bert_base_결과정리.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "res1 = '''\n",
    "1번째 모델\n",
    "[Epoch 1 Batch 50/482] loss=2.8177, lr=0.0000116667, acc=0.917\n",
    "[Epoch 1 Batch 100/482] loss=0.8137, lr=0.0000220833, acc=0.950\n",
    "[Epoch 1 Batch 150/482] loss=1.0392, lr=0.0000325000, acc=0.959\n",
    "[Epoch 1 Batch 200/482] loss=1.5004, lr=0.0000429167, acc=0.963\n",
    "[Epoch 1 Batch 250/482] loss=2.6393, lr=0.0000496307, acc=0.962\n",
    "[Epoch 1 Batch 300/482] loss=1.7526, lr=0.0000484765, acc=0.963\n",
    "[Epoch 1 Batch 350/482] loss=2.0649, lr=0.0000473223, acc=0.962\n",
    "[Epoch 1 Batch 400/482] loss=2.3171, lr=0.0000461681, acc=0.959\n",
    "[Epoch 1 Batch 450/482] loss=1.3803, lr=0.0000450139, acc=0.961\n",
    "\n",
    "Valid acc : 0.9848091404829914\n",
    "[Epoch 2 Batch 50/482] loss=1.0960, lr=0.0000431210, acc=0.982\n",
    "[Epoch 2 Batch 100/482] loss=1.4387, lr=0.0000419668, acc=0.974\n",
    "[Epoch 2 Batch 150/482] loss=0.8660, lr=0.0000408126, acc=0.978\n",
    "[Epoch 2 Batch 200/482] loss=0.4934, lr=0.0000396584, acc=0.981\n",
    "[Epoch 2 Batch 250/482] loss=1.2836, lr=0.0000385042, acc=0.978\n",
    "[Epoch 2 Batch 300/482] loss=1.5405, lr=0.0000373500, acc=0.976\n",
    "[Epoch 2 Batch 350/482] loss=0.7342, lr=0.0000361958, acc=0.977\n",
    "[Epoch 2 Batch 400/482] loss=0.7901, lr=0.0000350416, acc=0.978\n",
    "[Epoch 2 Batch 450/482] loss=0.5907, lr=0.0000338873, acc=0.980\n",
    "\n",
    "Valid acc : 0.990651778758764\n",
    "[Epoch 3 Batch 50/482] loss=0.6875, lr=0.0000319945, acc=0.989\n",
    "[Epoch 3 Batch 100/482] loss=0.7305, lr=0.0000308403, acc=0.988\n",
    "[Epoch 3 Batch 150/482] loss=0.7184, lr=0.0000296861, acc=0.989\n",
    "[Epoch 3 Batch 200/482] loss=0.6591, lr=0.0000285319, acc=0.989\n",
    "[Epoch 3 Batch 250/482] loss=0.5190, lr=0.0000273777, acc=0.990\n",
    "[Epoch 3 Batch 300/482] loss=0.5215, lr=0.0000262235, acc=0.990\n",
    "[Epoch 3 Batch 350/482] loss=0.3310, lr=0.0000250693, acc=0.990\n",
    "[Epoch 3 Batch 400/482] loss=0.7567, lr=0.0000239151, acc=0.990\n",
    "[Epoch 3 Batch 450/482] loss=0.5504, lr=0.0000227608, acc=0.990\n",
    "\n",
    "Valid acc : 0.9902622695403791\n",
    "[Epoch 4 Batch 50/482] loss=0.4662, lr=0.0000208680, acc=0.989\n",
    "[Epoch 4 Batch 100/482] loss=1.9933, lr=0.0000197138, acc=0.972\n",
    "[Epoch 4 Batch 150/482] loss=0.5637, lr=0.0000185596, acc=0.979\n",
    "[Epoch 4 Batch 200/482] loss=0.5407, lr=0.0000174054, acc=0.982\n",
    "[Epoch 4 Batch 250/482] loss=0.4607, lr=0.0000162512, acc=0.984\n",
    "[Epoch 4 Batch 300/482] loss=0.3885, lr=0.0000150970, acc=0.985\n",
    "[Epoch 4 Batch 350/482] loss=0.2421, lr=0.0000139428, acc=0.987\n",
    "[Epoch 4 Batch 400/482] loss=0.5118, lr=0.0000127886, acc=0.987\n",
    "[Epoch 4 Batch 450/482] loss=0.3347, lr=0.0000116343, acc=0.988\n",
    "\n",
    "Valid acc : 0.9919501428200468\n",
    "[Epoch 5 Batch 50/482] loss=0.4171, lr=0.0000097415, acc=0.993\n",
    "[Epoch 5 Batch 100/482] loss=0.2651, lr=0.0000085873, acc=0.994\n",
    "[Epoch 5 Batch 150/482] loss=0.2967, lr=0.0000074331, acc=0.994\n",
    "[Epoch 5 Batch 200/482] loss=0.4506, lr=0.0000062789, acc=0.994\n",
    "[Epoch 5 Batch 250/482] loss=0.2461, lr=0.0000051247, acc=0.994\n",
    "[Epoch 5 Batch 300/482] loss=0.3101, lr=0.0000039705, acc=0.994\n",
    "[Epoch 5 Batch 350/482] loss=0.2095, lr=0.0000028163, acc=0.995\n",
    "[Epoch 5 Batch 400/482] loss=0.5680, lr=0.0000016620, acc=0.994\n",
    "[Epoch 5 Batch 450/482] loss=0.3786, lr=0.0000005078, acc=0.994\n",
    "\n",
    "Valid acc : 0.994417034536484\n",
    "'''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss1 = [float(s.split('=')[1][:-1]) \n",
    "         for s in res1.split(' ') if 'loss' in s]\n",
    "acc1 = [float(s.split('=')[1][:5]) \n",
    "         for s in res1.split(' ') if 'acc=' in s]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "res2 = \"\"\"\n",
    "2번째 모델\n",
    "[Epoch 1 Batch 50/482] loss=3.5441, lr=0.0000104167, acc=0.875\n",
    "[Epoch 1 Batch 100/482] loss=1.0443, lr=0.0000208333, acc=0.925\n",
    "[Epoch 1 Batch 150/482] loss=1.1543, lr=0.0000312500, acc=0.943\n",
    "[Epoch 1 Batch 200/482] loss=0.7825, lr=0.0000416667, acc=0.953\n",
    "[Epoch 1 Batch 250/482] loss=1.0256, lr=0.0000497692, acc=0.958\n",
    "[Epoch 1 Batch 300/482] loss=0.6390, lr=0.0000486150, acc=0.963\n",
    "[Epoch 1 Batch 350/482] loss=0.7012, lr=0.0000474608, acc=0.966\n",
    "[Epoch 1 Batch 400/482] loss=1.0592, lr=0.0000463066, acc=0.969\n",
    "[Epoch 1 Batch 450/482] loss=0.8977, lr=0.0000451524, acc=0.970\n",
    "Valid acc : 0.9880550506361984\n",
    "\n",
    "[Epoch 2 Batch 50/482] loss=0.7283, lr=0.0000432595, acc=0.989\n",
    "[Epoch 2 Batch 100/482] loss=0.3041, lr=0.0000421053, acc=0.993\n",
    "[Epoch 2 Batch 150/482] loss=0.2937, lr=0.0000409511, acc=0.994\n",
    "[Epoch 2 Batch 200/482] loss=0.3019, lr=0.0000397969, acc=0.994\n",
    "[Epoch 2 Batch 250/482] loss=0.3084, lr=0.0000386427, acc=0.994\n",
    "[Epoch 2 Batch 300/482] loss=0.3414, lr=0.0000374885, acc=0.994\n",
    "[Epoch 2 Batch 350/482] loss=0.2156, lr=0.0000363343, acc=0.995\n",
    "[Epoch 2 Batch 400/482] loss=0.2637, lr=0.0000351801, acc=0.995\n",
    "[Epoch 2 Batch 450/482] loss=0.1297, lr=0.0000340259, acc=0.995\n",
    "Valid acc : 0.9977927810958193\n",
    "\n",
    "[Epoch 3 Batch 50/482] loss=0.1937, lr=0.0000321330, acc=0.998\n",
    "[Epoch 3 Batch 100/482] loss=0.2067, lr=0.0000309788, acc=0.997\n",
    "[Epoch 3 Batch 150/482] loss=0.2210, lr=0.0000298246, acc=0.997\n",
    "[Epoch 3 Batch 200/482] loss=0.1601, lr=0.0000286704, acc=0.997\n",
    "[Epoch 3 Batch 250/482] loss=0.1683, lr=0.0000275162, acc=0.997\n",
    "[Epoch 3 Batch 300/482] loss=0.0536, lr=0.0000263620, acc=0.997\n",
    "[Epoch 3 Batch 350/482] loss=0.1713, lr=0.0000252078, acc=0.997\n",
    "[Epoch 3 Batch 400/482] loss=0.1471, lr=0.0000240536, acc=0.997\n",
    "[Epoch 3 Batch 450/482] loss=0.1057, lr=0.0000228994, acc=0.998\n",
    "Valid acc : 0.9987016359387172\n",
    "\n",
    "[Epoch 4 Batch 50/482] loss=0.1292, lr=0.0000210065, acc=0.998\n",
    "[Epoch 4 Batch 100/482] loss=0.0740, lr=0.0000198523, acc=0.999\n",
    "[Epoch 4 Batch 150/482] loss=0.1718, lr=0.0000186981, acc=0.998\n",
    "[Epoch 4 Batch 200/482] loss=0.1616, lr=0.0000175439, acc=0.998\n",
    "[Epoch 4 Batch 250/482] loss=0.0678, lr=0.0000163897, acc=0.999\n",
    "[Epoch 4 Batch 300/482] loss=0.0460, lr=0.0000152355, acc=0.999\n",
    "[Epoch 4 Batch 350/482] loss=0.1444, lr=0.0000140813, acc=0.999\n",
    "[Epoch 4 Batch 400/482] loss=0.0651, lr=0.0000129271, acc=0.999\n",
    "[Epoch 4 Batch 450/482] loss=0.0661, lr=0.0000117729, acc=0.999\n",
    "Valid acc : 0.9984419631264606\n",
    "\n",
    "[Epoch 5 Batch 50/482] loss=0.0589, lr=0.0000098800, acc=0.999\n",
    "[Epoch 5 Batch 100/482] loss=0.0574, lr=0.0000087258, acc=0.999\n",
