, Averages:[0.06276], Metrics:[0.98535, 0.9857, 0.98593, 0.98547]\n",
      "Ep:5/5,Itr:104/469, Averages:[0.05981], Metrics:[0.97656, 0.97609, 0.97578, 0.97639]\n",
      "Ep:5/5,Itr:112/469, Averages:[0.06892], Metrics:[0.98047, 0.97995, 0.97986, 0.98003]\n",
      "Ep:5/5,Itr:120/469, Averages:[0.05113], Metrics:[0.98438, 0.98376, 0.98389, 0.98364]\n",
      "Ep:5/5,Itr:128/469, Averages:[0.0647], Metrics:[0.97754, 0.97719, 0.97754, 0.97683]\n",
      "Ep:5/5,Itr:136/469, Averages:[0.05028], Metrics:[0.9873, 0.98757, 0.98818, 0.98696]\n",
      "Ep:5/5,Itr:144/469, Averages:[0.06139], Metrics:[0.98047, 0.98009, 0.98045, 0.97972]\n",
      "Ep:5/5,Itr:153/469, Averages:[0.05287], Metrics:[0.98524, 0.98553, 0.9854, 0.98567]\n",
      "Ep:5/5,Itr:162/469, Averages:[0.074], Metrics:[0.9809, 0.98078, 0.98091, 0.98065]\n",
      "Ep:5/5,Itr:171/469, Averages:[0.06693], Metrics:[0.97569, 0.97577, 0.97528, 0.97626]\n",
      "Ep:5/5,Itr:180/469, Averages:[0.06548], Metrics:[0.98785, 0.9876, 0.98733, 0.98787]\n",
      "Ep:5/5,Itr:189/469, Averages:[0.06192], Metrics:[0.9809, 0.98103, 0.9814, 0.98066]\n",
      "Ep:5/5,Itr:198/469, Averages:[0.0775], Metrics:[0.97396, 0.97412, 0.97433, 0.97391]\n",
      "Ep:5/5,Itr:207/469, Averages:[0.07012], Metrics:[0.97309, 0.97315, 0.97307, 0.97324]\n",
      "Ep:5/5,Itr:216/469, Averages:[0.05547], Metrics:[0.98264, 0.98281, 0.98231, 0.9833]\n",
      "Ep:5/5,Itr:225/469, Averages:[0.06716], Metrics:[0.98351, 0.9834, 0.98335, 0.98345]\n",
      "Ep:5/5,Itr:234/469, Averages:[0.07044], Metrics:[0.97656, 0.97675, 0.97658, 0.97692]\n",
      "Ep:5/5,Itr:243/469, Averages:[0.09626], Metrics:[0.97917, 0.97922, 0.97915, 0.97928]\n",
      "Ep:5/5,Itr:252/469, Averages:[0.06899], Metrics:[0.9783, 0.97806, 0.97837, 0.97775]\n",
      "Ep:5/5,Itr:261/469, Averages:[0.0621], Metrics:[0.98264, 0.98234, 0.98285, 0.98184]\n",
      "Ep:5/5,Itr:270/469, Averages:[0.08093], Metrics:[0.9783, 0.97794, 0.9776, 0.97827]\n",
      "Ep:5/5,Itr:280/469, Averages:[0.09527], Metrics:[0.975, 0.97521, 0.975, 0.97541]\n",
      "Ep:5/5,Itr:290/469, Averages:[0.05753], Metrics:[0.97891, 0.97909, 0.97912, 0.97905]\n",
      "Ep:5/5,Itr:300/469, Averages:[0.05907], Metrics:[0.98516, 0.98456, 0.98381, 0.98531]\n",
      "Ep:5/5,Itr:310/469, Averages:[0.04005], Metrics:[0.9875, 0.98756, 0.9872, 0.98791]\n",
      "Ep:5/5,Itr:320/469, Averages:[0.0808], Metrics:[0.97578, 0.9755, 0.97498, 0.97601]\n",
      "Ep:5/5,Itr:330/469, Averages:[0.05318], Metrics:[0.98594, 0.98647, 0.98641, 0.98653]\n",
      "Ep:5/5,Itr:340/469, Averages:[0.06975], Metrics:[0.975, 0.97498, 0.97489, 0.97507]\n",
      "Ep:5/5,Itr:350/469, Averages:[0.07637], Metrics:[0.98047, 0.98032, 0.98003, 0.98061]\n",
      "Ep:5/5,Itr:360/469, Averages:[0.06617], Metrics:[0.97734, 0.97703, 0.97697, 0.97709]\n",
      "Ep:5/5,Itr:370/469, Averages:[0.0678], Metrics:[0.98125, 0.98084, 0.98145, 0.98023]\n",
      "Ep:5/5,Itr:380/469, Averages:[0.07465], Metrics:[0.97813, 0.978, 0.97763, 0.97837]\n",
      "Ep:5/5,Itr:390/469, Averages:[0.05893], Metrics:[0.97891, 0.97852, 0.97896, 0.97809]\n",
      "Ep:5/5,Itr:400/469, Averages:[0.0535], Metrics:[0.98125, 0.98121, 0.98119, 0.98124]\n",
      "Ep:5/5,Itr:410/469, Averages:[0.06248], Metrics:[0.98125, 0.98143, 0.98135, 0.98151]\n",
      "Ep:5/5,Itr:420/469, Averages:[0.08648], Metrics:[0.97813, 0.97862, 0.97836, 0.97887]\n",
      "Ep:5/5,Itr:430/469, Averages:[0.07207], Metrics:[0.98203, 0.98162, 0.98154, 0.9817]\n",
      "Ep:5/5,Itr:440/469, Averages:[0.08042], Metrics:[0.98125, 0.98083, 0.97999, 0.98167]\n",
      "Ep:5/5,Itr:450/469, Averages:[0.05768], Metrics:[0.98203, 0.98207, 0.98185, 0.98229]\n",
      "Ep:5/5,Itr:460/469, Averages:[0.06798], Metrics:[0.97578, 0.97571, 0.97593, 0.97548]\n",
      "Ep:5/5,Itr:469/469, Averages:[0.08775], Metrics:[0.97679, 0.97745, 0.9777, 0.9772]\n",
      "validation ...\n",
      " Itr:1/79, Averages:[0.00776], Metrics:[0.99219, 0.99468, 0.99412, 0.99524]\n",
      " Itr:2/79, Averages:[0.00129], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:6/79, Averages:[0.02205], Metrics:[0.99219, 0.9932, 0.99167, 0.99474]\n",
      " Itr:8/79, Averages:[0.18416], Metrics:[0.96875, 0.96456, 0.96564, 0.96348]\n",
      " Itr:12/79, Averages:[0.0723], Metrics:[0.97656, 0.97411, 0.9712, 0.97704]\n",
      " Itr:15/79, Averages:[0.08381], Metrics:[0.98438, 0.98276, 0.98516, 0.98036]\n",
      " Itr:20/79, Averages:[0.0867], Metrics:[0.97656, 0.97895, 0.97917, 0.97873]\n",
      " Itr:24/79, Averages:[0.05312], Metrics:[0.97656, 0.97647, 0.97731, 0.97564]\n",
      " Itr:30/79, Averages:[0.10608], Metrics:[0.99219, 0.99286, 0.99286, 0.99286]\n",
      " Itr:36/79, Averages:[0.10752], Metrics:[0.96875, 0.96749, 0.96843, 0.96655]\n",
      " Itr:42/79, Averages:[0.0004], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:49/79, Averages:[0.02024], Metrics:[0.99219, 0.99166, 0.99333, 0.99]\n",
      " Itr:56/79, Averages:[0.00243], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:63/79, Averages:[0.01253], Metrics:[0.99219, 0.99226, 0.99167, 0.99286]\n",
      " Itr:70/79, Averages:[0.00212], Metrics:[1.0, 1.0, 1.0, 1.0]\n",
      " Itr:77/79, Averages:[0.07714], Metrics:[0.98438, 0.98397, 0.98333, 0.98462]\n",
      "[SUCCESS]! experiment validation Averages:[0.03826], Metrics:[0.9887, 0.98868, 0.98866, 0.9887]\n",
      "[SUCCESS]!  *** Best Model Saved!!! *** : 0.988679444201353\n"
     ]
    }
   ],
   "source": [
    "dataloader_args = {'train': {'dataset': train_dataset},\n",
    "                   'validation': {'dataset': val_dataset}}\n",
    "runner = EasyTorch(phase='train', distributed_validation=True,\n",
    "                   batch_size=512, epochs=21,\n",
    "                   dataloader_args=dataloader_args,\n",
    "                   image_size=(28, 28))\n",
    "runner.run(MNISTTrainer)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Image('net_logs/experiment/experiment_train_log_0.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Image('net_logs/experiment/experiment_train_log_1.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Saved logs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "import json\n",
