C_data_synthesis.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Adversarial Targeted Dataset Simulation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Paths\n",
    "LAINET_PATH = 'lainet_trained.pth'\n",
    "NEURALADM_PATH = 'adv.pt'\n",
    "device = 'cuda'\n",
    "founders_vcf_file_path = 'founders_train.vcf'\n",
    "founders_map_file_path = 'founders_train.map'\n",
    "output_folder_prefix = 'output/'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Importing libraries, datasets, and networks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Start importing required libraries...\n",
      "Done importing, using pytorch version 1.12.1+cu102...\n"
     ]
    }
   ],
   "source": [
    "# Importing libraries\n",
    "print('Start importing required libraries...')\n",
    "import os, sys, time, allel, yaml, math, gzip, torch\n",
    "sys.path.append('../LAI-Net/')\n",
    "from tqdm.auto import tqdm\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "import torch.optim as optim\n",
    "from collections import Counter\n",
    "from scipy.interpolate import interp1d\n",
    "from matplotlib import cm\n",
    "from matplotlib.patches import Patch\n",
    "from matplotlib.lines import Line2D\n",
    "import pickle as pkl\n",
    "import numpy as np\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.colors import ListedColormap, LinearSegmentedColormap\n",
    "from sklearn.decomposition import PCA\n",
    "import sklearn\n",
    "from sklearn import neighbors, datasets\n",
    "\n",
    "from lainet.utils.reader import load_founders_from_vcf_and_map\n",
    "\n",
    "from helpers.method_inference import forward_haploid, forward_nadm, extend_if_lai, compute_pca, unique_to_percentage\n",
    "from helpers.binarize import binarize\n",
    "from helpers.adversarial_perturbation import cw_loss, subsample_adversarial_mask, eval_net_and_subsample, generate_adversarial_mask, run_robustness_analysis\n",
    "from helpers.adversarial_synthesis import generate_adversarial_population\n",
    "\n",
    "print('Done importing, using pytorch version {}...'.format(torch.__version__))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading vcf and .map files...\n",
      "Done loading vcf and .map files...\n",
      "A total of 3200 diploid individuals where found in the vcf and .map\n",
      "A total of 8 ancestries where found: ['AFR' 'AHG' 'AMR' 'EAS' 'EUR' 'OCE' 'SAS' 'WAS']\n",
      "(2880, 516800) (2880,) (2880, 516800) (2880,)\n"
     ]
    }
   ],
   "source": [
    "# Loading dataset\n",
    "snps_all, labels_all, _, _, ancestry_names, _ = load_founders_from_vcf_and_map(founders_vcf_file_path, founders_map_file_path, make_haploid=True, random_split=False, verbose=True)\n",
    "\n",
    "subset = 2\n",
    "snps, labels = snps_all[0::subset,:], labels_all[0::subset]\n",
    "snps_test, labels_test = snps_all[1::subset,:], labels_all[1::subset]\n",
    "print(snps.shape, labels.shape, snps_test.shape, labels_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading LAI-Net from /local-scratch/mrivas/dmasmont/misc/lnpfoo/lainet_trained.pth ...\n",
      "Done loading LAI-Net\n"
     ]
    }
   ],
   "source": [
    "# Loading LAI-Net\n",
    "import lainet\n",
    "print(f'Loading LAI-Net from {LAINET_PATH} ...')\n",
    "LN = torch.load(LAINET_PATH, map_location=device)\n",
    "LN.device = device\n",
    "LN = LN.to(device)\n",
    "LN = LN.eval()\n",
    "print('Done loading LAI-Net')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading Neural ADMIXTURE from /local-scratch/mrivas/dmasmont/misc/lnpfoo/adv.pt ...\n",
      "Done loading Neural ADMIXTURE\n"
     ]
    }
   ],
   "source": [
    "# Loading Neural ADMIXTURE\n",
    "from neural_admixture.model.neural_admixture import NeuralAdmixture\n",
    "print(f'Loading Neural ADMIXTURE from {NEURALADM_PATH} ...')\n",
    "NADM= NeuralAdmixture([8], num_features=516800)\n",
    "NADM.load_state_dict(torch.load(NEURALADM_PATH, map_location=device), strict=True)\n",
    "NADM.device = device\n",
    "NADM = NADM.to(device)\n",
    "NADM = NADM.eval()\n",
    "print('Done loading Neural ADMIXTURE')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Setup PCA\n",
    "pca = PCA(n_components=2)\n",
    "X_pca = pca.fit_transform(snps)\n",
    "W = torch.tensor(pca.components_)\n",
    "b =  torch.tensor(pca.mean_)\n",
    "\n",
    "pcat, _, _, _ = compute_pca(torch.tensor(snps), W, b)\n",
    "pcat = pcat.numpy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(n_neighbors=1)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Setup Fast K-NN\n",
    "knn_dim = 1024\n",
    "knn_W = torch.randn((snps.shape[1], knn_dim)) * (1/(2*torch.sqrt(torch.tensor(knn_dim))))\n",
    "KNN = neighbors.KNeighborsClassifier(n_neighbors=1)\n",
    "train_template = np.matmul(snps, knn_W.cpu().numpy())\n",
    "KNN.fit(train_template, labels)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Generate and store fake dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 422.16282166997433 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 302.87279166738034 2.062979221343994 43.42768859863281 30.371673583984375\n",
      "1 415.6721182443901 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 297.23010569594345 2.050555944442749 43.24954605102539 29.892364501953125\n",
      "2 409.12463918572126 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 291.53596845513044 2.0385913848876953 43.0701789855957 29.40972137451172\n",
      "3 402.74690254400923 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 285.9865129775877 2.0260393619537354 42.894527435302734 28.945295333862305\n",
      "4 396.4503164327722 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 280.49679923420956 2.0140297412872314 42.724708557128906 28.490070343017578\n",