    "[Epoch 5 Batch 150/482] loss=0.0975, lr=0.0000075716, acc=0.999\n",
    "[Epoch 5 Batch 200/482] loss=0.1672, lr=0.0000064174, acc=0.999\n",
    "[Epoch 5 Batch 250/482] loss=0.0471, lr=0.0000052632, acc=0.999\n",
    "[Epoch 5 Batch 300/482] loss=0.0332, lr=0.0000041090, acc=0.999\n",
    "[Epoch 5 Batch 350/482] loss=0.1358, lr=0.0000029548, acc=0.999\n",
    "[Epoch 5 Batch 400/482] loss=0.0193, lr=0.0000018006, acc=0.999\n",
    "[Epoch 5 Batch 450/482] loss=0.0220, lr=0.0000006464, acc=0.999\n",
    "Valid acc : 0.9990911451571021\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss2 = [float(s.split('=')[1][:-1]) \n",
    "         for s in res2.split(' ') if 'loss' in s]\n",
    "acc2 = [float(s.split('=')[1][:5]) \n",
    "         for s in res2.split(' ') if 'acc=' in s]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "res3 = \"\"\"\n",
    "3번째 모델\n",
    "[Epoch 1 Batch 50/482] loss=3.8170, lr=0.0000104167, acc=0.867\n",
    "[Epoch 1 Batch 100/482] loss=0.9981, lr=0.0000208333, acc=0.922\n",
    "[Epoch 1 Batch 150/482] loss=1.3358, lr=0.0000312500, acc=0.940\n",
    "[Epoch 1 Batch 200/482] loss=1.6636, lr=0.0000416667, acc=0.948\n",
    "[Epoch 1 Batch 250/482] loss=2.7325, lr=0.0000497692, acc=0.948\n",
    "[Epoch 1 Batch 300/482] loss=1.0373, lr=0.0000486150, acc=0.953\n",
    "[Epoch 1 Batch 350/482] loss=2.2668, lr=0.0000474608, acc=0.953\n",
    "[Epoch 1 Batch 400/482] loss=1.2171, lr=0.0000463066, acc=0.957\n",
    "[Epoch 1 Batch 450/482] loss=0.8506, lr=0.0000451524, acc=0.960\n",
    "Valid acc : 0.9909114515710206\n",
    "\n",
    "[Epoch 2 Batch 50/482] loss=0.6607, lr=0.0000432595, acc=0.990\n",
    "[Epoch 2 Batch 100/482] loss=0.6432, lr=0.0000421053, acc=0.990\n",
    "[Epoch 2 Batch 150/482] loss=0.8168, lr=0.0000409511, acc=0.989\n",
    "[Epoch 2 Batch 200/482] loss=0.6697, lr=0.0000397969, acc=0.989\n",
    "[Epoch 2 Batch 250/482] loss=0.5338, lr=0.0000386427, acc=0.990\n",
    "[Epoch 2 Batch 300/482] loss=0.3851, lr=0.0000374885, acc=0.990\n",
    "[Epoch 2 Batch 350/482] loss=0.6019, lr=0.0000363343, acc=0.991\n",
    "[Epoch 2 Batch 400/482] loss=0.4914, lr=0.0000351801, acc=0.991\n",
    "[Epoch 2 Batch 450/482] loss=0.2387, lr=0.0000340259, acc=0.991\n",
    "Valid acc : 0.9955855621916385\n",
    "\n",
    "[Epoch 3 Batch 50/482] loss=0.3225, lr=0.0000321330, acc=0.995\n",
    "[Epoch 3 Batch 100/482] loss=0.2251, lr=0.0000309788, acc=0.996\n",
    "[Epoch 3 Batch 150/482] loss=0.2363, lr=0.0000298246, acc=0.996\n",
    "[Epoch 3 Batch 200/482] loss=0.3101, lr=0.0000286704, acc=0.996\n",
    "[Epoch 3 Batch 250/482] loss=0.3617, lr=0.0000275162, acc=0.996\n",
    "[Epoch 3 Batch 300/482] loss=0.1518, lr=0.0000263620, acc=0.996\n",
    "[Epoch 3 Batch 350/482] loss=0.3892, lr=0.0000252078, acc=0.996\n",
    "[Epoch 3 Batch 400/482] loss=0.3505, lr=0.0000240536, acc=0.996\n",
    "[Epoch 3 Batch 450/482] loss=0.4472, lr=0.0000228994, acc=0.995\n",
    "Valid acc : 0.99623474422228\n",
    "\n",
    "[Epoch 4 Batch 50/482] loss=0.3545, lr=0.0000210065, acc=0.995\n",
    "[Epoch 4 Batch 100/482] loss=0.1820, lr=0.0000198523, acc=0.996\n",
    "[Epoch 4 Batch 150/482] loss=0.2593, lr=0.0000186981, acc=0.996\n",
    "[Epoch 4 Batch 200/482] loss=0.3618, lr=0.0000175439, acc=0.996\n",
    "[Epoch 4 Batch 250/482] loss=0.2971, lr=0.0000163897, acc=0.996\n",
    "[Epoch 4 Batch 300/482] loss=0.1173, lr=0.0000152355, acc=0.996\n",
    "[Epoch 4 Batch 350/482] loss=0.2699, lr=0.0000140813, acc=0.996\n",
    "[Epoch 4 Batch 400/482] loss=0.1911, lr=0.0000129271, acc=0.996\n",
    "[Epoch 4 Batch 450/482] loss=0.1706, lr=0.0000117729, acc=0.997\n",
    "Valid acc : 0.9964944170345365\n",
    "\n",
    "[Epoch 5 Batch 50/482] loss=0.1658, lr=0.0000098800, acc=0.997\n",
    "[Epoch 5 Batch 100/482] loss=0.1113, lr=0.0000087258, acc=0.997\n",
    "[Epoch 5 Batch 150/482] loss=0.1734, lr=0.0000075716, acc=0.997\n",
    "[Epoch 5 Batch 200/482] loss=0.1799, lr=0.0000064174, acc=0.997\n",
    "[Epoch 5 Batch 250/482] loss=0.2133, lr=0.0000052632, acc=0.997\n",
    "[Epoch 5 Batch 300/482] loss=0.0511, lr=0.0000041090, acc=0.997\n",
    "[Epoch 5 Batch 350/482] loss=0.0829, lr=0.0000029548, acc=0.998\n",
    "[Epoch 5 Batch 400/482] loss=0.1330, lr=0.0000018006, acc=0.998\n",
    "[Epoch 5 Batch 450/482] loss=0.0631, lr=0.0000006464, acc=0.998\n",
    "Valid acc : 0.9974032718774344\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss3 = [float(s.split('=')[1][:-1]) \n",
    "         for s in res3.split(' ') if 'loss' in s]\n",
    "acc3 = [float(s.split('=')[1][:5]) \n",
    "         for s in res3.split(' ') if 'acc=' in s]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "res4 = \"\"\"\n",
    "4번째 모델\n",
    "[Epoch 1 Batch 50/482] loss=3.8812, lr=0.0000104167, acc=0.853\n",
    "[Epoch 1 Batch 100/482] loss=0.8690, lr=0.0000208333, acc=0.918\n",
    "[Epoch 1 Batch 150/482] loss=1.0083, lr=0.0000312500, acc=0.939\n",
    "[Epoch 1 Batch 200/482] loss=0.8451, lr=0.0000416667, acc=0.950\n",
    "[Epoch 1 Batch 250/482] loss=0.7176, lr=0.0000497692, acc=0.958\n",
    "[Epoch 1 Batch 300/482] loss=0.5100, lr=0.0000486150, acc=0.963\n",
    "[Epoch 1 Batch 350/482] loss=0.5244, lr=0.0000474608, acc=0.967\n",
    "[Epoch 1 Batch 400/482] loss=0.3760, lr=0.0000463066, acc=0.970\n",
    "[Epoch 1 Batch 450/482] loss=0.6435, lr=0.0000451524, acc=0.973\n",
    "Valid acc : 0.994417034536484\n",
    "\n",
    "[Epoch 2 Batch 50/482] loss=0.4209, lr=0.0000432595, acc=0.994\n",
    "[Epoch 2 Batch 100/482] loss=0.2485, lr=0.0000421053, acc=0.996\n",
    "[Epoch 2 Batch 150/482] loss=0.4208, lr=0.0000409511, acc=0.995\n",
    "[Epoch 2 Batch 200/482] loss=0.3251, lr=0.0000397969, acc=0.995\n",
    "[Epoch 2 Batch 250/482] loss=3.5034, lr=0.0000386427, acc=0.984\n",
    "[Epoch 2 Batch 300/482] loss=1.4873, lr=0.0000374885, acc=0.983\n",
    "[Epoch 2 Batch 350/482] loss=0.2906, lr=0.0000363343, acc=0.985\n",
    "[Epoch 2 Batch 400/482] loss=0.2800, lr=0.0000351801, acc=0.987\n",
    "[Epoch 2 Batch 450/482] loss=0.2205, lr=0.0000340259, acc=0.988\n",
    "Valid acc : 0.9961049078161517\n",
    "\n",
    "[Epoch 3 Batch 50/482] loss=0.1944, lr=0.0000321330, acc=0.998\n",
    "[Epoch 3 Batch 100/482] loss=0.1501, lr=0.0000309788, acc=0.998\n",
    "[Epoch 3 Batch 150/482] loss=0.1397, lr=0.0000298246, acc=0.998\n",
    "[Epoch 3 Batch 200/482] loss=0.1423, lr=0.0000286704, acc=0.998\n",
    "[Epoch 3 Batch 250/482] loss=0.3607, lr=0.0000275162, acc=0.997\n",
    "[Epoch 3 Batch 300/482] loss=0.3089, lr=0.0000263620, acc=0.997\n",
    "[Epoch 3 Batch 350/482] loss=0.9833, lr=0.0000252078, acc=0.994\n",
    "[Epoch 3 Batch 400/482] loss=9.0017, lr=0.0000240536, acc=0.936\n",
    "[Epoch 3 Batch 450/482] loss=8.9251, lr=0.0000228994, acc=0.888\n",
    "Valid acc : 0.5212931706050377\n",
    "\n",
    "[Epoch 4 Batch 50/482] loss=8.8475, lr=0.0000210065, acc=0.510\n",