    "import pprint as ppr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "log = json.loads(open('net_logs/experiment/experiment_log.json').read())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'phase': 'train',\n",
       " 'batch_size': 128,\n",
       " 'epochs': 3,\n",
       " 'grad_accum_iters': 1,\n",
       " 'learning_rate': 0.001,\n",
       " 'gpus': [],\n",
       " 'pin_memory': False,\n",
       " 'num_workers': 0,\n",
       " 'dataset_dir': '',\n",
       " 'load_limit': None,\n",
       " 'log_dir': 'net_logs/experiment',\n",
       " 'pretrained_path': None,\n",
       " 'verbose': True,\n",
       " 'seed_all': False,\n",
       " 'force': False,\n",
       " 'patience': None,\n",
       " 'load_sparse': False,\n",
       " 'num_folds': None,\n",
       " 'split_ratio': None,\n",
       " 'seed': 9500374,\n",
       " 'global_test_metrics': [],\n",
       " 'log_header': 'Loss|Accuracy,F1,Precision,Recall',\n",
       " 'monitor_metric': 'f1',\n",
       " 'metric_direction': 'maximize',\n",
       " 'train_log': [[4.55726, 0.17188, 0.15343, 0.15437, 0.15251],\n",
       "  [8.36324, 0.11719, 0.15221, 0.11551, 0.22307],\n",
       "  [4.6089, 0.20312, 0.18334, 0.18453, 0.18216],\n",
       "  [3.00693, 0.38281, 0.31471, 0.3559, 0.28206],\n",
       "  [2.58417, 0.36719, 0.38099, 0.38982, 0.37256],\n",
       "  [2.54765, 0.28906, 0.32768, 0.33277, 0.32274],\n",
       "  [1.89565, 0.35938, 0.34514, 0.36923, 0.324],\n",
       "  [1.65683, 0.39062, 0.33505, 0.37506, 0.30275],\n",
       "  [1.60825, 0.49219, 0.41763, 0.43247, 0.40379],\n",
       "  [1.62466, 0.42969, 0.45435, 0.43465, 0.47592],\n",
       "  [1.60842, 0.42969, 0.3926, 0.4061, 0.37996],\n",
       "  [1.46395, 0.53906, 0.551, 0.53689, 0.56587],\n",
       "  [1.46863, 0.53125, 0.53597, 0.52167, 0.55108],\n",
       "  [1.30625, 0.53125, 0.50984, 0.49307, 0.5278],\n",
       "  [1.49132, 0.47656, 0.47239, 0.46228, 0.48295],\n",
       "  [1.20914, 0.60156, 0.57295, 0.56614, 0.57992],\n",
       "  [1.38259, 0.53906, 0.51787, 0.51276, 0.52308],\n",
       "  [1.47863, 0.54688, 0.53611, 0.53712, 0.5351],\n",
       "  [1.1292, 0.67969, 0.65715, 0.65469, 0.65962],\n",
       "  [1.04504, 0.65625, 0.64727, 0.62931, 0.66627],\n",
       "  [1.05973, 0.64062, 0.63603, 0.64603, 0.62634],\n",
       "  [1.14065, 0.58594, 0.58988, 0.59711, 0.58282],\n",
       "  [0.91851, 0.70312, 0.69953, 0.70368, 0.69544],\n",
       "  [0.95304, 0.66406, 0.6401, 0.63825, 0.64196],\n",
       "  [0.99082, 0.60938, 0.62537, 0.62035, 0.63048],\n",
       "  [0.99906, 0.69531, 0.68988, 0.68014, 0.69991],\n",
       "  [0.92053, 0.64062, 0.62599, 0.62173, 0.63031],\n",
       "  [0.84629, 0.72656, 0.73041, 0.72735, 0.7335],\n",
       "  [0.90382, 0.66406, 0.67775, 0.68763, 0.66815],\n",
       "  [0.70123, 0.73438, 0.73627, 0.74036, 0.73222],\n",
       "  [1.06761, 0.625, 0.65972, 0.67995, 0.64066],\n",
       "  [1.01731, 0.67188, 0.70703, 0.7151, 0.69914],\n",
       "  [0.7293, 0.75781, 0.73578, 0.73431, 0.73726],\n",
       "  [0.88746, 0.71094, 0.68358, 0.69252, 0.67488],\n",
       "  [0.71489, 0.83594, 0.83075, 0.83306, 0.82845],\n",
       "  [0.76198, 0.75, 0.74521, 0.73123, 0.75974],\n",
       "  [0.61733, 0.80469, 0.79798, 0.80448, 0.79159],\n",
       "  [0.71694, 0.78125, 0.75283, 0.7443, 0.76155],\n",
       "  [0.91543, 0.74219, 0.72892, 0.72961, 0.72824],\n",
       "  [0.54795, 0.82031, 0.82036, 0.81883, 0.8219],\n",
       "  [0.76018, 0.75781, 0.77007, 0.78508, 0.75563],\n",
       "  [0.59263, 0.79688, 0.7894, 0.7895, 0.78931],\n",
       "  [0.48664, 0.82031, 0.82184, 0.81907, 0.82463],\n",
       "  [0.69726, 0.80469, 0.80624, 0.80531, 0.80718],\n",
       "  [0.61484, 0.82031, 0.81896, 0.81981, 0.81811],\n",
       "  [0.70614, 0.77344, 0.78211, 0.79037, 0.77402],\n",
       "  [0.5746, 0.77344, 0.77114, 0.77149, 0.77079],\n",
       "  [0.62967, 0.82812, 0.82735, 0.82532, 0.82939],\n",
       "  [0.51804, 0.8125, 0.79856, 0.80475, 0.79247],\n",
       "  [0.5742, 0.79688, 0.81041, 0.80651, 0.81434],\n",
       "  [0.68084, 0.75781, 0.74834, 0.74964, 0.74704],\n",
       "  [0.60518, 0.76562, 0.75655, 0.7645, 0.74876],\n",
       "  [0.39087, 0.88281, 0.89159, 0.89547, 0.88774],\n",
       "  [0.34039, 0.91406, 0.90503, 0.903, 0.90707],\n",
       "  [0.64755, 0.80469, 0.79436, 0.79282, 0.7959],\n",
       "  [0.46053, 0.875, 0.86334, 0.86078, 0.86592],\n",
       "  [0.52184, 0.82031, 0.81112, 0.81504, 0.80725],\n",
       "  [0.48573, 0.84375, 0.84435, 0.85198, 0.83685],\n",
       "  [0.45234, 0.82812, 0.82948, 0.8167, 0.84268],\n",
       "  [0.55778, 0.80469, 0.79738, 0.80427, 0.79061],\n",
       "  [0.4739, 0.83594, 0.82514, 0.82193, 0.82837],\n",
       "  [0.56208, 0.79688, 0.79624, 0.79522, 0.79727],\n",
       "  [0.58344, 0.82812, 0.83993, 0.83557, 0.84434],\n",
       "  [0.62213, 0.82031, 0.82264, 0.82022, 0.82506],\n",
       "  [0.50476, 0.83594, 0.82915, 0.82643, 0.83188],\n",
       "  [0.46791, 0.83594, 0.83617, 0.82799, 0.84452],\n",
       "  [0.59363, 0.78125, 0.78773, 0.79271, 0.78282],\n",
       "  [0.46305, 0.86719, 0.86153, 0.86673, 0.8564],\n",
       "  [0.32909, 0.90625, 0.91085, 0.9117, 0.91],\n",
       "  [0.52382, 0.84375, 0.84983, 0.84694, 0.85273],\n",
       "  [0.61286, 0.82812, 0.81533, 0.81144, 0.81925],\n",
       "  [0.4883, 0.82812, 0.82977, 0.82487, 0.83473],\n",
       "  [0.33854, 0.90625, 0.90341, 0.89975, 0.9071],\n",
       "  [0.34392, 0.89844, 0.90217, 0.9037, 0.90064],\n",
       "  [0.4489, 0.89062, 0.87358, 0.85933, 0.88831],\n",
       "  [0.38631, 0.89062, 0.90162, 0.89552, 0.90779],\n",
       "  [0.44858, 0.875, 0.87311, 0.87737, 0.86889],\n",
       "  [0.58783, 0.82031, 0.80956, 0.80009, 0.81926],\n",
       "  [0.42713, 0.85156, 0.83317, 0.83811, 0.82829],\n",
       "  [0.49332, 0.84375, 0.84115, 0.83153, 0.851],\n",