      "5 390.22761353919293 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 275.0923665470817 2.001671552658081 42.55233383178711 28.028907775878906\n",
      "6 384.002775236675 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 269.700035497257 1.989770531654358 42.37834548950195 27.556278228759766\n",
      "7 377.9821862166031 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 264.4923514311417 1.9784765243530273 42.20244216918945 27.106473922729492\n",
      "8 371.91052802132 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 259.2480099968754 1.965832233428955 42.03496551513672 26.626754760742188\n",
      "9 365.8815301536602 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 254.02895985832635 1.9532058238983154 41.86223602294922 26.17489242553711\n",
      "10 360.0044725651726 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 248.93543310117573 1.9404727220535278 41.6977653503418 25.733036041259766\n",
      "11 354.2691663731898 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 243.9901489247645 1.9298734664916992 41.528621673583984 25.291900634765625\n",
      "12 348.5376167892862 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 239.0398401617456 1.9176908731460571 41.3590202331543 24.862045288085938\n",
      "13 342.7670423113538 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 234.03156844012273 1.9063360691070557 41.1894645690918 24.45020866394043\n",
      "14 337.03122992970987 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 229.07523150899453 1.895705223083496 41.01670837402344 24.02687644958496\n",
      "15 331.45028128470165 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 224.27632298315748 1.884310245513916 40.84168243408203 23.60628318786621\n",
      "16 326.1342401658625 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 219.7366788541407 1.8737118244171143 40.674259185791016 23.175331115722656\n",
      "17 320.80732969512485 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 215.1715249227478 1.8634430170059204 40.501808166503906 22.76874542236328\n",
      "18 315.4370865177014 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 210.523109490762 1.852291464805603 40.336814880371094 22.3880558013916\n",
      "19 310.25012767801104 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 206.08086765298665 1.8396873474121094 40.17014694213867 21.98927879333496\n",
      "20 305.08772883960523 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 201.6501518304042 1.8275420665740967 39.999778747558594 21.610477447509766\n",
      "21 299.89310592526783 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 197.20288235062944 1.8173197507858276 39.82771682739258 21.217470169067383\n",
      "22 294.8596733918502 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 192.90266896394036 1.8078383207321167 39.65376281738281 20.84164047241211\n",
      "23 289.84340453472964 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 188.59960031834473 1.7982685565948486 39.4856071472168 20.474321365356445\n",
      "24 284.9314420502077 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 184.41806843685018 1.786097764968872 39.3170166015625 20.093242645263672\n",
      "25 280.19398536029894 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 180.40056528392873 1.7755215167999268 39.152427673339844 19.713043212890625\n",
      "26 275.532779912265 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 176.43464754036253 1.7654144763946533 38.988685607910156 19.3553466796875\n",
      "27 270.82529847527064 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 172.42858199978386 1.7548021078109741 38.82132339477539 18.999267578125\n",
      "28 266.18120590688693 tensor(5, device='cuda:4') tensor(0, device='cuda:4') 168.4856169605445 1.7439268827438354 38.65850067138672 18.634660720825195\n",
      "29 261.7236781952047 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 164.71330555650226 1.732466220855713 38.49428176879883 18.289342880249023\n",
      "30 257.20347658779326 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 160.86864010479158 1.723304271697998 38.33256912231445 17.946393966674805\n",
      "31 252.98340495977072 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 157.3217533484426 1.713897705078125 38.16840362548828 17.610946655273438\n",
      "32 248.7636815063953 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 153.75662479329105 1.7041277885437012 38.00384521484375 17.295238494873047\n",
      "33 244.46752312381977 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 150.1341113110375 1.6928917169570923 37.83348083496094 16.97355842590332\n",
      "34 240.3625311881609 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 146.70703363719653 1.682448387145996 37.66632843017578 16.640392303466797\n",
      "35 236.39749427641652 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 143.376530488384 1.6704312562942505 37.50492477416992 16.340682983398438\n",
      "36 232.30461597546866 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 139.94450891122153 1.6579853296279907 37.3411750793457 16.019771575927734\n",
      "37 228.32940783247008 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 136.61196675046935 1.6481213569641113 37.17900085449219 15.711318016052246\n",
      "38 224.5259337657106 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 133.44430115157485 1.6383552551269531 37.01454544067383 15.414186477661133\n",
      "39 220.75270918659487 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 130.30404976658144 1.628828525543213 36.854644775390625 15.110541343688965\n",
      "40 217.00544560980276 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 127.18274284433798 1.617449164390564 36.693939208984375 14.817375183105469\n",
      "41 213.32409802944437 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 124.11480919391887 1.6072800159454346 36.533260345458984 14.53548812866211\n",