    "[Epoch 4 Batch 100/482] loss=8.7146, lr=0.0000198523, acc=0.514\n",
    "[Epoch 4 Batch 150/482] loss=8.7426, lr=0.0000186981, acc=0.511\n",
    "[Epoch 4 Batch 200/482] loss=8.6324, lr=0.0000175439, acc=0.512\n",
    "[Epoch 4 Batch 250/482] loss=3.2634, lr=0.0000163897, acc=0.592\n",
    "[Epoch 4 Batch 300/482] loss=0.6205, lr=0.0000152355, acc=0.658\n",
    "[Epoch 4 Batch 350/482] loss=0.1900, lr=0.0000140813, acc=0.707\n",
    "[Epoch 4 Batch 400/482] loss=0.1415, lr=0.0000129271, acc=0.743\n",
    "[Epoch 4 Batch 450/482] loss=0.1733, lr=0.0000117729, acc=0.771\n",
    "Valid acc : 0.9948065437548689\n",
    "\n",
    "[Epoch 5 Batch 50/482] loss=0.2898, lr=0.0000098800, acc=0.996\n",
    "[Epoch 5 Batch 100/482] loss=0.1309, lr=0.0000087258, acc=0.997\n",
    "[Epoch 5 Batch 150/482] loss=0.1833, lr=0.0000075716, acc=0.997\n",
    "[Epoch 5 Batch 200/482] loss=0.0473, lr=0.0000064174, acc=0.998\n",
    "[Epoch 5 Batch 250/482] loss=0.1615, lr=0.0000052632, acc=0.998\n",
    "[Epoch 5 Batch 300/482] loss=0.0906, lr=0.0000041090, acc=0.998\n",
    "[Epoch 5 Batch 350/482] loss=0.0678, lr=0.0000029548, acc=0.998\n",
    "[Epoch 5 Batch 400/482] loss=0.1051, lr=0.0000018006, acc=0.998\n",
    "[Epoch 5 Batch 450/482] loss=0.1045, lr=0.0000006464, acc=0.998\n",
    "Valid acc : 0.9961049078161517\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss4 = [float(s.split('=')[1][:-1]) \n",
    "         for s in res4.split(' ') if 'loss' in s]\n",
    "acc4 = [float(s.split('=')[1][:5]) \n",
    "         for s in res4.split(' ') if 'acc=' in s]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "res5 = \"\"\"\n",
    "5번째 모델\n",
    "[Epoch 1 Batch 50/482] loss=3.8099, lr=0.0000104167, acc=0.853\n",
    "[Epoch 1 Batch 100/482] loss=1.1901, lr=0.0000208333, acc=0.914\n",
    "[Epoch 1 Batch 150/482] loss=1.2645, lr=0.0000312500, acc=0.935\n",
    "[Epoch 1 Batch 200/482] loss=0.6752, lr=0.0000416667, acc=0.948\n",
    "[Epoch 1 Batch 250/482] loss=1.2337, lr=0.0000497692, acc=0.955\n",
    "[Epoch 1 Batch 300/482] loss=1.3267, lr=0.0000486150, acc=0.957\n",
    "[Epoch 1 Batch 350/482] loss=0.8844, lr=0.0000474608, acc=0.961\n",
    "[Epoch 1 Batch 400/482] loss=1.5784, lr=0.0000463066, acc=0.962\n",
    "[Epoch 1 Batch 450/482] loss=1.1958, lr=0.0000451524, acc=0.965\n",
    "Valid acc : 0.993897688911971\n",
    "\n",
    "[Epoch 2 Batch 50/482] loss=0.3180, lr=0.0000432595, acc=0.995\n",
    "[Epoch 2 Batch 100/482] loss=0.2949, lr=0.0000421053, acc=0.994\n",
    "[Epoch 2 Batch 150/482] loss=0.2772, lr=0.0000409511, acc=0.994\n",
    "[Epoch 2 Batch 200/482] loss=0.2677, lr=0.0000397969, acc=0.995\n",
    "[Epoch 2 Batch 250/482] loss=0.4357, lr=0.0000386427, acc=0.995\n",
    "[Epoch 2 Batch 300/482] loss=0.2068, lr=0.0000374885, acc=0.995\n",
    "[Epoch 2 Batch 350/482] loss=0.2843, lr=0.0000363343, acc=0.995\n",
    "[Epoch 2 Batch 400/482] loss=0.3077, lr=0.0000351801, acc=0.995\n",
    "[Epoch 2 Batch 450/482] loss=0.2628, lr=0.0000340259, acc=0.995\n",
    "Valid acc : 0.9963645806284082\n",
    "\n",
    "[Epoch 3 Batch 50/482] loss=0.2743, lr=0.0000321330, acc=0.995\n",
    "[Epoch 3 Batch 100/482] loss=0.1660, lr=0.0000309788, acc=0.996\n",
    "[Epoch 3 Batch 150/482] loss=0.2012, lr=0.0000298246, acc=0.996\n",
    "[Epoch 3 Batch 200/482] loss=0.0173, lr=0.0000286704, acc=0.997\n",
    "[Epoch 3 Batch 250/482] loss=0.2099, lr=0.0000275162, acc=0.997\n",
    "[Epoch 3 Batch 300/482] loss=0.0926, lr=0.0000263620, acc=0.997\n",
    "[Epoch 3 Batch 350/482] loss=0.1264, lr=0.0000252078, acc=0.997\n",
    "[Epoch 3 Batch 400/482] loss=0.1456, lr=0.0000240536, acc=0.997\n",
    "[Epoch 3 Batch 450/482] loss=0.0982, lr=0.0000228994, acc=0.998\n",
    "Valid acc : 0.9970137626590496\n",
    "\n",
    "[Epoch 4 Batch 50/482] loss=0.1433, lr=0.0000210065, acc=0.998\n",
    "[Epoch 4 Batch 100/482] loss=0.0414, lr=0.0000198523, acc=0.999\n",
    "[Epoch 4 Batch 150/482] loss=0.0444, lr=0.0000186981, acc=0.999\n",
    "[Epoch 4 Batch 200/482] loss=0.0688, lr=0.0000175439, acc=0.999\n",
    "[Epoch 4 Batch 250/482] loss=0.1600, lr=0.0000163897, acc=0.999\n",
    "[Epoch 4 Batch 300/482] loss=0.0619, lr=0.0000152355, acc=0.999\n",
    "[Epoch 4 Batch 350/482] loss=0.0599, lr=0.0000140813, acc=0.999\n",
    "[Epoch 4 Batch 400/482] loss=0.0266, lr=0.0000129271, acc=0.999\n",
    "[Epoch 4 Batch 450/482] loss=0.0347, lr=0.0000117729, acc=0.999\n",
    "Valid acc : 0.9981822903142041\n",
    "\n",
    "[Epoch 5 Batch 50/482] loss=0.0631, lr=0.0000098800, acc=0.999\n",
    "[Epoch 5 Batch 100/482] loss=0.0184, lr=0.0000087258, acc=1.000\n",
    "[Epoch 5 Batch 150/482] loss=0.0069, lr=0.0000075716, acc=1.000\n",
    "[Epoch 5 Batch 200/482] loss=0.0514, lr=0.0000064174, acc=1.000\n",
    "[Epoch 5 Batch 250/482] loss=0.1134, lr=0.0000052632, acc=0.999\n",
    "[Epoch 5 Batch 300/482] loss=0.0274, lr=0.0000041090, acc=0.999\n",
    "[Epoch 5 Batch 350/482] loss=0.0373, lr=0.0000029548, acc=0.999\n",
    "[Epoch 5 Batch 400/482] loss=0.0070, lr=0.0000018006, acc=1.000\n",
    "[Epoch 5 Batch 450/482] loss=0.0351, lr=0.0000006464, acc=1.000\n",
    "Valid acc : 0.997662944689691\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss5 = [float(s.split('=')[1][:-1]) \n",
    "         for s in res5.split(' ') if 'loss' in s]\n",
    "acc5 = [float(s.split('=')[1][:5]) \n",
    "         for s in res5.split(' ') if 'acc=' in s]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "res6 = \"\"\"\n",
    "6번째 모델\n",
    "[Epoch 1 Batch 50/482] loss=4.0271, lr=0.0000104167, acc=0.851\n",
    "[Epoch 1 Batch 100/482] loss=1.0716, lr=0.0000208333, acc=0.914\n",
    "[Epoch 1 Batch 150/482] loss=0.9675, lr=0.0000312500, acc=0.937\n",
    "[Epoch 1 Batch 200/482] loss=1.1207, lr=0.0000416667, acc=0.948\n",
    "[Epoch 1 Batch 250/482] loss=1.0070, lr=0.0000497692, acc=0.955\n",
    "[Epoch 1 Batch 300/482] loss=0.6044, lr=0.0000486150, acc=0.960\n",
    "[Epoch 1 Batch 350/482] loss=0.4437, lr=0.0000474608, acc=0.965\n",
    "[Epoch 1 Batch 400/482] loss=0.4285, lr=0.0000463066, acc=0.968\n",
    "[Epoch 1 Batch 450/482] loss=0.3808, lr=0.0000451524, acc=0.971\n",
    "Valid acc : 0.9954557257855102\n",
    "\n",
    "[Epoch 2 Batch 50/482] loss=0.3618, lr=0.0000432595, acc=0.994\n",
    "[Epoch 2 Batch 100/482] loss=0.1229, lr=0.0000421053, acc=0.996\n",
    "[Epoch 2 Batch 150/482] loss=0.2504, lr=0.0000409511, acc=0.996\n",
    "[Epoch 2 Batch 200/482] loss=0.1889, lr=0.0000397969, acc=0.996\n",
    "[Epoch 2 Batch 250/482] loss=0.2015, lr=0.0000386427, acc=0.996\n",
    "[Epoch 2 Batch 300/482] loss=0.1173, lr=0.0000374885, acc=0.997\n",
    "[Epoch 2 Batch 350/482] loss=0.2784, lr=0.0000363343, acc=0.996\n",
    "[Epoch 2 Batch 400/482] loss=0.1259, lr=0.0000351801, acc=0.997\n",
    "[Epoch 2 Batch 450/482] loss=0.1208, lr=0.0000340259, acc=0.997\n",
    "Valid acc : 0.9972734354713062\n",
    "\n",
    "[Epoch 3 Batch 50/482] loss=0.2252, lr=0.0000321330, acc=0.997\n",