       "  [0.57049, 0.84375, 0.84254, 0.84656, 0.83855],\n",
       "  [0.41842, 0.875, 0.86535, 0.86325, 0.86747],\n",
       "  [0.36822, 0.89062, 0.89344, 0.89266, 0.89421],\n",
       "  [0.55312, 0.79688, 0.80565, 0.80472, 0.80658],\n",
       "  [0.32056, 0.85156, 0.84661, 0.84482, 0.8484],\n",
       "  [0.33257, 0.9375, 0.93643, 0.92968, 0.94327],\n",
       "  [0.37054, 0.88281, 0.88015, 0.87955, 0.88076],\n",
       "  [0.27989, 0.89844, 0.90005, 0.8971, 0.90301],\n",
       "  [0.40642, 0.88281, 0.88892, 0.88119, 0.8968],\n",
       "  [0.31637, 0.89844, 0.89052, 0.8853, 0.8958],\n",
       "  [0.41347, 0.88281, 0.88638, 0.88562, 0.88714],\n",
       "  [0.43575, 0.875, 0.8832, 0.88253, 0.88388],\n",
       "  [0.37658, 0.875, 0.8761, 0.86969, 0.88261],\n",
       "  [0.33486, 0.88281, 0.88206, 0.89059, 0.8737],\n",
       "  [0.41043, 0.88281, 0.88985, 0.88279, 0.89702],\n",
       "  [0.31754, 0.91406, 0.91873, 0.91248, 0.92506],\n",
       "  [0.24146, 0.94531, 0.94906, 0.95079, 0.94734],\n",
       "  [0.40553, 0.85156, 0.84943, 0.84972, 0.84915],\n",
       "  [0.33391, 0.92188, 0.91513, 0.90973, 0.92059],\n",
       "  [0.34153, 0.89062, 0.88708, 0.89063, 0.88356],\n",
       "  [0.25019, 0.92188, 0.91968, 0.92511, 0.91432],\n",
       "  [0.54053, 0.86719, 0.87316, 0.8696, 0.87675],\n",
       "  [0.39861, 0.89062, 0.89136, 0.88043, 0.90256],\n",
       "  [0.27615, 0.9375, 0.94429, 0.94238, 0.94622],\n",
       "  [0.28959, 0.9375, 0.93249, 0.93565, 0.92935],\n",
       "  [0.35122, 0.88281, 0.88744, 0.89192, 0.883],\n",
       "  [0.34019, 0.90625, 0.90069, 0.90973, 0.89183],\n",
       "  [0.40008, 0.89844, 0.89297, 0.89267, 0.89327],\n",
       "  [0.24099, 0.9375, 0.93583, 0.9325, 0.93919],\n",
       "  [0.29627, 0.91406, 0.91715, 0.93023, 0.90444],\n",
       "  [0.22556, 0.9375, 0.93579, 0.93218, 0.93942],\n",
       "  [0.28459, 0.90625, 0.90664, 0.91266, 0.9007],\n",
       "  [0.22913, 0.92969, 0.9323, 0.93599, 0.92863],\n",
       "  [0.2352, 0.92969, 0.9165, 0.91529, 0.9177],\n",
       "  [0.45679, 0.89062, 0.8841, 0.88444, 0.88375],\n",
       "  [0.51113, 0.83594, 0.83672, 0.8381, 0.83534],\n",
       "  [0.30046, 0.92188, 0.92713, 0.93133, 0.92297],\n",
       "  [0.40456, 0.86719, 0.86886, 0.87627, 0.86158],\n",
       "  [0.20109, 0.92188, 0.92053, 0.92204, 0.91903],\n",
       "  [0.30779, 0.875, 0.86685, 0.88486, 0.84957],\n",
       "  [0.21922, 0.92969, 0.92031, 0.91923, 0.92139],\n",
       "  [0.2054, 0.92969, 0.92856, 0.9235, 0.93367],\n",
       "  [0.38806, 0.85938, 0.86843, 0.87704, 0.85998],\n",
       "  [0.28912, 0.90625, 0.91514, 0.92003, 0.91031],\n",
       "  [0.27311, 0.89844, 0.90306, 0.90383, 0.90228],\n",
       "  [0.37065, 0.875, 0.87115, 0.87232, 0.86999],\n",
       "  [0.31851, 0.89844, 0.89386, 0.89719, 0.89055],\n",
       "  [0.49657, 0.875, 0.87024, 0.86108, 0.8796],\n",
       "  [0.35754, 0.88281, 0.88511, 0.88324, 0.88698],\n",
       "  [0.31994, 0.89062, 0.88433, 0.88583, 0.88284],\n",
       "  [0.48412, 0.91406, 0.92473, 0.91658, 0.93303],\n",
       "  [0.20983, 0.92188, 0.92384, 0.92712, 0.92057],\n",
       "  [0.26637, 0.92188, 0.91744, 0.9152, 0.91968],\n",
       "  [0.397, 0.86719, 0.88536, 0.88961, 0.88116],\n",
       "  [0.29868, 0.89844, 0.89953, 0.89452, 0.90458],\n",
       "  [0.19516, 0.9375, 0.94107, 0.93688, 0.9453],\n",
       "  [0.3092, 0.875, 0.87668, 0.87522, 0.87814],\n",
       "  [0.21454, 0.92188, 0.91497, 0.914, 0.91595],\n",
       "  [0.19517, 0.9375, 0.93459, 0.93311, 0.93608],\n",
       "  [0.30153, 0.90625, 0.90886, 0.9119, 0.90584],\n",
       "  [0.23673, 0.92188, 0.91927, 0.92254, 0.91603],\n",
       "  [0.31746, 0.89844, 0.90447, 0.89924, 0.90976],\n",
       "  [0.25487, 0.92969, 0.92153, 0.92825, 0.9149],\n",
       "  [0.24312, 0.92969, 0.92575, 0.92398, 0.92753],\n",
       "  [0.43757, 0.875, 0.87564, 0.87403, 0.87725],\n",
       "  [0.36199, 0.91406, 0.89295, 0.89727, 0.88867],\n",
       "  [0.22777, 0.94531, 0.95264, 0.95584, 0.94946],\n",
       "  [0.1938, 0.92188, 0.92297, 0.92369, 0.92225],\n",
       "  [0.21777, 0.9375, 0.93961, 0.94214, 0.9371],\n",
       "  [0.15881, 0.96094, 0.96031, 0.95983, 0.96078],\n",
       "  [0.24807, 0.91406, 0.91796, 0.93873, 0.89809],\n",
       "  [0.35268, 0.92188, 0.93031, 0.93446, 0.92619],\n",
       "  [0.43743, 0.91406, 0.91271, 0.91348, 0.91194],\n",
       "  [0.35672, 0.90625, 0.90708, 0.90608, 0.90808],\n",
       "  [0.29125, 0.91406, 0.91244, 0.92247, 0.90262],\n",
       "  [0.23182, 0.92188, 0.92085, 0.91939, 0.92231],\n",
       "  [0.27417, 0.91406, 0.91572, 0.91885, 0.91261],\n",
       "  [0.19415, 0.92969, 0.92424, 0.91547, 0.93319],\n",
       "  [0.18007, 0.92969, 0.92242, 0.92713, 0.91777],\n",
       "  [0.21388, 0.94531, 0.94946, 0.95212, 0.94682],\n",
       "  [0.20559, 0.91406, 0.91006, 0.91828, 0.90198],\n",
       "  [0.21755, 0.9375, 0.91586, 0.91133, 0.92044],\n",
       "  [0.1581, 0.96094, 0.95922, 0.96452, 0.95397],\n",
       "  [0.17907, 0.96094, 0.95678, 0.95918, 0.95439],\n",
       "  [0.34572, 0.89844, 0.90671, 0.9025, 0.91096],\n",
       "  [0.1992, 0.9375, 0.93658, 0.93749, 0.93567],\n",
       "  [0.22074, 0.9375, 0.93405, 0.92163, 0.94681],\n",
       "  [0.16332, 0.9375, 0.94414, 0.94741, 0.94089],\n",
       "  [0.29575, 0.91406, 0.91488, 0.91081, 0.91898],\n",
       "  [0.25549, 0.91406, 0.91404, 0.91579, 0.9123],\n",
       "  [0.30042, 0.92188, 0.91788, 0.91619, 0.91958],\n",
       "  [0.0921, 0.99219, 0.99143, 0.99, 0.99286],\n",
       "  [0.28434, 0.875, 0.87064, 0.86639, 0.87494],\n",
       "  [0.31905, 0.94531, 0.94607, 0.94202, 0.95016],\n",
       "  [0.39319, 0.91406, 0.91459, 0.90971, 0.91952],\n",
       "  [0.36033, 0.90625, 0.90747, 0.91048, 0.90447],\n",
       "  [0.14102, 0.96094, 0.95978, 0.95551, 0.96408],\n",
       "  [0.17517, 0.96094, 0.9639, 0.96388, 0.96393],\n",
       "  [0.26547, 0.91406, 0.92283, 0.92271, 0.92294],\n",
       "  [0.23303, 0.92188, 0.92586, 0.92466, 0.92707],\n",
       "  [0.34401, 0.91406, 0.92089, 0.91906, 0.92273],\n",