      "42 209.77477075313757 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 121.18730451320837 1.59765625 36.37226486206055 14.245280265808105\n",
      "43 206.2720630257648 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 118.28502429418978 1.5867387056350708 36.2094841003418 13.981331825256348\n",
      "44 202.74431241595425 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 115.37143791758695 1.5785353183746338 36.04530334472656 13.70373249053955\n",
      "45 199.3436646453711 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 112.56854379099335 1.5691224336624146 35.887054443359375 13.431889533996582\n",
      "46 195.98867733569784 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 109.79125173183125 1.5589005947113037 35.7342414855957 13.170042037963867\n",
      "47 192.65956304228268 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 107.0553833023306 1.5505822896957397 35.572418212890625 12.908761024475098\n",
      "48 189.37267391758812 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 104.35320966320884 1.5418651103973389 35.410640716552734 12.656317710876465\n",
      "49 186.2102903485607 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 101.75148836377373 1.5329978466033936 35.2613410949707 12.403121948242188\n",
      "50 183.13756120817618 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 99.24761093275502 1.5238738059997559 35.106937408447266 12.152201652526855\n",
      "51 180.12240769070965 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 96.8041857640205 1.5148371458053589 34.948246002197266 11.906892776489258\n",
      "52 177.2265828926519 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 94.4491499741033 1.5058293342590332 34.79570770263672 11.680188179016113\n",
      "53 174.2511576153061 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 92.04842481386872 1.4949977397918701 34.63267517089844 11.442384719848633\n",
      "54 171.2877057194414 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 89.65121108290575 1.4852075576782227 34.47343063354492 11.204425811767578\n",
      "55 168.4516423184021 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 87.36673466745452 1.4759494066238403 34.316864013671875 10.97523021697998\n",
      "56 165.7564421471372 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 85.20286090934422 1.4677391052246094 34.16556167602539 10.754718780517578\n",
      "57 163.03254218222048 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 83.00914569021607 1.458573341369629 34.01374816894531 10.53732681274414\n",
      "58 160.3180725232957 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 80.82985910338148 1.4494465589523315 33.85823059082031 10.322305679321289\n",
      "59 157.6752974836449 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 78.7234775869469 1.441016674041748 33.699668884277344 10.111465454101562\n",
      "60 155.10200873857468 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 76.67427162176104 1.431918978691101 33.54734802246094 9.901122093200684\n",
      "61 152.5635168638562 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 74.65679348376737 1.4232696294784546 33.39155197143555 9.700349807739258\n",
      "62 150.06599297384548 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 72.68030550340939 1.4153400659561157 33.235816955566406 9.498713493347168\n",
      "63 147.59453936331363 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 70.71905453913301 1.4061883687973022 33.082008361816406 9.305279731750488\n",
      "64 145.2516136526744 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 68.87174586054489 1.3983433246612549 32.933353424072266 9.11481761932373\n",
      "65 142.98919125604962 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 67.10579451132153 1.3901478052139282 32.7817268371582 8.929795265197754\n",
      "66 140.73434697124523 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 65.35675619575542 1.382576584815979 32.63169860839844 8.731616973876953\n",
      "67 138.55303243210716 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 63.66733089497488 1.3734554052352905 32.48478698730469 8.542672157287598\n",
      "68 136.4135049917232 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 62.00674906749834 1.3646941184997559 32.337852478027344 8.36635684967041\n",
      "69 134.28052790808061 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 60.358514625304736 1.356755256652832 32.18964385986328 8.185970306396484\n",
      "70 132.1644240599657 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 58.7283434849764 1.3480802774429321 32.03795623779297 8.01208782196045\n",
      "71 130.140225640479 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 57.17555688971608 1.3394432067871094 31.891874313354492 7.841476917266846\n",
      "72 128.18684638783716 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 55.6750677947738 1.331777811050415 31.749826431274414 7.680347919464111\n",
      "73 126.2486130743525 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 54.20281977944588 1.3233630657196045 31.601089477539062 7.520251274108887\n",
      "74 124.31991138866437 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 52.72867575099957 1.3156838417053223 31.455944061279297 7.363663673400879\n",
      "75 122.48429932241199 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 51.33059742097613 1.3081990480422974 31.316913604736328 7.211675643920898\n",
      "76 120.60127455662787 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 49.90688366364539 1.300282597541809 31.167503356933594 7.059101581573486\n",
      "77 118.840391353527 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 48.58007438556282 1.2939082384109497 31.028589248657227 6.9092302322387695\n",