    "[Epoch 3 Batch 100/482] loss=0.0354, lr=0.0000309788, acc=0.998\n",
    "[Epoch 3 Batch 150/482] loss=0.0798, lr=0.0000298246, acc=0.998\n",
    "[Epoch 3 Batch 200/482] loss=0.0867, lr=0.0000286704, acc=0.998\n",
    "[Epoch 3 Batch 250/482] loss=0.0832, lr=0.0000275162, acc=0.998\n",
    "[Epoch 3 Batch 300/482] loss=0.0778, lr=0.0000263620, acc=0.998\n",
    "[Epoch 3 Batch 350/482] loss=0.0583, lr=0.0000252078, acc=0.998\n",
    "[Epoch 3 Batch 400/482] loss=0.0150, lr=0.0000240536, acc=0.998\n",
    "[Epoch 3 Batch 450/482] loss=0.0496, lr=0.0000228994, acc=0.998\n",
    "Valid acc : 0.9983121267203324\n",
    "\n",
    "[Epoch 4 Batch 50/482] loss=0.0413, lr=0.0000210065, acc=0.999\n",
    "[Epoch 4 Batch 100/482] loss=0.0197, lr=0.0000198523, acc=1.000\n",
    "[Epoch 4 Batch 150/482] loss=0.0385, lr=0.0000186981, acc=0.999\n",
    "[Epoch 4 Batch 200/482] loss=0.0012, lr=0.0000175439, acc=1.000\n",
    "[Epoch 4 Batch 250/482] loss=0.0095, lr=0.0000163897, acc=1.000\n",
    "[Epoch 4 Batch 300/482] loss=0.0112, lr=0.0000152355, acc=1.000\n",
    "[Epoch 4 Batch 350/482] loss=0.0016, lr=0.0000140813, acc=1.000\n",
    "[Epoch 4 Batch 400/482] loss=0.0014, lr=0.0000129271, acc=1.000\n",
    "[Epoch 4 Batch 450/482] loss=0.0426, lr=0.0000117729, acc=1.000\n",
    "Valid acc : 0.9990911451571021\n",
    "\n",
    "[Epoch 5 Batch 50/482] loss=0.0580, lr=0.0000098800, acc=0.999\n",
    "[Epoch 5 Batch 100/482] loss=0.0013, lr=0.0000087258, acc=1.000\n",
    "[Epoch 5 Batch 150/482] loss=0.0052, lr=0.0000075716, acc=1.000\n",
    "[Epoch 5 Batch 200/482] loss=0.0009, lr=0.0000064174, acc=1.000\n",
    "[Epoch 5 Batch 250/482] loss=0.0031, lr=0.0000052632, acc=1.000\n",
    "[Epoch 5 Batch 300/482] loss=0.0056, lr=0.0000041090, acc=1.000\n",
    "[Epoch 5 Batch 350/482] loss=0.0143, lr=0.0000029548, acc=1.000\n",
    "[Epoch 5 Batch 400/482] loss=0.0006, lr=0.0000018006, acc=1.000\n",
    "[Epoch 5 Batch 450/482] loss=0.0008, lr=0.0000006464, acc=1.000\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "# It is Good!!\n",
    "loss6 = [float(s.split('=')[1][:-1]) \n",
    "         for s in res6.split(' ') if 'loss' in s]\n",
    "acc6 = [float(s.split('=')[1][:5]) \n",
    "         for s in res6.split(' ') if 'acc=' in s]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "res7 = \"\"\"\n",
    "7번째 모델\n",
    "[Epoch 1 Batch 50/482] loss=3.0777, lr=0.0000104167, acc=0.882\n",
    "[Epoch 1 Batch 100/482] loss=0.6366, lr=0.0000208333, acc=0.933\n",
    "[Epoch 1 Batch 150/482] loss=0.9645, lr=0.0000312500, acc=0.948\n",
    "[Epoch 1 Batch 200/482] loss=0.4753, lr=0.0000416667, acc=0.958\n",
    "[Epoch 1 Batch 250/482] loss=0.5887, lr=0.0000497692, acc=0.964\n",
    "[Epoch 1 Batch 300/482] loss=0.9321, lr=0.0000486150, acc=0.968\n",
    "[Epoch 1 Batch 350/482] loss=0.8215, lr=0.0000474608, acc=0.971\n",
    "[Epoch 1 Batch 400/482] loss=1.5146, lr=0.0000463066, acc=0.970\n",
    "[Epoch 1 Batch 450/482] loss=2.3125, lr=0.0000451524, acc=0.964\n",
    "Valid acc : 0.9846793040768631\n",
    "\n",
    "[Epoch 2 Batch 50/482] loss=0.6378, lr=0.0000432595, acc=0.989\n",
    "[Epoch 2 Batch 100/482] loss=0.3549, lr=0.0000421053, acc=0.991\n",
    "[Epoch 2 Batch 150/482] loss=0.1809, lr=0.0000409511, acc=0.993\n",
    "[Epoch 2 Batch 200/482] loss=0.2128, lr=0.0000397969, acc=0.994\n",
    "[Epoch 2 Batch 250/482] loss=0.4988, lr=0.0000386427, acc=0.994\n",
    "[Epoch 2 Batch 300/482] loss=0.5001, lr=0.0000374885, acc=0.994\n",
    "[Epoch 2 Batch 350/482] loss=0.2846, lr=0.0000363343, acc=0.994\n",
    "[Epoch 2 Batch 400/482] loss=0.2769, lr=0.0000351801, acc=0.994\n",
    "[Epoch 2 Batch 450/482] loss=0.2855, lr=0.0000340259, acc=0.995\n",
    "Valid acc : 0.9961049078161517\n",
    "\n",
    "[Epoch 3 Batch 50/482] loss=0.0933, lr=0.0000321330, acc=0.999\n",
    "[Epoch 3 Batch 100/482] loss=0.2136, lr=0.0000309788, acc=0.997\n",
    "[Epoch 3 Batch 150/482] loss=0.1468, lr=0.0000298246, acc=0.997\n",
    "[Epoch 3 Batch 200/482] loss=0.1008, lr=0.0000286704, acc=0.998\n",
    "[Epoch 3 Batch 250/482] loss=0.1644, lr=0.0000275162, acc=0.998\n",
    "[Epoch 3 Batch 300/482] loss=0.2260, lr=0.0000263620, acc=0.998\n",
    "[Epoch 3 Batch 350/482] loss=0.1623, lr=0.0000252078, acc=0.998\n",
    "[Epoch 3 Batch 400/482] loss=0.1321, lr=0.0000240536, acc=0.998\n",
    "[Epoch 3 Batch 450/482] loss=0.1952, lr=0.0000228994, acc=0.998\n",
    "Valid acc : 0.9972734354713062\n",
    "\n",
    "[Epoch 4 Batch 50/482] loss=0.0492, lr=0.0000210065, acc=0.999\n",
    "[Epoch 4 Batch 100/482] loss=0.0701, lr=0.0000198523, acc=0.999\n",
    "[Epoch 4 Batch 150/482] loss=0.0956, lr=0.0000186981, acc=0.999\n",
    "[Epoch 4 Batch 200/482] loss=0.0415, lr=0.0000175439, acc=0.999\n",
    "[Epoch 4 Batch 250/482] loss=0.0713, lr=0.0000163897, acc=0.999\n",
    "[Epoch 4 Batch 300/482] loss=0.1001, lr=0.0000152355, acc=0.999\n",
    "[Epoch 4 Batch 350/482] loss=0.1187, lr=0.0000140813, acc=0.999\n",
    "[Epoch 4 Batch 400/482] loss=0.0497, lr=0.0000129271, acc=0.999\n",
    "[Epoch 4 Batch 450/482] loss=0.1553, lr=0.0000117729, acc=0.999\n",
    "Valid acc : 0.9974032718774344\n",
    "\n",
    "[Epoch 5 Batch 50/482] loss=0.0072, lr=0.0000098800, acc=1.000\n",
    "[Epoch 5 Batch 100/482] loss=0.0266, lr=0.0000087258, acc=1.000\n",
    "[Epoch 5 Batch 150/482] loss=0.0502, lr=0.0000075716, acc=1.000\n",
    "[Epoch 5 Batch 200/482] loss=0.0205, lr=0.0000064174, acc=1.000\n",
    "[Epoch 5 Batch 250/482] loss=0.1121, lr=0.0000052632, acc=0.999\n",
    "[Epoch 5 Batch 300/482] loss=0.0828, lr=0.0000041090, acc=0.999\n",
    "[Epoch 5 Batch 350/482] loss=0.0329, lr=0.0000029548, acc=0.999\n",
    "[Epoch 5 Batch 400/482] loss=0.0195, lr=0.0000018006, acc=0.999\n",
    "[Epoch 5 Batch 450/482] loss=0.0066, lr=0.0000006464, acc=0.999\n",
    "Valid acc : 0.997662944689691\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss7 = [float(s.split('=')[1][:-1]) \n",
    "         for s in res7.split(' ') if 'loss' in s]\n",
    "acc7 = [float(s.split('=')[1][:5]) \n",
    "         for s in res7.split(' ') if 'acc=' in s]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "res8 = \"\"\"\n",
    "8번째 모델\n",
    "[Epoch 1 Batch 50/482] loss=4.0845, lr=0.0000104167, acc=0.845\n",
    "[Epoch 1 Batch 100/482] loss=1.0578, lr=0.0000208333, acc=0.910\n",
    "[Epoch 1 Batch 150/482] loss=0.9985, lr=0.0000312500, acc=0.933\n",
    "[Epoch 1 Batch 200/482] loss=0.7639, lr=0.0000416667, acc=0.946\n",
    "[Epoch 1 Batch 250/482] loss=1.0080, lr=0.0000497692, acc=0.954\n",
    "[Epoch 1 Batch 300/482] loss=0.9886, lr=0.0000486150, acc=0.959\n",
    "[Epoch 1 Batch 350/482] loss=1.9851, lr=0.0000474608, acc=0.959\n",
    "[Epoch 1 Batch 400/482] loss=2.0315, lr=0.0000463066, acc=0.959\n",
    "[Epoch 1 Batch 450/482] loss=1.4507, lr=0.0000451524, acc=0.960\n",
    "Valid acc : 0.966891716437289\n",