       "  [0.2184, 0.91406, 0.9137, 0.91688, 0.91054],\n",
       "  [0.17658, 0.94531, 0.94688, 0.94831, 0.94546],\n",
       "  [0.23009, 0.94531, 0.94434, 0.94924, 0.93949],\n",
       "  [0.21827, 0.94531, 0.95085, 0.95269, 0.94902],\n",
       "  [0.33699, 0.91406, 0.91006, 0.91378, 0.90637],\n",
       "  [0.25978, 0.90625, 0.91108, 0.90709, 0.9151],\n",
       "  [0.37901, 0.90625, 0.90577, 0.91054, 0.90104],\n",
       "  [0.22946, 0.9375, 0.93588, 0.94602, 0.92595],\n",
       "  [0.20426, 0.92188, 0.91785, 0.91572, 0.92],\n",
       "  [0.33834, 0.90625, 0.89977, 0.8985, 0.90105],\n",
       "  [0.24321, 0.9375, 0.94046, 0.93962, 0.9413],\n",
       "  [0.17215, 0.9375, 0.9306, 0.9201, 0.94135],\n",
       "  [0.27571, 0.94531, 0.95017, 0.94739, 0.95296],\n",
       "  [0.18813, 0.96094, 0.96338, 0.96, 0.96679],\n",
       "  [0.23866, 0.92969, 0.92734, 0.9258, 0.92888],\n",
       "  [0.2323, 0.92969, 0.92128, 0.92475, 0.91783],\n",
       "  [0.20815, 0.95312, 0.9546, 0.95289, 0.95632],\n",
       "  [0.34287, 0.91406, 0.91132, 0.91094, 0.9117],\n",
       "  [0.33923, 0.91406, 0.90601, 0.90294, 0.90909],\n",
       "  [0.29273, 0.91406, 0.91866, 0.91549, 0.92185],\n",
       "  [0.16269, 0.94531, 0.94746, 0.94653, 0.9484],\n",
       "  [0.21126, 0.92969, 0.93882, 0.93801, 0.93964],\n",
       "  [0.123, 0.97656, 0.97656, 0.97564, 0.97747],\n",
       "  [0.28032, 0.89062, 0.88641, 0.88411, 0.88871],\n",
       "  [0.16599, 0.94531, 0.93453, 0.9353, 0.93375],\n",
       "  [0.25192, 0.94531, 0.94279, 0.94256, 0.94301],\n",
       "  [0.1151, 0.94531, 0.94247, 0.94563, 0.93933],\n",
       "  [0.22335, 0.91406, 0.90652, 0.90735, 0.90569],\n",
       "  [0.31556, 0.85156, 0.85111, 0.85522, 0.84704],\n",
       "  [0.26178, 0.92188, 0.92108, 0.92633, 0.91589],\n",
       "  [0.18829, 0.9375, 0.93681, 0.9345, 0.93913],\n",
       "  [0.28237, 0.92969, 0.93219, 0.93124, 0.93314],\n",
       "  [0.15927, 0.92969, 0.93501, 0.93435, 0.93567],\n",
       "  [0.12188, 0.96875, 0.97276, 0.97063, 0.9749],\n",
       "  [0.36004, 0.86719, 0.85961, 0.84243, 0.8775],\n",
       "  [0.22925, 0.9375, 0.94192, 0.9453, 0.93856],\n",
       "  [0.18132, 0.94531, 0.94027, 0.94637, 0.93426],\n",
       "  [0.31899, 0.92188, 0.91293, 0.9137, 0.91216],\n",
       "  [0.29253, 0.89062, 0.89327, 0.89369, 0.89285],\n",
       "  [0.27586, 0.90625, 0.91186, 0.90737, 0.9164],\n",
       "  [0.27672, 0.89844, 0.90128, 0.89585, 0.90679],\n",
       "  [0.117, 0.96875, 0.96735, 0.96619, 0.9685],\n",
       "  [0.13872, 0.95312, 0.94724, 0.94167, 0.95287],\n",
       "  [0.22676, 0.94531, 0.9499, 0.95207, 0.94774],\n",
       "  [0.14039, 0.96094, 0.95657, 0.95813, 0.95502],\n",
       "  [0.16744, 0.9375, 0.93405, 0.93366, 0.93443],\n",
       "  [0.10918, 0.96875, 0.96739, 0.96482, 0.96997],\n",
       "  [0.28271, 0.92969, 0.93016, 0.92965, 0.93068],\n",
       "  [0.28988, 0.9375, 0.94032, 0.9433, 0.93737],\n",
       "  [0.2315, 0.94531, 0.92653, 0.93065, 0.92245],\n",
       "  [0.15288, 0.96875, 0.97061, 0.96662, 0.97464],\n",
       "  [0.2122, 0.91406, 0.91706, 0.92204, 0.91213],\n",
       "  [0.20062, 0.9375, 0.93183, 0.93372, 0.92994],\n",
       "  [0.28083, 0.91406, 0.90968, 0.90786, 0.91151],\n",
       "  [0.19526, 0.92188, 0.91896, 0.91706, 0.92087],\n",
       "  [0.11988, 0.94531, 0.9459, 0.94043, 0.95142],\n",
       "  [0.19889, 0.9375, 0.93932, 0.93708, 0.94157],\n",
       "  [0.2702, 0.89844, 0.89079, 0.89965, 0.8821],\n",
       "  [0.12941, 0.94531, 0.94559, 0.94364, 0.94756],\n",
       "  [0.19157, 0.95312, 0.96066, 0.96106, 0.96026],\n",
       "  [0.29164, 0.92188, 0.92702, 0.93214, 0.92196],\n",
       "  [0.30067, 0.92969, 0.92837, 0.93439, 0.92243],\n",
       "  [0.11571, 0.97656, 0.97863, 0.97745, 0.97981],\n",
       "  [0.19714, 0.9375, 0.94096, 0.9443, 0.93764],\n",
       "  [0.21781, 0.9375, 0.93196, 0.93268, 0.93125],\n",
       "  [0.10245, 0.96875, 0.96901, 0.96869, 0.96933],\n",
       "  [0.22243, 0.92969, 0.91546, 0.91708, 0.91384],\n",
       "  [0.17383, 0.95312, 0.94511, 0.94265, 0.94759],\n",
       "  [0.17243, 0.96094, 0.9643, 0.96418, 0.96443],\n",
       "  [0.25085, 0.92188, 0.91827, 0.91876, 0.91779],\n",
       "  [0.19008, 0.92969, 0.93702, 0.93395, 0.9401],\n",
       "  [0.182, 0.92969, 0.93248, 0.92901, 0.93598],\n",
       "  [0.18737, 0.94531, 0.95089, 0.94627, 0.95556],\n",
       "  [0.1326, 0.94531, 0.94454, 0.934, 0.95533],\n",
       "  [0.25907, 0.91406, 0.91345, 0.92761, 0.89971],\n",
       "  [0.14608, 0.96875, 0.96826, 0.97154, 0.965],\n",
       "  [0.18255, 0.96094, 0.96013, 0.96269, 0.95758],\n",
       "  [0.23126, 0.92188, 0.92937, 0.92268, 0.93616],\n",
       "  [0.10641, 0.97656, 0.97531, 0.97313, 0.9775],\n",
       "  [0.25282, 0.90625, 0.90591, 0.90564, 0.90618],\n",
       "  [0.16101, 0.9375, 0.93644, 0.93336, 0.93955],\n",
       "  [0.19938, 0.95312, 0.9561, 0.96071, 0.95153],\n",
       "  [0.26996, 0.92969, 0.93175, 0.93389, 0.92963],\n",
       "  [0.11488, 0.95312, 0.95537, 0.95531, 0.95543],\n",
       "  [0.12186, 0.97656, 0.97829, 0.98369, 0.97295],\n",
       "  [0.12325, 0.97656, 0.97367, 0.96869, 0.97871],\n",
       "  [0.28873, 0.89844, 0.89729, 0.9012, 0.89341],\n",
       "  [0.27654, 0.92969, 0.93082, 0.93351, 0.92815],\n",
       "  [0.27479, 0.92188, 0.91624, 0.90935, 0.92324],\n",
       "  [0.22102, 0.92969, 0.91626, 0.91533, 0.9172],\n",
       "  [0.12522, 0.96875, 0.96998, 0.96966, 0.9703],\n",
       "  [0.26228, 0.92969, 0.93819, 0.93606, 0.94033],\n",
       "  [0.25714, 0.91406, 0.92772, 0.92842, 0.92703],\n",
       "  [0.14246, 0.9375, 0.93561, 0.93495, 0.93626],\n",
       "  [0.17326, 0.9375, 0.93521, 0.93923, 0.93123],\n",
       "  [0.25451, 0.92969, 0.92325, 0.93584, 0.91099],\n",
       "  [0.23536, 0.91406, 0.92019, 0.92083, 0.91955],\n",
       "  [0.14282, 0.95312, 0.96065, 0.96046, 0.96083],\n",
       "  [0.21925, 0.94531, 0.94419, 0.94233, 0.94606],\n",
       "  [0.20952, 0.94531, 0.94957, 0.95078, 0.94836],\n",