      "78 117.0785047412482 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 47.25219386821658 1.2844793796539307 30.886655807495117 6.768519878387451\n",
      "79 115.35265000749068 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 45.97057255197005 1.2775142192840576 30.74307632446289 6.618410587310791\n",
      "80 113.65134818549139 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 44.70986623759252 1.2704538106918335 30.60187530517578 6.467277526855469\n",
      "81 112.05079580213108 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 43.531588801398655 1.2621264457702637 30.464378356933594 6.328323841094971\n",
      "82 110.4061577481724 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 42.325320856044414 1.2554985284805298 30.31747817993164 6.19038200378418\n",
      "83 108.85159043308326 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 41.19284287925788 1.2485641241073608 30.17237091064453 6.065441608428955\n",
      "84 107.31720064499896 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 40.07465766289752 1.2414414882659912 30.032503128051758 5.936095237731934\n",
      "85 105.79829871872687 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 38.98116445759559 1.2339752912521362 29.888290405273438 5.806578159332275\n",
      "86 104.35432560325293 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 37.94507820487646 1.2261929512023926 29.750978469848633 5.681097507476807\n",
      "87 102.91065231167838 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 36.90373411976859 1.2191298007965088 29.615774154663086 5.556240081787109\n",
      "88 101.46237498052547 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 35.861423308925126 1.2134299278259277 29.475013732910156 5.437494277954102\n",
      "89 100.07372247059752 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 34.880984295205366 1.207391381263733 29.332351684570312 5.320643424987793\n",
      "90 98.75932544050914 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 33.95513599692088 1.2001025676727295 29.19507598876953 5.213934898376465\n",
      "91 97.46182646043869 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 33.04403068311782 1.1926718950271606 29.059646606445312 5.105830669403076\n",
      "92 96.21495313226674 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 32.181579776401264 1.1871380805969238 28.923812866210938 4.99860954284668\n",
      "93 94.90814968017146 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 31.27347727683589 1.1789448261260986 28.78435516357422 4.887017250061035\n",
      "94 93.65651271351362 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 30.416117763600596 1.1720861196517944 28.644514083862305 4.779280662536621\n",
      "95 92.41902131132184 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 29.56725615552961 1.1654534339904785 28.508678436279297 4.668954849243164\n",
      "96 91.26410371382323 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 28.785000625444322 1.1594476699829102 28.376489639282227 4.566676139831543\n",
      "97 90.13060370954207 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 28.025832572304523 1.1523733139038086 28.243026733398438 4.466344356536865\n",
      "98 89.028281131409 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 27.30157629551728 1.1463172435760498 28.10637664794922 4.367634296417236\n",
      "99 87.93365590491194 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 26.575976657055804 1.139017939567566 27.975074768066406 4.268511772155762\n",
      "100 86.85850701464626 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 25.882782675161092 1.1336208581924438 27.837106704711914 4.1678900718688965\n",
      "101 85.80767013442404 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 25.20399512660391 1.1266781091690063 27.701801300048828 4.073394298553467\n",
      "102 84.79399543309651 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 24.548937015776254 1.1211055517196655 27.569175720214844 3.9856014251708984\n",
      "103 83.81655497152877 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 23.926708765817402 1.1153002977371216 27.438539505004883 3.8974668979644775\n",
      "104 82.83888573452177 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 23.29080719753446 1.107713222503662 27.31256675720215 3.8152318000793457\n",
      "105 81.89501261441944 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 22.692192551782107 1.1014962196350098 27.18280029296875 3.7357232570648193\n",
      "106 80.96246834183074 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 22.114247039798265 1.0947201251983643 27.050209045410156 3.653083086013794\n",
      "107 80.03980602534716 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 21.5517250469536 1.0899171829223633 26.914390563964844 3.569382667541504\n",
      "108 79.15665448611968 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 21.011625418126513 1.0838730335235596 26.783065795898438 3.4950244426727295\n",
      "109 78.25732203212699 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 20.45760782447776 1.0769826173782349 26.65119171142578 3.4203481674194336\n",
      "110 77.40312237397427 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 19.952052899753994 1.070501446723938 26.519575119018555 3.3414177894592285\n",
      "111 76.57919856396438 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 19.457254017264855 1.0639716386795044 26.395620346069336 3.2667322158813477\n",
      "112 75.7594053624869 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 18.966467010855737 1.0570257902145386 26.269182205200195 3.1975481510162354\n",
      "113 74.96657458408288 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 18.511307512358933 1.0498894453048706 26.139331817626953 3.126713991165161\n",
      "114 74.1922639572561 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 18.063040372127322 1.044400691986084 26.012319564819336 3.0601837635040283\n",