    "\n",
    "[Epoch 2 Batch 50/482] loss=1.1074, lr=0.0000432595, acc=0.980\n",
    "[Epoch 2 Batch 100/482] loss=0.9131, lr=0.0000421053, acc=0.982\n",
    "[Epoch 2 Batch 150/482] loss=0.8680, lr=0.0000409511, acc=0.984\n",
    "[Epoch 2 Batch 200/482] loss=0.6553, lr=0.0000397969, acc=0.985\n",
    "[Epoch 2 Batch 250/482] loss=0.4964, lr=0.0000386427, acc=0.987\n",
    "[Epoch 2 Batch 300/482] loss=0.4414, lr=0.0000374885, acc=0.988\n",
    "[Epoch 2 Batch 350/482] loss=0.3751, lr=0.0000363343, acc=0.989\n",
    "[Epoch 2 Batch 400/482] loss=0.4120, lr=0.0000351801, acc=0.989\n",
    "[Epoch 2 Batch 450/482] loss=0.2841, lr=0.0000340259, acc=0.990\n",
    "Valid acc : 0.9971435990651779\n",
    "\n",
    "[Epoch 3 Batch 50/482] loss=0.2210, lr=0.0000321330, acc=0.996\n",
    "[Epoch 3 Batch 100/482] loss=0.0950, lr=0.0000309788, acc=0.997\n",
    "[Epoch 3 Batch 150/482] loss=0.1345, lr=0.0000298246, acc=0.997\n",
    "[Epoch 3 Batch 200/482] loss=0.1714, lr=0.0000286704, acc=0.997\n",
    "[Epoch 3 Batch 250/482] loss=0.1864, lr=0.0000275162, acc=0.997\n",
    "[Epoch 3 Batch 300/482] loss=0.2517, lr=0.0000263620, acc=0.997\n",
    "[Epoch 3 Batch 350/482] loss=0.2048, lr=0.0000252078, acc=0.997\n",
    "[Epoch 3 Batch 400/482] loss=0.2129, lr=0.0000240536, acc=0.997\n",
    "[Epoch 3 Batch 450/482] loss=0.1032, lr=0.0000228994, acc=0.997\n",
    "Valid acc : 0.9972734354713062\n",
    "\n",
    "[Epoch 4 Batch 50/482] loss=0.1069, lr=0.0000210065, acc=0.998\n",
    "[Epoch 4 Batch 100/482] loss=0.0354, lr=0.0000198523, acc=0.998\n",
    "[Epoch 4 Batch 150/482] loss=0.0885, lr=0.0000186981, acc=0.998\n",
    "[Epoch 4 Batch 200/482] loss=0.1135, lr=0.0000175439, acc=0.998\n",
    "[Epoch 4 Batch 250/482] loss=0.0930, lr=0.0000163897, acc=0.999\n",
    "[Epoch 4 Batch 300/482] loss=0.0792, lr=0.0000152355, acc=0.999\n",
    "[Epoch 4 Batch 350/482] loss=0.1179, lr=0.0000140813, acc=0.999\n",
    "[Epoch 4 Batch 400/482] loss=0.1687, lr=0.0000129271, acc=0.998\n",
    "[Epoch 4 Batch 450/482] loss=0.1106, lr=0.0000117729, acc=0.998\n",
    "Valid acc : 0.9974032718774344\n",
    "\n",
    "[Epoch 5 Batch 50/482] loss=0.0552, lr=0.0000098800, acc=0.999\n",
    "[Epoch 5 Batch 100/482] loss=0.0111, lr=0.0000087258, acc=0.999\n",
    "[Epoch 5 Batch 150/482] loss=0.0202, lr=0.0000075716, acc=0.999\n",
    "[Epoch 5 Batch 200/482] loss=0.0897, lr=0.0000064174, acc=0.999\n",
    "[Epoch 5 Batch 250/482] loss=0.0310, lr=0.0000052632, acc=0.999\n",
    "[Epoch 5 Batch 300/482] loss=0.0548, lr=0.0000041090, acc=0.999\n",
    "[Epoch 5 Batch 350/482] loss=0.0694, lr=0.0000029548, acc=0.999\n",
    "[Epoch 5 Batch 400/482] loss=0.0432, lr=0.0000018006, acc=0.999\n",
    "[Epoch 5 Batch 450/482] loss=0.0550, lr=0.0000006464, acc=0.999\n",
    "Valid acc : 0.997662944689691\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss8 = [float(s.split('=')[1][:-1]) \n",
    "         for s in res8.split(' ') if 'loss' in s]\n",
    "acc8 = [float(s.split('=')[1][:5]) \n",
    "         for s in res8.split(' ') if 'acc=' in s]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "res9 = \"\"\"\n",
    "9번째 모델\n",
    "[Epoch 1 Batch 50/482] loss=3.9611, lr=0.0000104167, acc=0.845\n",
    "[Epoch 1 Batch 100/482] loss=1.3811, lr=0.0000208333, acc=0.907\n",
    "[Epoch 1 Batch 150/482] loss=1.2201, lr=0.0000312500, acc=0.929\n",
    "[Epoch 1 Batch 200/482] loss=1.1854, lr=0.0000416667, acc=0.942\n",
    "[Epoch 1 Batch 250/482] loss=0.5175, lr=0.0000497692, acc=0.952\n",
    "[Epoch 1 Batch 300/482] loss=0.3711, lr=0.0000486150, acc=0.958\n",
    "[Epoch 1 Batch 350/482] loss=0.8173, lr=0.0000474608, acc=0.962\n",
    "[Epoch 1 Batch 400/482] loss=0.9540, lr=0.0000463066, acc=0.965\n",
    "[Epoch 1 Batch 450/482] loss=0.4038, lr=0.0000451524, acc=0.968\n",
    "Valid acc : 0.9945468709426123\n",
    "\n",
    "[Epoch 2 Batch 50/482] loss=0.7820, lr=0.0000432595, acc=0.989\n",
    "[Epoch 2 Batch 100/482] loss=0.3475, lr=0.0000421053, acc=0.992\n",
    "[Epoch 2 Batch 150/482] loss=0.6587, lr=0.0000409511, acc=0.991\n",
    "[Epoch 2 Batch 200/482] loss=0.7468, lr=0.0000397969, acc=0.990\n",
    "[Epoch 2 Batch 250/482] loss=0.2762, lr=0.0000386427, acc=0.992\n",
    "[Epoch 2 Batch 300/482] loss=0.2713, lr=0.0000374885, acc=0.992\n",
    "[Epoch 2 Batch 350/482] loss=0.2324, lr=0.0000363343, acc=0.993\n",
    "[Epoch 2 Batch 400/482] loss=1.0558, lr=0.0000351801, acc=0.992\n",
    "[Epoch 2 Batch 450/482] loss=0.2317, lr=0.0000340259, acc=0.992\n",
    "Valid acc : 0.9948065437548689\n",
    "\n",
    "[Epoch 3 Batch 50/482] loss=0.2652, lr=0.0000321330, acc=0.997\n",
    "[Epoch 3 Batch 100/482] loss=0.2646, lr=0.0000309788, acc=0.996\n",
    "[Epoch 3 Batch 150/482] loss=0.2653, lr=0.0000298246, acc=0.997\n",
    "[Epoch 3 Batch 200/482] loss=0.4518, lr=0.0000286704, acc=0.996\n",
    "[Epoch 3 Batch 250/482] loss=0.0875, lr=0.0000275162, acc=0.997\n",
    "[Epoch 3 Batch 300/482] loss=0.1259, lr=0.0000263620, acc=0.997\n",
    "[Epoch 3 Batch 350/482] loss=0.1758, lr=0.0000252078, acc=0.997\n",
    "[Epoch 3 Batch 400/482] loss=0.2577, lr=0.0000240536, acc=0.997\n",
    "[Epoch 3 Batch 450/482] loss=0.1267, lr=0.0000228994, acc=0.997\n",
    "Valid acc : 0.9971435990651779\n",
    "\n",
    "[Epoch 4 Batch 50/482] loss=0.0967, lr=0.0000210065, acc=0.999\n",
    "[Epoch 4 Batch 100/482] loss=0.1709, lr=0.0000198523, acc=0.998\n",
    "[Epoch 4 Batch 150/482] loss=0.3285, lr=0.0000186981, acc=0.998\n",
    "[Epoch 4 Batch 200/482] loss=0.3037, lr=0.0000175439, acc=0.997\n",
    "[Epoch 4 Batch 250/482] loss=0.1207, lr=0.0000163897, acc=0.997\n",
    "[Epoch 4 Batch 300/482] loss=0.1002, lr=0.0000152355, acc=0.998\n",
    "[Epoch 4 Batch 350/482] loss=0.0990, lr=0.0000140813, acc=0.998\n",
    "[Epoch 4 Batch 400/482] loss=0.1787, lr=0.0000129271, acc=0.998\n",
    "[Epoch 4 Batch 450/482] loss=0.1241, lr=0.0000117729, acc=0.998\n",
    "Valid acc : 0.997662944689691\n",
    "\n",
    "[Epoch 5 Batch 50/482] loss=0.1055, lr=0.0000098800, acc=0.999\n",
    "[Epoch 5 Batch 100/482] loss=0.1551, lr=0.0000087258, acc=0.998\n",
    "[Epoch 5 Batch 150/482] loss=0.2334, lr=0.0000075716, acc=0.998\n",
    "[Epoch 5 Batch 200/482] loss=0.2024, lr=0.0000064174, acc=0.998\n",
    "[Epoch 5 Batch 250/482] loss=0.0469, lr=0.0000052632, acc=0.998\n",
    "[Epoch 5 Batch 300/482] loss=0.1023, lr=0.0000041090, acc=0.998\n",
    "[Epoch 5 Batch 350/482] loss=0.0976, lr=0.0000029548, acc=0.998\n",
    "[Epoch 5 Batch 400/482] loss=0.2008, lr=0.0000018006, acc=0.998\n",
    "[Epoch 5 Batch 450/482] loss=0.1194, lr=0.0000006464, acc=0.998\n",
    "Valid acc : 0.9979226175019476\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss9 = [float(s.split('=')[1][:-1]) \n",
    "         for s in res9.split(' ') if 'loss' in s]\n",
    "acc9 = [float(s.split('=')[1][:5]) \n",
    "         for s in res9.split(' ') if 'acc=' in s]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "res10 = \"\"\"\n",