       "  [0.24493, 0.92969, 0.93749, 0.9368, 0.93819],\n",
       "  [0.2207, 0.91406, 0.91433, 0.91549, 0.91317],\n",
       "  [0.10745, 0.96094, 0.96256, 0.96014, 0.965],\n",
       "  [0.38211, 0.91406, 0.9097, 0.90013, 0.91948],\n",
       "  [0.14526, 0.96875, 0.97072, 0.96787, 0.97359],\n",
       "  [0.25421, 0.90625, 0.91269, 0.90683, 0.91863],\n",
       "  [0.14118, 0.95312, 0.95374, 0.96083, 0.94674],\n",
       "  [0.23718, 0.95312, 0.95536, 0.95588, 0.95483],\n",
       "  [0.17515, 0.95312, 0.95386, 0.95872, 0.94904],\n",
       "  [0.20999, 0.94531, 0.93528, 0.932, 0.93859],\n",
       "  [0.13349, 0.95312, 0.94796, 0.93817, 0.95795],\n",
       "  [0.24357, 0.92188, 0.9274, 0.92554, 0.92927],\n",
       "  [0.28966, 0.89062, 0.88985, 0.89454, 0.88522],\n",
       "  [0.20285, 0.92969, 0.9331, 0.92767, 0.93859],\n",
       "  [0.16415, 0.9375, 0.93131, 0.92682, 0.93584],\n",
       "  [0.14632, 0.9375, 0.93136, 0.93763, 0.92518],\n",
       "  [0.19759, 0.94531, 0.94836, 0.94541, 0.95133],\n",
       "  [0.2593, 0.92969, 0.93631, 0.93792, 0.93472],\n",
       "  [0.25093, 0.92188, 0.92779, 0.915, 0.94095],\n",
       "  [0.2355, 0.96094, 0.9611, 0.9598, 0.96241],\n",
       "  [0.15226, 0.98438, 0.98526, 0.98889, 0.98167],\n",
       "  [0.19977, 0.95312, 0.95145, 0.94594, 0.95702],\n",
       "  [0.18385, 0.9375, 0.94144, 0.93789, 0.94501],\n",
       "  [0.23532, 0.94531, 0.94309, 0.94016, 0.94604],\n",
       "  [0.14173, 0.95312, 0.95691, 0.95698, 0.95684],\n",
       "  [0.15092, 0.95312, 0.95672, 0.9577, 0.95575],\n",
       "  [0.10705, 0.96094, 0.96268, 0.96533, 0.96005],\n",
       "  [0.14559, 0.95312, 0.95756, 0.95976, 0.95536],\n",
       "  [0.15783, 0.94531, 0.94136, 0.94002, 0.94272],\n",
       "  [0.09122, 0.96094, 0.95848, 0.95844, 0.95851],\n",
       "  [0.15039, 0.94531, 0.94306, 0.9435, 0.94262],\n",
       "  [0.19215, 0.91406, 0.92079, 0.92101, 0.92057],\n",
       "  [0.1323, 0.96875, 0.97272, 0.96944, 0.97601],\n",
       "  [0.11213, 0.98438, 0.98373, 0.98182, 0.98565],\n",
       "  [0.15898, 0.95312, 0.94715, 0.94831, 0.94599],\n",
       "  [0.10679, 0.95312, 0.94359, 0.94064, 0.94655],\n",
       "  [0.23292, 0.91406, 0.90633, 0.90927, 0.9034],\n",
       "  [0.16158, 0.94531, 0.95007, 0.95273, 0.94742],\n",
       "  [0.14013, 0.94531, 0.93811, 0.93392, 0.94234],\n",
       "  [0.11076, 0.96875, 0.96598, 0.97072, 0.96128],\n",
       "  [0.22302, 0.92969, 0.92084, 0.92052, 0.92116],\n",
       "  [0.17323, 0.95312, 0.96015, 0.96338, 0.95693],\n",
       "  [0.08297, 0.95312, 0.9561, 0.96059, 0.95165],\n",
       "  [0.21048, 0.92188, 0.92136, 0.92352, 0.91921],\n",
       "  [0.0349, 0.99219, 0.99479, 0.99545, 0.99412],\n",
       "  [0.1548, 0.94531, 0.94628, 0.94349, 0.94908],\n",
       "  [0.13307, 0.95312, 0.95193, 0.95467, 0.94921],\n",
       "  [0.11833, 0.95312, 0.95271, 0.94605, 0.95946],\n",
       "  [0.14751, 0.96875, 0.9693, 0.97113, 0.96747],\n",
       "  [0.18688, 0.9375, 0.93132, 0.93127, 0.93137],\n",
       "  [0.18435, 0.94531, 0.94883, 0.94841, 0.94924],\n",
       "  [0.16153, 0.94531, 0.94836, 0.9473, 0.94942],\n",
       "  [0.17848, 0.95312, 0.94662, 0.94886, 0.94439],\n",
       "  [0.11819, 0.96875, 0.9689, 0.9655, 0.97233],\n",
       "  [0.13447, 0.96875, 0.97067, 0.96859, 0.97276],\n",
       "  [0.31845, 0.92188, 0.91668, 0.92358, 0.90988],\n",
       "  [0.22313, 0.92969, 0.93166, 0.93547, 0.92789],\n",
       "  [0.26942, 0.91406, 0.91641, 0.91538, 0.91744],\n",
       "  [0.18469, 0.9375, 0.93404, 0.94443, 0.92388],\n",
       "  [0.08206, 0.98438, 0.98474, 0.98571, 0.98377],\n",
       "  [0.23729, 0.94531, 0.94774, 0.93962, 0.95599],\n",
       "  [0.13194, 0.96094, 0.9627, 0.96726, 0.95817],\n",
       "  [0.20729, 0.92188, 0.92252, 0.925, 0.92005],\n",
       "  [0.13778, 0.95312, 0.95794, 0.95659, 0.95928],\n",
       "  [0.1898, 0.94531, 0.94712, 0.95102, 0.94326],\n",
       "  [0.17177, 0.95312, 0.9535, 0.95445, 0.95256],\n",
       "  [0.18164, 0.94531, 0.95039, 0.95452, 0.94628],\n",
       "  [0.22233, 0.91406, 0.9139, 0.91181, 0.916],\n",
       "  [0.1465, 0.96875, 0.96759, 0.96536, 0.96983],\n",
       "  [0.11992, 0.95312, 0.95276, 0.94505, 0.96059],\n",
       "  [0.15651, 0.95312, 0.9547, 0.95258, 0.95683],\n",
       "  [0.10608, 0.97656, 0.97655, 0.97758, 0.97552],\n",
       "  [0.13858, 0.95312, 0.95986, 0.95777, 0.96195],\n",
       "  [0.20717, 0.9375, 0.94285, 0.93767, 0.9481],\n",
       "  [0.19256, 0.96094, 0.96117, 0.96398, 0.95837],\n",
       "  [0.24065, 0.90625, 0.90632, 0.90558, 0.90706],\n",
       "  [0.13865, 0.95312, 0.95711, 0.95047, 0.96384],\n",
       "  [0.08323, 0.97656, 0.97501, 0.96857, 0.98153],\n",
       "  [0.20164, 0.9375, 0.93152, 0.92919, 0.93386],\n",
       "  [0.11417, 0.96094, 0.96552, 0.96428, 0.96675],\n",
       "  [0.09235, 0.96094, 0.96023, 0.96266, 0.9578],\n",
       "  [0.3371, 0.89844, 0.90067, 0.90894, 0.89253],\n",
       "  [0.10144, 0.96094, 0.96359, 0.9674, 0.95981],\n",
       "  [0.21961, 0.92188, 0.92025, 0.91651, 0.92402],\n",
       "  [0.16092, 0.95312, 0.95243, 0.95498, 0.94989],\n",
       "  [0.1728, 0.94531, 0.94516, 0.94448, 0.94585],\n",
       "  [0.15006, 0.94531, 0.94459, 0.94773, 0.94147],\n",
       "  [0.20104, 0.94531, 0.94038, 0.9429, 0.93786],\n",
       "  [0.23058, 0.92188, 0.93145, 0.93089, 0.93201],\n",
       "  [0.13308, 0.9375, 0.9432, 0.94123, 0.94517],\n",
       "  [0.20423, 0.96875, 0.97175, 0.97669, 0.96685],\n",
       "  [0.15437, 0.96094, 0.95594, 0.95652, 0.95536],\n",
       "  [0.14458, 0.95312, 0.95588, 0.95504, 0.95672],\n",
       "  [0.16475, 0.92969, 0.93316, 0.93591, 0.93044],\n",
       "  [0.22312, 0.92188, 0.92274, 0.92177, 0.9237],\n",
       "  [0.19394, 0.96094, 0.95934, 0.96065, 0.95803],\n",
       "  [0.11761, 0.94531, 0.94942, 0.95876, 0.94026],\n",
       "  [0.09094, 0.97656, 0.98075, 0.98042, 0.98109],\n",
       "  [0.20895, 0.9375, 0.94017, 0.9392, 0.94114],\n",
       "  [0.262, 0.9375, 0.93848, 0.93292, 0.94411],\n",