      "115 73.42580869915966 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 17.62780514481548 1.0386897325515747 25.88397789001465 2.9913580417633057\n",
      "116 72.69402561402309 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 17.223931171655533 1.033137559890747 25.751962661743164 2.9330315589904785\n",
      "117 71.96382698528892 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 16.81219837181693 1.0272079706192017 25.63025665283203 2.8639073371887207\n",
      "118 71.23954769709192 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 16.41261228659235 1.0209308862686157 25.503820419311523 2.79836368560791\n",
      "119 70.56755764688278 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 16.051742984903104 1.0150398015975952 25.382265090942383 2.7362446784973145\n",
      "120 69.85252035085483 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 15.660610560815766 1.0096275806427002 25.25383949279785 2.674603223800659\n",
      "121 69.18610776093749 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 15.316483503849788 1.0036336183547974 25.12572479248047 2.6145410537719727\n",
      "122 68.52448821418025 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 14.982013884708667 0.9965284466743469 24.994583129882812 2.556779623031616\n",
      "123 67.89404369142524 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 14.655747540264047 0.990082859992981 24.8713436126709 2.505526065826416\n",
      "124 67.26229703925605 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 14.33127898477073 0.9850384593009949 24.74628257751465 2.4534144401550293\n",
      "125 66.6641319789094 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 14.031284550587555 0.9794584512710571 24.626585006713867 2.400218963623047\n",
      "126 66.03235194040923 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 13.728755980268996 0.9734339118003845 24.491724014282227 2.3467140197753906\n",
      "127 65.4236607795148 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 13.433020079651367 0.9673560261726379 24.363780975341797 2.295722723007202\n",
      "128 64.85513160404867 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 13.165811457011163 0.9619849324226379 24.240978240966797 2.2453787326812744\n",
      "129 64.27149339474602 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 12.899609823598094 0.9566468596458435 24.111976623535156 2.1912834644317627\n",
      "130 63.71393549098691 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 12.642444487950414 0.9517989158630371 23.988588333129883 2.1425154209136963\n",
      "131 63.17990363847324 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 12.404387009254187 0.9463986158370972 23.864952087402344 2.0992138385772705\n",
      "132 62.65318298149323 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 12.178421316625826 0.9417144656181335 23.739580154418945 2.053886890411377\n",
      "133 62.15549218280522 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 11.970636006428876 0.9365997314453125 23.61730194091797 2.0136525630950928\n",
      "134 61.65593001360782 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 11.758361906956516 0.9311891794204712 23.49807357788086 1.9702317714691162\n",
      "135 61.160428287944555 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 11.551337006053686 0.9257736206054688 23.377408981323242 1.928499698638916\n",
      "136 60.66687287457454 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 11.34831304915416 0.9205145239830017 23.254859924316406 1.8883254528045654\n",
      "137 60.18283469013441 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 11.1528763227867 0.9152634143829346 23.132448196411133 1.849798560142517\n",
      "138 59.714917866245976 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 10.972445158020726 0.9098462462425232 23.00936508178711 1.8138962984085083\n",
      "139 59.243200348128994 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 10.794683502425869 0.9034906625747681 22.884384155273438 1.776257872581482\n",
      "140 58.788975203592656 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 10.62161072548139 0.8985552787780762 22.767061233520508 1.7346867322921753\n",
      "141 58.34615505710019 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 10.457083771375551 0.8928988575935364 22.64852523803711 1.6991219520568848\n",
      "142 57.91333610913627 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 10.303946207975077 0.8874621391296387 22.528932571411133 1.6640626192092896\n",
      "143 57.4846817320604 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 10.152879781128853 0.8827436566352844 22.410287857055664 1.6284825801849365\n",
      "144 57.061572497408484 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 10.005105083506201 0.8776555061340332 22.292316436767578 1.5941790342330933\n",
      "145 56.653662623203935 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 9.869297088421526 0.8724220991134644 22.175588607788086 1.5607662200927734\n",
      "146 56.23400751917451 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 9.73354600041002 0.8672037720680237 22.052146911621094 1.5289639234542847\n",
      "147 55.82714706468854 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 9.608411846640443 0.8610337972640991 21.92865753173828 1.500386357307434\n",
      "148 55.4298167387861 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 9.485674456776973 0.8555960059165955 21.8098201751709 1.4689059257507324\n",
      "149 55.046877190286224 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 9.371408685857357 0.8505058884620667 21.693819046020508 1.4373245239257812\n",
      "150 54.66253628957775 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 9.262268926983149 0.8450113534927368 21.573780059814453 1.4076958894729614\n",