    "10번째 모델\n",
    "[Epoch 1 Batch 50/482] loss=4.2806, lr=0.0000104167, acc=0.839\n",
    "[Epoch 1 Batch 100/482] loss=1.1826, lr=0.0000208333, acc=0.909\n",
    "[Epoch 1 Batch 150/482] loss=1.0535, lr=0.0000312500, acc=0.932\n",
    "[Epoch 1 Batch 200/482] loss=0.9291, lr=0.0000416667, acc=0.945\n",
    "[Epoch 1 Batch 250/482] loss=0.7484, lr=0.0000497692, acc=0.953\n",
    "[Epoch 1 Batch 300/482] loss=2.0532, lr=0.0000486150, acc=0.955\n",
    "[Epoch 1 Batch 350/482] loss=3.6109, lr=0.0000474608, acc=0.948\n",
    "[Epoch 1 Batch 400/482] loss=5.6462, lr=0.0000463066, acc=0.928\n",
    "[Epoch 1 Batch 450/482] loss=2.7738, lr=0.0000451524, acc=0.929\n",
    "Valid acc : 0.9657231887821345\n",
    "\n",
    "[Epoch 2 Batch 50/482] loss=2.5548, lr=0.0000432595, acc=0.945\n",
    "[Epoch 2 Batch 100/482] loss=2.5951, lr=0.0000421053, acc=0.948\n",
    "[Epoch 2 Batch 150/482] loss=1.5108, lr=0.0000409511, acc=0.957\n",
    "[Epoch 2 Batch 200/482] loss=0.9801, lr=0.0000397969, acc=0.964\n",
    "[Epoch 2 Batch 250/482] loss=0.7584, lr=0.0000386427, acc=0.969\n",
    "[Epoch 2 Batch 300/482] loss=0.7742, lr=0.0000374885, acc=0.972\n",
    "[Epoch 2 Batch 350/482] loss=0.7253, lr=0.0000363343, acc=0.975\n",
    "[Epoch 2 Batch 400/482] loss=0.4543, lr=0.0000351801, acc=0.977\n",
    "[Epoch 2 Batch 450/482] loss=0.4451, lr=0.0000340259, acc=0.979\n",
    "Valid acc : 0.9933783432874578\n",
    "\n",
    "[Epoch 3 Batch 50/482] loss=0.4898, lr=0.0000321330, acc=0.992\n",
    "[Epoch 3 Batch 100/482] loss=0.5548, lr=0.0000309788, acc=0.992\n",
    "[Epoch 3 Batch 150/482] loss=0.3056, lr=0.0000298246, acc=0.993\n",
    "[Epoch 3 Batch 200/482] loss=0.2836, lr=0.0000286704, acc=0.994\n",
    "[Epoch 3 Batch 250/482] loss=0.3870, lr=0.0000275162, acc=0.994\n",
    "[Epoch 3 Batch 300/482] loss=0.4346, lr=0.0000263620, acc=0.994\n",
    "[Epoch 3 Batch 350/482] loss=0.3548, lr=0.0000252078, acc=0.994\n",
    "[Epoch 3 Batch 400/482] loss=0.3191, lr=0.0000240536, acc=0.994\n",
    "[Epoch 3 Batch 450/482] loss=0.2405, lr=0.0000228994, acc=0.995\n",
    "Valid acc : 0.9970137626590496\n",
    "\n",
    "[Epoch 4 Batch 50/482] loss=0.2599, lr=0.0000210065, acc=0.997\n",
    "[Epoch 4 Batch 100/482] loss=0.1820, lr=0.0000198523, acc=0.997\n",
    "[Epoch 4 Batch 150/482] loss=0.1320, lr=0.0000186981, acc=0.998\n",
    "[Epoch 4 Batch 200/482] loss=0.0875, lr=0.0000175439, acc=0.998\n",
    "[Epoch 4 Batch 250/482] loss=0.3268, lr=0.0000163897, acc=0.997\n",
    "[Epoch 4 Batch 300/482] loss=0.1387, lr=0.0000152355, acc=0.998\n",
    "[Epoch 4 Batch 350/482] loss=0.2819, lr=0.0000140813, acc=0.997\n",
    "[Epoch 4 Batch 400/482] loss=0.1954, lr=0.0000129271, acc=0.997\n",
    "[Epoch 4 Batch 450/482] loss=0.1686, lr=0.0000117729, acc=0.997\n",
    "Valid acc : 0.9971435990651779\n",
    "\n",
    "[Epoch 5 Batch 50/482] loss=0.1668, lr=0.0000098800, acc=0.998\n",
    "[Epoch 5 Batch 100/482] loss=0.1999, lr=0.0000087258, acc=0.997\n",
    "[Epoch 5 Batch 150/482] loss=0.1259, lr=0.0000075716, acc=0.998\n",
    "[Epoch 5 Batch 200/482] loss=0.0679, lr=0.0000064174, acc=0.998\n",
    "[Epoch 5 Batch 250/482] loss=0.2251, lr=0.0000052632, acc=0.998\n",
    "[Epoch 5 Batch 300/482] loss=0.1222, lr=0.0000041090, acc=0.998\n",
    "[Epoch 5 Batch 350/482] loss=0.2099, lr=0.0000029548, acc=0.998\n",
    "[Epoch 5 Batch 400/482] loss=0.1494, lr=0.0000018006, acc=0.998\n",
    "[Epoch 5 Batch 450/482] loss=0.1430, lr=0.0000006464, acc=0.998\n",
    "Valid acc : 0.9972734354713062\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss10 = [float(s.split('=')[1][:-1]) \n",
    "         for s in res10.split(' ') if 'loss' in s]\n",
    "acc10 = [float(s.split('=')[1][:5]) \n",
    "         for s in res10.split(' ') if 'acc=' in s]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "res11 = \"\"\"\n",
    "11번째 모델\n",
    "[Epoch 1 Batch 50/482] loss=3.4300, lr=0.0000104167, acc=0.868\n",
    "[Epoch 1 Batch 100/482] loss=0.9659, lr=0.0000208333, acc=0.922\n",
    "[Epoch 1 Batch 150/482] loss=0.9275, lr=0.0000312500, acc=0.942\n",
    "[Epoch 1 Batch 200/482] loss=1.0997, lr=0.0000416667, acc=0.952\n",
    "[Epoch 1 Batch 250/482] loss=0.6810, lr=0.0000497692, acc=0.959\n",
    "[Epoch 1 Batch 300/482] loss=3.4885, lr=0.0000486150, acc=0.937\n",
    "[Epoch 1 Batch 350/482] loss=8.9202, lr=0.0000474608, acc=0.875\n",
    "[Epoch 1 Batch 400/482] loss=5.2570, lr=0.0000463066, acc=0.865\n",
    "[Epoch 1 Batch 450/482] loss=9.5234, lr=0.0000451524, acc=0.826\n",
    "Valid acc : 0.4857179953258894\n",
    "\n",
    "[Epoch 2 Batch 50/482] loss=8.7889, lr=0.0000432595, acc=0.503\n",
    "[Epoch 2 Batch 100/482] loss=8.8263, lr=0.0000421053, acc=0.498\n",
    "[Epoch 2 Batch 150/482] loss=8.7484, lr=0.0000409511, acc=0.501\n",
    "[Epoch 2 Batch 200/482] loss=8.7575, lr=0.0000397969, acc=0.502\n",
    "[Epoch 2 Batch 250/482] loss=8.7555, lr=0.0000386427, acc=0.500\n",
    "[Epoch 2 Batch 300/482] loss=8.7711, lr=0.0000374885, acc=0.501\n",
    "[Epoch 2 Batch 350/482] loss=8.7152, lr=0.0000363343, acc=0.503\n",
    "[Epoch 2 Batch 400/482] loss=8.6789, lr=0.0000351801, acc=0.506\n",
    "[Epoch 2 Batch 450/482] loss=5.4142, lr=0.0000340259, acc=0.535\n",
    "Valid acc : 0.9261230849130097\n",
    "\n",
    "[Epoch 3 Batch 50/482] loss=5.0596, lr=0.0000321330, acc=0.909\n",
    "[Epoch 3 Batch 100/482] loss=1.9305, lr=0.0000309788, acc=0.935\n",
    "[Epoch 3 Batch 150/482] loss=2.7202, lr=0.0000298246, acc=0.934\n",
    "[Epoch 3 Batch 200/482] loss=2.2536, lr=0.0000286704, acc=0.937\n",
    "[Epoch 3 Batch 250/482] loss=2.9386, lr=0.0000275162, acc=0.940\n",
    "[Epoch 3 Batch 300/482] loss=1.2048, lr=0.0000263620, acc=0.947\n",
    "[Epoch 3 Batch 350/482] loss=0.8091, lr=0.0000252078, acc=0.952\n",
    "[Epoch 3 Batch 400/482] loss=0.4630, lr=0.0000240536, acc=0.957\n",
    "[Epoch 3 Batch 450/482] loss=0.3702, lr=0.0000228994, acc=0.961\n",
    "Valid acc : 0.990651778758764\n",
    "\n",
    "[Epoch 4 Batch 50/482] loss=0.2151, lr=0.0000210065, acc=0.995\n",
    "[Epoch 4 Batch 100/482] loss=0.3994, lr=0.0000198523, acc=0.993\n",
    "[Epoch 4 Batch 150/482] loss=0.5205, lr=0.0000186981, acc=0.992\n",
    "[Epoch 4 Batch 200/482] loss=0.4955, lr=0.0000175439, acc=0.992\n",
    "[Epoch 4 Batch 250/482] loss=0.2630, lr=0.0000163897, acc=0.992\n",
    "[Epoch 4 Batch 300/482] loss=0.2189, lr=0.0000152355, acc=0.993\n",
    "[Epoch 4 Batch 350/482] loss=0.2819, lr=0.0000140813, acc=0.993\n",
    "[Epoch 4 Batch 400/482] loss=0.1801, lr=0.0000129271, acc=0.994\n",
    "[Epoch 4 Batch 450/482] loss=0.2178, lr=0.0000117729, acc=0.994\n",
    "Valid acc : 0.9932485068813295\n",
    "\n",
    "[Epoch 5 Batch 50/482] loss=0.1623, lr=0.0000098800, acc=0.997\n",
    "[Epoch 5 Batch 100/482] loss=0.2463, lr=0.0000087258, acc=0.996\n",
    "[Epoch 5 Batch 150/482] loss=0.2804, lr=0.0000075716, acc=0.996\n",