       "  [0.20293, 0.91406, 0.91227, 0.91068, 0.91387],\n",
       "  [0.191, 0.94531, 0.95294, 0.95524, 0.95065],\n",
       "  [0.28739, 0.94531, 0.94445, 0.9358, 0.95325],\n",
       "  [0.11322, 0.98438, 0.98439, 0.98182, 0.98697],\n",
       "  [0.15415, 0.96875, 0.97084, 0.97015, 0.97152],\n",
       "  [0.09883, 0.97656, 0.97609, 0.97468, 0.97752],\n",
       "  [0.11595, 0.96875, 0.96909, 0.96953, 0.96864],\n",
       "  [0.22231, 0.92969, 0.9227, 0.92099, 0.92443],\n",
       "  [0.11513, 0.96094, 0.9656, 0.96673, 0.96447],\n",
       "  [0.17659, 0.94531, 0.9466, 0.94511, 0.9481],\n",
       "  [0.19009, 0.96094, 0.96207, 0.96081, 0.96333],\n",
       "  [0.15143, 0.96094, 0.96025, 0.96258, 0.95794],\n",
       "  [0.08029, 0.97656, 0.97714, 0.97788, 0.97639],\n",
       "  [0.17032, 0.95312, 0.94304, 0.94295, 0.94314],\n",
       "  [0.28702, 0.91406, 0.91484, 0.91518, 0.91451],\n",
       "  [0.14164, 0.96875, 0.96894, 0.96578, 0.97212],\n",
       "  [0.14983, 0.95312, 0.95627, 0.96231, 0.95031],\n",
       "  [0.18635, 0.94531, 0.94841, 0.94688, 0.94994],\n",
       "  [0.11034, 0.96875, 0.96782, 0.9736, 0.9621],\n",
       "  [0.14598, 0.9375, 0.93905, 0.94122, 0.9369],\n",
       "  [0.04328, 0.98438, 0.98556, 0.98452, 0.98661],\n",
       "  [0.15751, 0.94531, 0.94234, 0.94321, 0.94147],\n",
       "  [0.10848, 0.96875, 0.96881, 0.96766, 0.96997],\n",
       "  [0.16496, 0.96094, 0.95543, 0.95927, 0.95162],\n",
       "  [0.25764, 0.92188, 0.9229, 0.91988, 0.92595],\n",
       "  [0.23372, 0.9375, 0.93763, 0.93753, 0.93773],\n",
       "  [0.15563, 0.95312, 0.95517, 0.95688, 0.95347],\n",
       "  [0.28427, 0.95312, 0.95664, 0.95232, 0.96099],\n",
       "  [0.28529, 0.9375, 0.93589, 0.93585, 0.93593],\n",
       "  [0.18746, 0.94531, 0.94624, 0.95014, 0.94238],\n",
       "  [0.1794, 0.91406, 0.90952, 0.90594, 0.91313],\n",
       "  [0.06108, 0.98438, 0.98536, 0.98412, 0.98661],\n",
       "  [0.06068, 0.99219, 0.99226, 0.99286, 0.99167],\n",
       "  [0.19956, 0.94531, 0.93377, 0.93453, 0.933],\n",
       "  [0.20901, 0.9375, 0.94009, 0.94116, 0.93903],\n",
       "  [0.06967, 0.98438, 0.98463, 0.98503, 0.98424],\n",
       "  [0.15047, 0.96094, 0.96227, 0.96984, 0.95482],\n",
       "  [0.11898, 0.96875, 0.96438, 0.97336, 0.95556],\n",
       "  [0.05881, 0.98438, 0.98037, 0.97333, 0.9875],\n",
       "  [0.14661, 0.9375, 0.93386, 0.92913, 0.93864],\n",
       "  [0.17507, 0.9375, 0.93284, 0.93197, 0.93372],\n",
       "  [0.21881, 0.94531, 0.94281, 0.93839, 0.94728],\n",
       "  [0.1029, 0.97656, 0.97869, 0.98009, 0.97731],\n",
       "  [0.08928, 0.97656, 0.97781, 0.98031, 0.97531],\n",
       "  [0.25943, 0.9375, 0.93718, 0.93272, 0.94169],\n",
       "  [0.13768, 0.94531, 0.93797, 0.93426, 0.9417],\n",
       "  [0.18633, 0.96094, 0.96621, 0.96271, 0.96974],\n",
       "  [0.17053, 0.95312, 0.95251, 0.95556, 0.94948],\n",
       "  [0.08435, 0.96875, 0.96906, 0.97161, 0.96653],\n",
       "  [0.12241, 0.95312, 0.95031, 0.94495, 0.95573],\n",
       "  [0.07569, 0.98438, 0.98605, 0.98333, 0.98879],\n",
       "  [0.08691, 0.95312, 0.94368, 0.94498, 0.94239],\n",
       "  [0.20141, 0.9375, 0.93941, 0.94778, 0.93119],\n",
       "  [0.25562, 0.95312, 0.9544, 0.95655, 0.95227],\n",
       "  [0.07333, 0.97656, 0.97794, 0.9755, 0.9804],\n",
       "  [0.23732, 0.95312, 0.95645, 0.95503, 0.95787],\n",
       "  [0.23254, 0.92188, 0.91837, 0.92035, 0.91639],\n",
       "  [0.24567, 0.92969, 0.92627, 0.92387, 0.9287],\n",
       "  [0.20252, 0.92969, 0.93446, 0.93237, 0.93656],\n",
       "  [0.08086, 0.98438, 0.98389, 0.98444, 0.98333],\n",
       "  [0.13522, 0.95312, 0.95257, 0.95266, 0.95248],\n",
       "  [0.24517, 0.92969, 0.93192, 0.93958, 0.92439],\n",
       "  [0.16372, 0.94531, 0.9429, 0.94336, 0.94244],\n",
       "  [0.22025, 0.92188, 0.925, 0.92531, 0.9247],\n",
       "  [0.09775, 0.98438, 0.98243, 0.98542, 0.97946],\n",
       "  [0.13982, 0.94531, 0.94462, 0.94567, 0.94357],\n",
       "  [0.18941, 0.95312, 0.95491, 0.95643, 0.9534],\n",
       "  [0.17739, 0.96875, 0.96411, 0.96474, 0.96348],\n",
       "  [0.2178, 0.91406, 0.91321, 0.90723, 0.91927],\n",
       "  [0.10647, 0.96875, 0.97142, 0.97388, 0.96898],\n",
       "  [0.15768, 0.9375, 0.93043, 0.93239, 0.92847],\n",
       "  [0.12063, 0.95312, 0.94987, 0.94877, 0.95097],\n",
       "  [0.11943, 0.96875, 0.96677, 0.96579, 0.96774],\n",
       "  [0.1284, 0.96094, 0.96211, 0.95833, 0.96591],\n",
       "  [0.1467, 0.96094, 0.9646, 0.96583, 0.96337],\n",
       "  [0.07602, 0.97656, 0.97376, 0.97495, 0.97258],\n",
       "  [0.13234, 0.96094, 0.96205, 0.96372, 0.96039],\n",
       "  [0.20963, 0.92969, 0.92983, 0.93844, 0.92137],\n",
       "  [0.13183, 0.94531, 0.94232, 0.9456, 0.93906],\n",
       "  [0.25682, 0.95312, 0.94604, 0.93896, 0.95322],\n",
       "  [0.11606, 0.98438, 0.98564, 0.98462, 0.98667],\n",
       "  [0.18859, 0.92188, 0.91979, 0.92531, 0.91433],\n",
       "  [0.10189, 0.95312, 0.9562, 0.95453, 0.95788],\n",
       "  [0.05781, 0.97656, 0.97599, 0.97341, 0.97857],\n",
       "  [0.19479, 0.92969, 0.92299, 0.91855, 0.92748],\n",
       "  [0.07935, 0.99219, 0.99045, 0.99, 0.99091],\n",
       "  [0.21211, 0.92969, 0.92864, 0.92646, 0.93083],\n",
       "  [0.16648, 0.91667, 0.92553, 0.91982, 0.9313],\n",
       "  [0.13992, 0.96875, 0.96766, 0.96739, 0.96794],\n",
       "  [0.11843, 0.96875, 0.96944, 0.96313, 0.97583],\n",
       "  [0.19397, 0.92969, 0.92549, 0.91484, 0.93639],\n",
       "  [0.0982, 0.97656, 0.97734, 0.97877, 0.97591],\n",
       "  [0.20356, 0.94922, 0.94508, 0.94581, 0.94434],\n",
       "  [0.15983, 0.95312, 0.95195, 0.94964, 0.95428],\n",
       "  [0.1438, 0.95703, 0.95737, 0.95662, 0.95811],\n",
       "  [0.17115, 0.95703, 0.95789, 0.95861, 0.95718],\n",
       "  [0.09668, 0.96875, 0.96803, 0.96839, 0.96766],\n",
       "  [0.18907, 0.96094, 0.95861, 0.9623, 0.95494],\n",
       "  [0.18896, 0.9375, 0.93692, 0.9383, 0.93555],\n",
       "  [0.13659, 0.96094, 0.96093, 0.96026, 0.96161],\n",