      "151 54.290305911249185 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 9.158725737279916 0.8403500914573669 21.455904006958008 1.3794220685958862\n",
      "152 53.9199958118839 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 9.056655076448415 0.8348814249038696 21.3387393951416 1.350980520248413\n",
      "153 53.563342751964996 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 8.958407284721805 0.8290624022483826 21.225400924682617 1.3250712156295776\n",
      "154 53.20411120759611 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 8.86740611898069 0.8241095542907715 21.104829788208008 1.3029359579086304\n",
      "155 52.865977112814974 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 8.782212917372776 0.818840503692627 20.992403030395508 1.2801176309585571\n",
      "156 52.52466671436345 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 8.700686323337909 0.8135127425193787 20.878511428833008 1.253444790840149\n",
      "157 52.19854312944326 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 8.6244117598525 0.8086917400360107 20.768932342529297 1.2275749444961548\n",
      "158 51.87243343235022 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 8.551720627555909 0.8037616014480591 20.656553268432617 1.203844666481018\n",
      "159 51.55210694136086 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 8.481929982323084 0.7990833520889282 20.54305648803711 1.1849806308746338\n",
      "160 51.208727824036735 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 8.411146210972923 0.7934635281562805 20.420398712158203 1.1633206605911255\n",
      "161 50.897431883563726 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 8.345349344958992 0.7886412143707275 20.310840606689453 1.1417601108551025\n",
      "162 50.58717989875238 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 8.284231543077638 0.7833895683288574 20.20025062561035 1.119057536125183\n",
      "163 50.28052149290339 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 8.225243132059731 0.7786169648170471 20.09084701538086 1.0949673652648926\n",
      "164 49.97958673024243 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 8.170356283426937 0.7734538316726685 19.979490280151367 1.076796054840088\n",
      "165 49.676346918855316 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 8.116447588716156 0.7694499492645264 19.866043090820312 1.0583631992340088\n",
      "166 49.37554246027058 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 8.066530226059609 0.7644156813621521 19.752517700195312 1.039561152458191\n",
      "167 49.08479689446517 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 8.019865382586206 0.7595884799957275 19.6427059173584 1.0199311971664429\n",
      "168 48.790028420807495 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.976184812904968 0.755544900894165 19.528989791870117 1.0003191232681274\n",
      "169 48.50245780297161 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.933598464211238 0.7507541179656982 19.417713165283203 0.9826788902282715\n",
      "170 48.22865092526108 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.894329605204163 0.7464185357093811 19.31151580810547 0.9648711681365967\n",
      "171 47.95357152071503 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.8568757399609375 0.7415539026260376 19.2037353515625 0.9476711750030518\n",
      "172 47.67659734035104 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.823749466186459 0.7368831038475037 19.092012405395508 0.931939959526062\n",
      "173 47.40847437708559 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.794773450781728 0.7319336533546448 18.982540130615234 0.91668701171875\n",
      "174 47.145847255037495 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.763752037332722 0.7278709411621094 18.876508712768555 0.9012068510055542\n",
      "175 46.881172279691825 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.7367241781712845 0.7231118083000183 18.768451690673828 0.8844329118728638\n",
      "176 46.618215453996726 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.710734439268176 0.7194045782089233 18.660184860229492 0.8677067160606384\n",
      "177 46.36571644256733 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.686818465692023 0.7151013016700745 18.55538558959961 0.853025496006012\n",
      "178 46.11186849333051 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.66334991909269 0.7105743885040283 18.449947357177734 0.8380494713783264\n",
      "179 45.85908766763173 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.6400095440622415 0.7054523229598999 18.34423828125 0.8251492381095886\n",
      "180 45.612016171207 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.619506925334501 0.700600266456604 18.240392684936523 0.8111236095428467\n",
      "181 45.36227962457634 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.59943792306877 0.6959055066108704 18.134265899658203 0.7984043955802917\n",
      "182 45.12554890625237 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.582152135298236 0.6918830871582031 18.03261375427246 0.7862861752510071\n",
      "183 44.880740064380255 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.563973265884009 0.6881020069122314 17.927392959594727 0.7738788723945618\n",
      "184 44.636072673271244 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.549387969444336 0.6840501427650452 17.8209171295166 0.760800302028656\n",
      "185 44.40543211352499 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.535488673564506 0.6801092624664307 17.722261428833008 0.7453113198280334\n",
      "186 44.17443257436864 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.522203980542343 0.6761401891708374 17.620624542236328 0.7348393201828003\n",
      "187 43.94045800570635 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.510117000381033 0.6717134118080139 17.517406463623047 0.7238146662712097\n",