    "[Epoch 5 Batch 200/482] loss=0.2318, lr=0.0000064174, acc=0.996\n",
    "[Epoch 5 Batch 250/482] loss=0.1175, lr=0.0000052632, acc=0.996\n",
    "[Epoch 5 Batch 300/482] loss=0.1283, lr=0.0000041090, acc=0.996\n",
    "[Epoch 5 Batch 350/482] loss=0.1323, lr=0.0000029548, acc=0.997\n",
    "[Epoch 5 Batch 400/482] loss=0.0871, lr=0.0000018006, acc=0.997\n",
    "[Epoch 5 Batch 450/482] loss=0.1688, lr=0.0000006464, acc=0.997\n",
    "Valid acc : 0.9963645806284082\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss11 = [float(s.split('=')[1][:-1]) \n",
    "         for s in res11.split(' ') if 'loss' in s]\n",
    "acc11 = [float(s.split('=')[1][:5]) \n",
    "         for s in res11.split(' ') if 'acc=' in s]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "res12 = \"\"\"\n",
    "12번째 모델\n",
    "[Epoch 1 Batch 50/482] loss=3.3269, lr=0.0000104167, acc=0.879\n",
    "[Epoch 1 Batch 100/482] loss=1.2529, lr=0.0000208333, acc=0.925\n",
    "[Epoch 1 Batch 150/482] loss=0.8157, lr=0.0000312500, acc=0.944\n",
    "[Epoch 1 Batch 200/482] loss=0.6348, lr=0.0000416667, acc=0.955\n",
    "[Epoch 1 Batch 250/482] loss=0.6600, lr=0.0000497692, acc=0.962\n",
    "[Epoch 1 Batch 300/482] loss=0.7342, lr=0.0000486150, acc=0.966\n",
    "[Epoch 1 Batch 350/482] loss=0.5806, lr=0.0000474608, acc=0.969\n",
    "[Epoch 1 Batch 400/482] loss=0.5801, lr=0.0000463066, acc=0.972\n",
    "[Epoch 1 Batch 450/482] loss=0.4648, lr=0.0000451524, acc=0.974\n",
    "Valid acc : 0.99623474422228\n",
    "\n",
    "[Epoch 2 Batch 50/482] loss=0.2891, lr=0.0000432595, acc=0.995\n",
    "[Epoch 2 Batch 100/482] loss=0.1096, lr=0.0000421053, acc=0.996\n",
    "[Epoch 2 Batch 150/482] loss=0.2260, lr=0.0000409511, acc=0.997\n",
    "[Epoch 2 Batch 200/482] loss=0.1259, lr=0.0000397969, acc=0.997\n",
    "[Epoch 2 Batch 250/482] loss=0.2979, lr=0.0000386427, acc=0.996\n",
    "[Epoch 2 Batch 300/482] loss=0.2110, lr=0.0000374885, acc=0.996\n",
    "[Epoch 2 Batch 350/482] loss=0.1227, lr=0.0000363343, acc=0.997\n",
    "[Epoch 2 Batch 400/482] loss=0.2190, lr=0.0000351801, acc=0.996\n",
    "[Epoch 2 Batch 450/482] loss=0.1416, lr=0.0000340259, acc=0.996\n",
    "Valid acc : 0.994417034536484\n",
    "\n",
    "[Epoch 3 Batch 50/482] loss=0.1709, lr=0.0000321330, acc=0.997\n",
    "[Epoch 3 Batch 100/482] loss=0.0552, lr=0.0000309788, acc=0.998\n",
    "[Epoch 3 Batch 150/482] loss=0.1037, lr=0.0000298246, acc=0.998\n",
    "[Epoch 3 Batch 200/482] loss=0.0629, lr=0.0000286704, acc=0.998\n",
    "[Epoch 3 Batch 250/482] loss=0.0916, lr=0.0000275162, acc=0.998\n",
    "[Epoch 3 Batch 300/482] loss=0.0954, lr=0.0000263620, acc=0.998\n",
    "[Epoch 3 Batch 350/482] loss=0.0536, lr=0.0000252078, acc=0.998\n",
    "[Epoch 3 Batch 400/482] loss=0.1229, lr=0.0000240536, acc=0.998\n",
    "[Epoch 3 Batch 450/482] loss=0.0182, lr=0.0000228994, acc=0.998\n",
    "Valid acc : 0.9984419631264606\n",
    "\n",
    "[Epoch 4 Batch 50/482] loss=0.0014, lr=0.0000210065, acc=1.000\n",
    "[Epoch 4 Batch 100/482] loss=0.0017, lr=0.0000198523, acc=1.000\n",
    "[Epoch 4 Batch 150/482] loss=0.0185, lr=0.0000186981, acc=1.000\n",
    "[Epoch 4 Batch 200/482] loss=0.0011, lr=0.0000175439, acc=1.000\n",
    "[Epoch 4 Batch 250/482] loss=0.0172, lr=0.0000163897, acc=1.000\n",
    "[Epoch 4 Batch 300/482] loss=0.0441, lr=0.0000152355, acc=1.000\n",
    "[Epoch 4 Batch 350/482] loss=0.0094, lr=0.0000140813, acc=1.000\n",
    "[Epoch 4 Batch 400/482] loss=0.0144, lr=0.0000129271, acc=1.000\n",
    "[Epoch 4 Batch 450/482] loss=0.0007, lr=0.0000117729, acc=1.000\n",
    "Valid acc : 0.9983121267203324\n",
    "\n",
    "[Epoch 5 Batch 50/482] loss=0.0142, lr=0.0000098800, acc=1.000\n",
    "[Epoch 5 Batch 100/482] loss=0.0159, lr=0.0000087258, acc=1.000\n",
    "[Epoch 5 Batch 150/482] loss=0.0005, lr=0.0000075716, acc=1.000\n",
    "[Epoch 5 Batch 200/482] loss=0.0007, lr=0.0000064174, acc=1.000\n",
    "[Epoch 5 Batch 250/482] loss=0.0005, lr=0.0000052632, acc=1.000\n",
    "[Epoch 5 Batch 300/482] loss=0.0490, lr=0.0000041090, acc=1.000\n",
    "[Epoch 5 Batch 350/482] loss=0.0012, lr=0.0000029548, acc=1.000\n",
    "[Epoch 5 Batch 400/482] loss=0.0007, lr=0.0000018006, acc=1.000\n",
    "[Epoch 5 Batch 450/482] loss=0.0005, lr=0.0000006464, acc=1.000\n",
    "Valid acc : 0.9984419631264606\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss12 = [float(s.split('=')[1][:-1]) \n",
    "         for s in res12.split(' ') if 'loss' in s]\n",
    "acc12 = [float(s.split('=')[1][:5]) \n",
    "         for s in res12.split(' ') if 'acc=' in s]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "res13 = \"\"\"\n",
    "13번째 모델\n",
    "[Epoch 1 Batch 50/482] loss=3.7528, lr=0.0000104167, acc=0.855\n",
    "[Epoch 1 Batch 100/482] loss=1.2487, lr=0.0000208333, acc=0.912\n",
    "[Epoch 1 Batch 150/482] loss=0.7341, lr=0.0000312500, acc=0.937\n",
    "[Epoch 1 Batch 200/482] loss=0.7848, lr=0.0000416667, acc=0.949\n",
    "[Epoch 1 Batch 250/482] loss=0.7147, lr=0.0000497692, acc=0.957\n",
    "[Epoch 1 Batch 300/482] loss=0.6266, lr=0.0000486150, acc=0.962\n",
    "[Epoch 1 Batch 350/482] loss=0.3855, lr=0.0000474608, acc=0.966\n",
    "[Epoch 1 Batch 400/482] loss=0.4965, lr=0.0000463066, acc=0.969\n",
    "[Epoch 1 Batch 450/482] loss=0.5793, lr=0.0000451524, acc=0.972\n",
    "Valid acc : 0.9922098156323033\n",
    "\n",
    "[Epoch 2 Batch 50/482] loss=0.2144, lr=0.0000432595, acc=0.995\n",
    "[Epoch 2 Batch 100/482] loss=0.3609, lr=0.0000421053, acc=0.994\n",
    "[Epoch 2 Batch 150/482] loss=0.3072, lr=0.0000409511, acc=0.994\n",
    "[Epoch 2 Batch 200/482] loss=0.2063, lr=0.0000397969, acc=0.995\n",
    "[Epoch 2 Batch 250/482] loss=0.1182, lr=0.0000386427, acc=0.996\n",
    "[Epoch 2 Batch 300/482] loss=0.1226, lr=0.0000374885, acc=0.996\n",
    "[Epoch 2 Batch 350/482] loss=0.1257, lr=0.0000363343, acc=0.996\n",
    "[Epoch 2 Batch 400/482] loss=0.1506, lr=0.0000351801, acc=0.996\n",
    "[Epoch 2 Batch 450/482] loss=0.2309, lr=0.0000340259, acc=0.996\n",
    "Valid acc : 0.9971435990651779\n",
    "\n",
    "[Epoch 3 Batch 50/482] loss=0.1950, lr=0.0000321330, acc=0.996\n",
    "[Epoch 3 Batch 100/482] loss=0.0484, lr=0.0000309788, acc=0.997\n",
    "[Epoch 3 Batch 150/482] loss=0.0823, lr=0.0000298246, acc=0.998\n",
    "[Epoch 3 Batch 200/482] loss=0.1077, lr=0.0000286704, acc=0.998\n",
    "[Epoch 3 Batch 250/482] loss=0.1248, lr=0.0000275162, acc=0.998\n",
    "[Epoch 3 Batch 300/482] loss=0.0754, lr=0.0000263620, acc=0.998\n",
    "[Epoch 3 Batch 350/482] loss=0.0329, lr=0.0000252078, acc=0.998\n",
    "[Epoch 3 Batch 400/482] loss=0.0782, lr=0.0000240536, acc=0.998\n",
    "[Epoch 3 Batch 450/482] loss=0.0865, lr=0.0000228994, acc=0.998\n",
    "Valid acc : 0.9975331082835627\n",
    "\n",
    "[Epoch 4 Batch 50/482] loss=0.0193, lr=0.0000210065, acc=0.999\n",
    "[Epoch 4 Batch 100/482] loss=0.0501, lr=0.0000198523, acc=0.999\n",