       "  [0.16445, 0.95833, 0.95927, 0.96102, 0.95753],\n",
       "  [0.1145, 0.96354, 0.96366, 0.96383, 0.96349],\n",
       "  [0.1475, 0.95833, 0.95807, 0.95821, 0.95793],\n",
       "  [0.11427, 0.96615, 0.96614, 0.96696, 0.96533],\n",
       "  [0.16231, 0.93229, 0.93102, 0.93208, 0.92996],\n",
       "  [0.0895, 0.98177, 0.98243, 0.98245, 0.98241],\n",
       "  [0.12313, 0.96875, 0.9706, 0.96918, 0.97202],\n",
       "  [0.15282, 0.96094, 0.96051, 0.96045, 0.96058],\n",
       "  [0.12901, 0.96354, 0.96433, 0.96314, 0.96551],\n",
       "  [0.12493, 0.95833, 0.95754, 0.95519, 0.95991],\n",
       "  [0.08909, 0.97917, 0.97886, 0.9785, 0.97921],\n",
       "  [0.13305, 0.96615, 0.96596, 0.964, 0.96792],\n",
       "  [0.11055, 0.97135, 0.97113, 0.96978, 0.97249],\n",
       "  [0.11744, 0.96354, 0.96508, 0.96394, 0.96622],\n",
       "  [0.12423, 0.96354, 0.96279, 0.96312, 0.96246],\n",
       "  [0.13601, 0.94531, 0.94543, 0.94577, 0.94508],\n",
       "  [0.09769, 0.97917, 0.97884, 0.97821, 0.97948],\n",
       "  [0.16818, 0.95833, 0.95964, 0.96078, 0.9585],\n",
       "  [0.09001, 0.96875, 0.96692, 0.96091, 0.97302],\n",
       "  [0.14717, 0.95703, 0.95597, 0.95753, 0.95442],\n",
       "  [0.15452, 0.95703, 0.95799, 0.95708, 0.95891],\n",
       "  [0.09498, 0.97656, 0.97711, 0.97649, 0.97774],\n",
       "  [0.18981, 0.94922, 0.94745, 0.9472, 0.94769],\n",
       "  [0.12628, 0.96484, 0.96531, 0.96487, 0.96576],\n",
       "  [0.1397, 0.95898, 0.95872, 0.95852, 0.95892],\n",
       "  [0.13701, 0.95703, 0.95724, 0.95801, 0.95648],\n",
       "  [0.11803, 0.96289, 0.96281, 0.96252, 0.96309],\n",
       "  [0.13142, 0.9668, 0.96748, 0.96721, 0.96774],\n",
       "  [0.13429, 0.95312, 0.95272, 0.95333, 0.9521],\n",
       "  [0.12258, 0.9668, 0.96839, 0.96903, 0.96774],\n",
       "  [0.13197, 0.96484, 0.96486, 0.96593, 0.9638],\n",
       "  [0.10128, 0.97266, 0.97312, 0.97236, 0.97389],\n",
       "  [0.10097, 0.96484, 0.96544, 0.96441, 0.96647],\n",
       "  [0.10733, 0.9668, 0.9679, 0.96784, 0.96796],\n",
       "  [0.14485, 0.95703, 0.95667, 0.95577, 0.95757],\n",
       "  [0.16623, 0.94336, 0.94205, 0.9418, 0.94229],\n",
       "  [0.13822, 0.96289, 0.9622, 0.96328, 0.96112],\n",
       "  [0.09232, 0.98047, 0.98071, 0.98054, 0.98088],\n",
       "  [0.13265, 0.96484, 0.96369, 0.96311, 0.96428],\n",
       "  [0.08942, 0.97656, 0.97708, 0.97653, 0.97763],\n",
       "  [0.12686, 0.96289, 0.96416, 0.963, 0.96534],\n",
       "  [0.11639, 0.9668, 0.96642, 0.96511, 0.96772],\n",
       "  [0.128, 0.96289, 0.96396, 0.96269, 0.96525],\n",
       "  [0.14284, 0.96289, 0.96355, 0.96236, 0.96475],\n",
       "  [0.10597, 0.9668, 0.96487, 0.964, 0.96574],\n",
       "  [0.10964, 0.96094, 0.96208, 0.96214, 0.96201],\n",
       "  [0.10337, 0.96094, 0.96178, 0.96194, 0.96163],\n",
       "  [0.16793, 0.95508, 0.95494, 0.95434, 0.95554],\n",
       "  [0.12354, 0.96094, 0.96193, 0.96262, 0.96124],\n",
       "  [0.12315, 0.96289, 0.96377, 0.96423, 0.96331],\n",
       "  [0.17844, 0.95898, 0.95998, 0.95815, 0.96182],\n",
       "  [0.15105, 0.95508, 0.95519, 0.95539, 0.95499],\n",
       "  [0.15476, 0.95312, 0.95224, 0.95255, 0.95194],\n",
       "  [0.10408, 0.96484, 0.9622, 0.96095, 0.96344],\n",
       "  [0.12927, 0.96484, 0.96485, 0.96459, 0.96512],\n",
       "  [0.123, 0.96094, 0.96047, 0.95986, 0.96109],\n",
       "  [0.13803, 0.97852, 0.97818, 0.97758, 0.97878],\n",
       "  [0.08592, 0.97461, 0.97222, 0.97064, 0.97381],\n",
       "  [0.13614, 0.96289, 0.96102, 0.96192, 0.96013],\n",
       "  [0.14543, 0.96289, 0.96169, 0.96066, 0.96272],\n",
       "  [0.0717, 0.98047, 0.98025, 0.97994, 0.98056],\n",
       "  [0.10462, 0.96875, 0.96901, 0.96842, 0.9696],\n",
       "  [0.16641, 0.96484, 0.96593, 0.96528, 0.96658],\n",
       "  [0.11548, 0.97461, 0.97303, 0.97274, 0.97333],\n",
       "  [0.09797, 0.96875, 0.96674, 0.96551, 0.96797],\n",
       "  [0.11391, 0.9707, 0.97046, 0.97094, 0.96998],\n",
       "  [0.1338, 0.95703, 0.95925, 0.95983, 0.95867],\n",
       "  [0.08111, 0.97461, 0.97418, 0.9733, 0.97507],\n",
       "  [0.10359, 0.96289, 0.9631, 0.96289, 0.96332],\n",
       "  [0.14427, 0.94727, 0.94639, 0.94513, 0.94765],\n",
       "  [0.09106, 0.9707, 0.97221, 0.97178, 0.97263],\n",
       "  [0.12568, 0.96289, 0.96213, 0.96162, 0.96265],\n",
       "  [0.12275, 0.9668, 0.96665, 0.96549, 0.96782],\n",
       "  [0.11987, 0.95312, 0.95488, 0.95444, 0.95532],\n",
       "  [0.14927, 0.95117, 0.95108, 0.94991, 0.95225],\n",
       "  [0.12747, 0.94922, 0.94913, 0.94811, 0.95015],\n",
       "  [0.1553, 0.94922, 0.94857, 0.94922, 0.94792],\n",
       "  [0.10742, 0.95898, 0.95946, 0.96026, 0.95866],\n",
       "  [0.08176, 0.97461, 0.97276, 0.97255, 0.97297],\n",
       "  [0.13562, 0.95508, 0.95324, 0.95309, 0.9534],\n",
       "  [0.12478, 0.96875, 0.9693, 0.96903, 0.96957],\n",
       "  [0.11048, 0.96719, 0.96517, 0.96555, 0.96478],\n",
       "  [0.13109, 0.96406, 0.96492, 0.96451, 0.96533],\n",
       "  [0.10587, 0.96875, 0.96895, 0.96766, 0.97023],\n",
       "  [0.08998, 0.97969, 0.97812, 0.97792, 0.97832],\n",
       "  [0.10673, 0.97031, 0.96906, 0.9675, 0.97063],\n",
       "  [0.09457, 0.97656, 0.9766, 0.97557, 0.97762],\n",
       "  [0.08672, 0.97656, 0.97665, 0.9767, 0.9766],\n",
       "  [0.1194, 0.96094, 0.96176, 0.96404, 0.95949],\n",
       "  [0.09701, 0.97031, 0.96895, 0.96953, 0.96837],\n",
       "  [0.13419, 0.95781, 0.95813, 0.95918, 0.95708],\n",
       "  [0.12233, 0.96562, 0.96486, 0.96469, 0.96504],\n",
       "  [0.10948, 0.97031, 0.97071, 0.96891, 0.97251],\n",
       "  [0.08726, 0.975, 0.9749, 0.97535, 0.97445],\n",
       "  [0.13356, 0.95469, 0.9551, 0.95583, 0.95438],\n",
       "  [0.09729, 0.96719, 0.96673, 0.96627, 0.96719],\n",
       "  [0.10651, 0.96406, 0.96432, 0.96322, 0.96542],\n",
       "  [0.09892, 0.97188, 0.97126, 0.97196, 0.97056],\n",
       "  [0.08586, 0.97188, 0.97204, 0.97181, 0.97227],\n",
       "  [0.09452, 0.96719, 0.96657, 0.96638, 0.96677],\n",