      "188 43.7152497654398 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4997336154420875 0.6671764850616455 17.417861938476562 0.7126157879829407\n",
      "189 43.484601647624615 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.489042014846443 0.6626540422439575 17.316125869750977 0.7006538510322571\n",
      "190 43.25943332698484 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4798459389462035 0.6586698889732361 17.214502334594727 0.691912829875946\n",
      "191 43.04260508086529 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.470720091126829 0.6550018191337585 17.11857032775879 0.6797425150871277\n",
      "192 42.827280735971556 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.463200664522276 0.6508432626724243 17.02092933654785 0.6713781356811523\n",
      "193 42.614725436977515 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.456297781280648 0.6462528109550476 16.924583435058594 0.6630079746246338\n",
      "194 42.38686196452481 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.449823423684659 0.6422901749610901 16.820812225341797 0.6531239151954651\n",
      "195 42.179108122517114 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.445028165508752 0.6387898921966553 16.725624084472656 0.644041895866394\n",
      "196 41.97523898433569 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.440053766478329 0.6345674991607666 16.632295608520508 0.6360265016555786\n",
      "197 41.76490817505233 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.435486535609879 0.630759596824646 16.535499572753906 0.6276628971099854\n",
      "198 41.560710869752256 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.431138239823669 0.627534806728363 16.441513061523438 0.6190117001533508\n",
      "199 41.35931584573676 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.427992198237674 0.6232698559761047 16.348329544067383 0.6113947033882141\n",
      "200 41.15664190458657 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.42397636579873 0.6197974681854248 16.255268096923828 0.6023318767547607\n",
      "201 40.9570477629702 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.421722176174393 0.6165955066680908 16.163074493408203 0.5925810933113098\n",
      "202 40.763439778923356 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.419594053863845 0.6134514212608337 16.0732479095459 0.5838984847068787\n",
      "203 40.563453732211585 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.417833803374762 0.6094117760658264 15.979835510253906 0.5765371322631836\n",
      "204 40.36259899422478 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.416668870895624 0.6056748628616333 15.885662078857422 0.5689311027526855\n",
      "205 40.16923833387056 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.415725116898209 0.601682186126709 15.794378280639648 0.5630744695663452\n",
      "206 39.98322058736795 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.415239821071567 0.598639965057373 15.706157684326172 0.5570254325866699\n",
      "207 39.79238905384256 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.414563257238404 0.5949083566665649 15.616353988647461 0.5502094626426697\n",
      "208 39.60335430282251 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.413719145237367 0.5911656022071838 15.52816390991211 0.5421417355537415\n",
      "209 39.40631283239139 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.413563677823671 0.5875739455223083 15.436131477355957 0.5329122543334961\n",
      "210 39.21635485212814 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.413443034100645 0.5834997296333313 15.346257209777832 0.526897668838501\n",
      "211 39.03918804771399 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.413995524711368 0.5794539451599121 15.262722969055176 0.5202926397323608\n",
      "212 38.8614114298707 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.41464982174692 0.5762955546379089 15.177672386169434 0.5151212811470032\n",
      "213 38.687431906725884 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.415371214415551 0.5733711123466492 15.095315933227539 0.508057713508606\n",
      "214 38.51136128349266 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.416819316577527 0.5700252652168274 15.012212753295898 0.5000911951065063\n",
      "215 38.33324717633763 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.418229154277233 0.5665314793586731 14.927055358886719 0.4943758249282837\n",
      "216 38.158072759385576 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.419538308854573 0.563428521156311 14.843362808227539 0.4883803129196167\n",
      "217 37.986663196437824 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.421209279649723 0.5605058073997498 14.760043144226074 0.484861820936203\n",
      "218 37.813325801658266 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.422155955600377 0.5575987696647644 14.677088737487793 0.4793936014175415\n",
      "219 37.64188875391951 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.423495696334747 0.5544713735580444 14.594831466674805 0.47425875067710876\n",
      "220 37.47682228445838 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.42506137251685 0.5512105822563171 14.514978408813477 0.47059351205825806\n",
      "221 37.31056227608878 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.427328669273445 0.5478312373161316 14.435260772705078 0.4648808240890503\n",
      "222 37.14528953942764 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.429221718640207 0.5444357395172119 14.355712890625 0.4602063000202179\n",
      "223 36.98420912591137 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.430858021651879 0.5412301421165466 14.278223037719727 0.4556748867034912\n",