    "[Epoch 4 Batch 150/482] loss=0.0281, lr=0.0000186981, acc=0.999\n",
    "[Epoch 4 Batch 200/482] loss=0.0741, lr=0.0000175439, acc=0.999\n",
    "[Epoch 4 Batch 250/482] loss=0.0296, lr=0.0000163897, acc=0.999\n",
    "[Epoch 4 Batch 300/482] loss=0.0499, lr=0.0000152355, acc=0.999\n",
    "[Epoch 4 Batch 350/482] loss=0.0264, lr=0.0000140813, acc=0.999\n",
    "[Epoch 4 Batch 400/482] loss=0.0351, lr=0.0000129271, acc=0.999\n",
    "[Epoch 4 Batch 450/482] loss=0.0815, lr=0.0000117729, acc=0.999\n",
    "Valid acc : 0.9983121267203324\n",
    "\n",
    "[Epoch 5 Batch 50/482] loss=0.0028, lr=0.0000098800, acc=1.000\n",
    "[Epoch 5 Batch 100/482] loss=0.0206, lr=0.0000087258, acc=1.000\n",
    "[Epoch 5 Batch 150/482] loss=0.0156, lr=0.0000075716, acc=1.000\n",
    "[Epoch 5 Batch 200/482] loss=0.0614, lr=0.0000064174, acc=1.000\n",
    "[Epoch 5 Batch 250/482] loss=0.0019, lr=0.0000052632, acc=1.000\n",
    "[Epoch 5 Batch 300/482] loss=0.0112, lr=0.0000041090, acc=1.000\n",
    "[Epoch 5 Batch 350/482] loss=0.0364, lr=0.0000029548, acc=1.000\n",
    "[Epoch 5 Batch 400/482] loss=0.0029, lr=0.0000018006, acc=1.000\n",
    "[Epoch 5 Batch 450/482] loss=0.0411, lr=0.0000006464, acc=1.000\n",
    "Valid acc : 0.9985717995325889\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss13 = [float(s.split('=')[1][:-1]) \n",
    "         for s in res13.split(' ') if 'loss' in s]\n",
    "acc13 = [float(s.split('=')[1][:5]) \n",
    "         for s in res13.split(' ') if 'acc=' in s]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "res14 = \"\"\"\n",
    "14번째 모델\n",
    "[Epoch 1 Batch 50/482] loss=3.2999, lr=0.0000104167, acc=0.874\n",
    "[Epoch 1 Batch 100/482] loss=1.0743, lr=0.0000208333, acc=0.924\n",
    "[Epoch 1 Batch 150/482] loss=1.0664, lr=0.0000312500, acc=0.943\n",
    "[Epoch 1 Batch 200/482] loss=0.9332, lr=0.0000416667, acc=0.953\n",
    "[Epoch 1 Batch 250/482] loss=0.7026, lr=0.0000497692, acc=0.960\n",
    "[Epoch 1 Batch 300/482] loss=1.4410, lr=0.0000486150, acc=0.963\n",
    "[Epoch 1 Batch 350/482] loss=1.4836, lr=0.0000474608, acc=0.964\n",
    "[Epoch 1 Batch 400/482] loss=0.4177, lr=0.0000463066, acc=0.968\n",
    "[Epoch 1 Batch 450/482] loss=0.3077, lr=0.0000451524, acc=0.971\n",
    "Valid acc : 0.9933783432874578\n",
    "\n",
    "[Epoch 2 Batch 50/482] loss=0.6474, lr=0.0000432595, acc=0.992\n",
    "[Epoch 2 Batch 100/482] loss=0.4733, lr=0.0000421053, acc=0.992\n",
    "[Epoch 2 Batch 150/482] loss=0.3642, lr=0.0000409511, acc=0.993\n",
    "[Epoch 2 Batch 200/482] loss=0.2900, lr=0.0000397969, acc=0.994\n",
    "[Epoch 2 Batch 250/482] loss=0.1701, lr=0.0000386427, acc=0.994\n",
    "[Epoch 2 Batch 300/482] loss=0.3370, lr=0.0000374885, acc=0.994\n",
    "[Epoch 2 Batch 350/482] loss=0.3044, lr=0.0000363343, acc=0.995\n",
    "[Epoch 2 Batch 400/482] loss=0.1856, lr=0.0000351801, acc=0.995\n",
    "[Epoch 2 Batch 450/482] loss=0.2458, lr=0.0000340259, acc=0.995\n",
    "Valid acc : 0.9972734354713062\n",
    "\n",
    "[Epoch 3 Batch 50/482] loss=0.2248, lr=0.0000321330, acc=0.996\n",
    "[Epoch 3 Batch 100/482] loss=0.3010, lr=0.0000309788, acc=0.995\n",
    "[Epoch 3 Batch 150/482] loss=0.0719, lr=0.0000298246, acc=0.996\n",
    "[Epoch 3 Batch 200/482] loss=0.0435, lr=0.0000286704, acc=0.997\n",
    "[Epoch 3 Batch 250/482] loss=0.0645, lr=0.0000275162, acc=0.997\n",
    "[Epoch 3 Batch 300/482] loss=0.1775, lr=0.0000263620, acc=0.997\n",
    "[Epoch 3 Batch 350/482] loss=0.1415, lr=0.0000252078, acc=0.997\n",
    "[Epoch 3 Batch 400/482] loss=0.1107, lr=0.0000240536, acc=0.997\n",
    "[Epoch 3 Batch 450/482] loss=0.0650, lr=0.0000228994, acc=0.997\n",
    "Valid acc : 0.9971435990651779\n",
    "\n",
    "[Epoch 4 Batch 50/482] loss=0.0750, lr=0.0000210065, acc=0.999\n",
    "[Epoch 4 Batch 100/482] loss=0.0793, lr=0.0000198523, acc=0.999\n",
    "[Epoch 4 Batch 150/482] loss=0.0664, lr=0.0000186981, acc=0.999\n",
    "[Epoch 4 Batch 200/482] loss=0.0206, lr=0.0000175439, acc=0.999\n",
    "[Epoch 4 Batch 250/482] loss=0.0783, lr=0.0000163897, acc=0.999\n",
    "[Epoch 4 Batch 300/482] loss=0.0875, lr=0.0000152355, acc=0.999\n",
    "[Epoch 4 Batch 350/482] loss=0.0508, lr=0.0000140813, acc=0.999\n",
    "[Epoch 4 Batch 400/482] loss=0.0802, lr=0.0000129271, acc=0.999\n",
    "[Epoch 4 Batch 450/482] loss=0.0135, lr=0.0000117729, acc=0.999\n",
    "Valid acc : 0.996754089846793\n",
    "\n",
    "[Epoch 5 Batch 50/482] loss=0.0205, lr=0.0000098800, acc=0.999\n",
    "[Epoch 5 Batch 100/482] loss=0.0029, lr=0.0000087258, acc=1.000\n",
    "[Epoch 5 Batch 150/482] loss=0.0465, lr=0.0000075716, acc=1.000\n",
    "[Epoch 5 Batch 200/482] loss=0.0028, lr=0.0000064174, acc=1.000\n",
    "[Epoch 5 Batch 250/482] loss=0.0121, lr=0.0000052632, acc=1.000\n",
    "[Epoch 5 Batch 300/482] loss=0.0013, lr=0.0000041090, acc=1.000\n",
    "[Epoch 5 Batch 350/482] loss=0.0558, lr=0.0000029548, acc=1.000\n",
    "[Epoch 5 Batch 400/482] loss=0.0201, lr=0.0000018006, acc=1.000\n",
    "[Epoch 5 Batch 450/482] loss=0.0018, lr=0.0000006464, acc=1.000\n",
    "Valid acc : 0.9988314723448455\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss14 = [float(s.split('=')[1][:-1]) \n",
    "         for s in res14.split(' ') if 'loss' in s]\n",
    "acc14 = [float(s.split('=')[1][:5]) \n",
    "         for s in res14.split(' ') if 'acc=' in s]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib as mpl\n",
    "mpl.style.use('fivethirtyeight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sys\n",
    "mod = sys.modules[__name__]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10, 6), facecolor='w')\n",
    "for i in range(6):\n",
    "    plt.axvline(9*i-.5, ls=':', color='r', alpha=.7)\n",
    "\n",
    "for i in range(1, 15):\n",
    "    if i <= 6: ls = None\n",
    "    elif i <= 12: ls = '-.'\n",
    "    else: ls = ':'\n",
    "    plt.plot(getattr(mod, 'loss{}'.format(i)),\n",
    "             marker='o', label='loss{}'.format(i), ls=ls)\n",
    "\n",
    "plt.legend(loc='upper right', fontsize=14)\n",
    "plt.title('Loss (BERTClassifier 1~14)', fontsize=18)\n",
    "plt.xlim([-1, 45])\n",
    "# plt.ylim([-.01, 5.01])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10, 6), facecolor='w')\n",
    "for i in range(6):\n",
    "    plt.axvline(9*i-.5, ls=':', color='r', alpha=.7)\n",
    "\n",
    "for i in range(1, 15):\n",
    "    if i <= 6: ls = None\n",
    "    elif i <= 12: ls = '-.'\n",
    "    else: ls = ':'\n",
    "    plt.plot(getattr(mod, 'acc{}'.format(i)),\n",
    "             marker='s', label='acc{}'.format(i), ls=ls)\n",
    "\n",
    "plt.legend(loc='lower right', fontsize=14)\n",
    "plt.title('Accuracy', fontsize=18)\n",
    "plt.xlim([-1, 45])\n",
    "# plt.ylim([0.84, 1.01])\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "basic",
   "language": "python",
   "name": "basic"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}