       "  [0.09648, 0.97344, 0.97358, 0.97337, 0.9738],\n",
       "  [0.09244, 0.97188, 0.97049, 0.97034, 0.97064],\n",
       "  [0.08571, 0.98125, 0.98139, 0.98216, 0.98062],\n",
       "  [0.11731, 0.97188, 0.97216, 0.97186, 0.97245],\n",
       "  [0.13956, 0.95625, 0.95663, 0.95737, 0.95588],\n",
       "  [0.10259, 0.97344, 0.97169, 0.97146, 0.97192],\n",
       "  [0.1233, 0.95625, 0.95681, 0.95756, 0.95607],\n",
       "  [0.08694, 0.97813, 0.97777, 0.97843, 0.9771],\n",
       "  [0.11854, 0.97031, 0.97042, 0.97108, 0.96976],\n",
       "  [0.12015, 0.9625, 0.96276, 0.96072, 0.96481],\n",
       "  [0.08896, 0.97656, 0.97847, 0.97857, 0.97837],\n",
       "  [0.08223, 0.96094, 0.95781, 0.95874, 0.95688],\n",
       "  [0.08005, 0.97656, 0.97762, 0.97911, 0.97613],\n",
       "  [0.1118, 0.97656, 0.97713, 0.97643, 0.97783],\n",
       "  [0.05693, 0.98177, 0.98195, 0.98144, 0.98246],\n",
       "  [0.09384, 0.97396, 0.97346, 0.97506, 0.97187],\n",
       "  [0.13936, 0.95312, 0.95284, 0.95222, 0.95347],\n",
       "  [0.11841, 0.96094, 0.96195, 0.95827, 0.96565],\n",
       "  [0.12071, 0.9668, 0.96791, 0.96768, 0.96815],\n",
       "  [0.0583, 0.98242, 0.98077, 0.98119, 0.98035],\n",
       "  [0.09423, 0.98047, 0.98051, 0.98112, 0.9799],\n",
       "  [0.07484, 0.97656, 0.9755, 0.97529, 0.97571],\n",
       "  [0.07146, 0.97852, 0.97785, 0.9777, 0.978],\n",
       "  [0.10423, 0.95898, 0.95793, 0.95802, 0.95784],\n",
       "  [0.11734, 0.96875, 0.96816, 0.96833, 0.96798],\n",
       "  [0.05907, 0.97969, 0.98012, 0.98059, 0.97966],\n",
       "  [0.07617, 0.98125, 0.98078, 0.98109, 0.98046],\n",
       "  [0.07388, 0.97969, 0.97967, 0.97974, 0.97959],\n",
       "  [0.08774, 0.96875, 0.96815, 0.96752, 0.96878],\n",
       "  [0.10402, 0.96562, 0.96692, 0.96622, 0.96761],\n",
       "  [0.08029, 0.97344, 0.97328, 0.97265, 0.97392],\n",
       "  [0.07854, 0.975, 0.97547, 0.97484, 0.9761],\n",
       "  [0.06083, 0.98438, 0.98407, 0.98471, 0.98343],\n",
       "  [0.09328, 0.97188, 0.97114, 0.96978, 0.9725],\n",
       "  [0.06193, 0.98177, 0.98117, 0.98103, 0.98131],\n",
       "  [0.11135, 0.96745, 0.96692, 0.96677, 0.96706],\n",
       "  [0.07812, 0.98047, 0.98064, 0.98124, 0.98004],\n",
       "  [0.08862, 0.97656, 0.9769, 0.9768, 0.977],\n",
       "  [0.10211, 0.97786, 0.97704, 0.97692, 0.97716],\n",
       "  [0.08939, 0.96484, 0.96512, 0.96558, 0.96467],\n",
       "  [0.11977, 0.97266, 0.97293, 0.97364, 0.97223],\n",
       "  [0.07653, 0.97266, 0.97147, 0.97118, 0.97177],\n",
       "  [0.07723, 0.97786, 0.9782, 0.9785, 0.9779],\n",
       "  [0.07174, 0.98828, 0.98828, 0.98809, 0.98847],\n",
       "  [0.11259, 0.96875, 0.96812, 0.96699, 0.96924],\n",
       "  [0.12657, 0.96094, 0.9579, 0.9584, 0.9574],\n",
       "  [0.08554, 0.97526, 0.97501, 0.97521, 0.97481],\n",
       "  [0.08284, 0.97135, 0.97176, 0.97211, 0.9714],\n",
       "  [0.10644, 0.96875, 0.96812, 0.9674, 0.96884],\n",
       "  [0.09739, 0.96615, 0.9654, 0.96577, 0.96502],\n",
       "  [0.05463, 0.98828, 0.98841, 0.98854, 0.98829],\n",
       "  [0.04794, 0.98177, 0.98184, 0.98261, 0.98107],\n",
       "  [0.06995, 0.98047, 0.98049, 0.98077, 0.98021],\n",
       "  [0.06608, 0.97526, 0.97562, 0.97511, 0.97613],\n",
       "  [0.08218, 0.97786, 0.97738, 0.9766, 0.97816],\n",
       "  [0.06854, 0.97396, 0.97379, 0.97411, 0.97347],\n",
       "  [0.09164, 0.97917, 0.97905, 0.97919, 0.97891],\n",
       "  [0.12255, 0.96354, 0.96326, 0.96348, 0.96303],\n",
       "  [0.09219, 0.97656, 0.97659, 0.97711, 0.97607],\n",
       "  [0.09164, 0.97433, 0.97473, 0.97416, 0.9753],\n",
       "  [0.07541, 0.97768, 0.97758, 0.97768, 0.97748],\n",
       "  [0.06854, 0.97768, 0.97758, 0.97774, 0.97741],\n",
       "  [0.091, 0.9721, 0.97121, 0.9704, 0.97202],\n",
       "  [0.1104, 0.97098, 0.97004, 0.96976, 0.97032],\n",
       "  [0.05989, 0.98438, 0.98386, 0.98393, 0.9838],\n",
       "  [0.1102, 0.95871, 0.95741, 0.95745, 0.95737],\n",
       "  [0.06598, 0.98103, 0.9809, 0.98126, 0.98055],\n",
       "  [0.10018, 0.96763, 0.96727, 0.96616, 0.96839],\n",
       "  [0.09177, 0.97321, 0.97258, 0.972, 0.97316],\n",
       "  [0.10003, 0.96763, 0.96723, 0.96716, 0.9673],\n",
       "  [0.08671, 0.98103, 0.97889, 0.97937, 0.97842],\n",
       "  [0.06911, 0.97768, 0.97753, 0.97748, 0.97759],\n",
       "  [0.10564, 0.96763, 0.96866, 0.9681, 0.96922],\n",
       "  [0.07505, 0.97768, 0.97797, 0.97725, 0.9787],\n",
       "  [0.0763, 0.97433, 0.97317, 0.97266, 0.97369],\n",
       "  [0.12637, 0.95647, 0.95669, 0.95661, 0.95676],\n",
       "  [0.08314, 0.97768, 0.97692, 0.97684, 0.977],\n",
       "  [0.07282, 0.97656, 0.97663, 0.97626, 0.977],\n",
       "  [0.10312, 0.97098, 0.97104, 0.97159, 0.97049],\n",
       "  [0.11643, 0.9654, 0.96575, 0.96592, 0.96559],\n",
       "  [0.09604, 0.97098, 0.971, 0.97045, 0.97156],\n",
       "  [0.10897, 0.96875, 0.9689, 0.96907, 0.96873],\n",
       "  [0.11337, 0.9721, 0.97287, 0.97329, 0.97244],\n",
       "  [0.09706, 0.97433, 0.97336, 0.97294, 0.97378],\n",
       "  [0.07713, 0.97879, 0.97897, 0.9795, 0.97845],\n",
       "  [0.09566, 0.96875, 0.96896, 0.96877, 0.96915],\n",
       "  [0.08336, 0.97545, 0.97576, 0.97597, 0.97554],\n",
       "  [0.0874, 0.97433, 0.9739, 0.97382, 0.97397],\n",
       "  [0.08213, 0.97098, 0.97073, 0.97113, 0.97034],\n",
       "  [0.07714, 0.97545, 0.97473, 0.9748, 0.97467],\n",
       "  [0.08668, 0.97545, 0.97553, 0.97587, 0.97519],\n",
       "  [0.08101, 0.9838, 0.98359, 0.9837, 0.98348]],\n",
       " 'validation_log': [],\n",
       " 'test_metrics': [],\n",
       " 'experiment_id': 'experiment',\n",
       " 'best_checkpoint': 'best_experiment_chk.tar',\n",
       " 'latest_checkpoint': 'latest_experiment_chk.tar',\n",
       " 'best_val_epoch': 0,\n",
       " 'best_val_score': 0.0,\n",
       " 'name': 'experiment',\n",
       " 'split_dir': 'net_logs/experiment/splits'}"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "log"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}