      "224 36.819729781948574 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.432645238243589 0.5382884740829468 14.198195457458496 0.4524051547050476\n",
      "225 36.66448904251155 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.434585880985818 0.5351704955101013 14.123889923095703 0.44695281982421875\n",
      "226 36.50832860012834 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4371332063276 0.5324887633323669 14.048038482666016 0.4426296651363373\n",
      "227 36.347365536434666 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.439473428471102 0.5297684073448181 13.970417022705078 0.43728965520858765\n",
      "228 36.19060857716117 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.441711918740612 0.5264984369277954 13.895325660705566 0.4317469000816345\n",
      "229 36.036885515495 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.443928346677479 0.5239935517311096 13.819774627685547 0.4294143617153168\n",
      "230 35.8813124957355 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.446029639509529 0.521173357963562 13.7442626953125 0.4255841076374054\n",
      "231 35.72877600475254 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.447880503377343 0.5181083679199219 13.671082496643066 0.4206221401691437\n",
      "232 35.58090470197601 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.450373450545497 0.5155351758003235 13.599778175354004 0.4154397249221802\n",
      "233 35.43468599783241 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.453351877797239 0.5125528573989868 13.528377532958984 0.41202619671821594\n",
      "234 35.28806709879643 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.456047471702165 0.5093885660171509 13.456066131591797 0.41049879789352417\n",
      "235 35.14018329886186 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.458829044377678 0.5069668292999268 13.383896827697754 0.40659376978874207\n",
      "236 34.988260302603734 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.461255554020894 0.5045004487037659 13.310376167297363 0.40175196528434753\n",
      "237 34.83953004676152 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.463534410775191 0.5020797848701477 13.238299369812012 0.397317111492157\n",
      "238 34.6906569699036 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.465885291455876 0.4998053312301636 13.16550064086914 0.39396506547927856\n",
      "239 34.544578993695524 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.468982601540832 0.4969423711299896 13.09414291381836 0.39036819338798523\n",
      "240 34.40292270585853 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.470101044624624 0.4946596026420593 13.025948524475098 0.38626500964164734\n",
      "241 34.257248954163806 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.472540066825171 0.49241164326667786 12.954689025878906 0.38291919231414795\n",
      "242 34.116194285644575 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.474563785089535 0.4897848069667816 12.885930061340332 0.3799855709075928\n",
      "243 33.97622862886459 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.476802085151977 0.4869447946548462 12.816902160644531 0.3786774277687073\n",
      "244 33.83412648035581 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.478467889562202 0.484551340341568 12.747932434082031 0.3752423822879791\n",
      "245 33.695000887442454 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.480647117424833 0.48168444633483887 12.679803848266602 0.3730616271495819\n",
      "246 33.55347443717038 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.483640562467928 0.47935280203819275 12.609916687011719 0.37064769864082336\n",
      "247 33.41483340622874 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.485921362657267 0.4762175381183624 12.54199504852295 0.3687044084072113\n",
      "248 33.2752303756244 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.488610205746399 0.4734388589859009 12.47391128540039 0.36535874009132385\n",
      "249 33.1405136607578 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.490622045843855 0.47106608748435974 12.408397674560547 0.36203017830848694\n",
      "250 32.99765458276758 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.492852355747315 0.4681033194065094 12.337799072265625 0.3611007630825043\n",
      "251 32.856817566458396 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.494494788948705 0.4657730162143707 12.268692970275879 0.3591638207435608\n",
      "252 32.715361724014535 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.497056523676171 0.46365463733673096 12.199016571044922 0.35661742091178894\n",
      "253 32.576496436415766 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.498877330838299 0.4614220857620239 12.130712509155273 0.35477200150489807\n",
      "254 32.44473422275172 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.501006342035395 0.4592396020889282 12.065844535827637 0.3527992069721222\n",
      "255 32.30953292761024 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.502461599447095 0.45682865381240845 12.000100135803223 0.3500424027442932\n",
      "256 32.173480452098545 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.504292459287345 0.45464423298835754 11.933501243591309 0.34754127264022827\n",
      "257 32.0361477945632 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.507166615564252 0.45205575227737427 11.865747451782227 0.34543052315711975\n",
      "258 31.90392543963651 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.509459605734149 0.4496021270751953 11.800724983215332 0.34341374039649963\n",
      "259 31.768177218437394 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.511379159689149 0.44714656472206116 11.733708381652832 0.34223473072052\n",
      "260 31.633261878883168 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.513457883989139 0.4450156092643738 11.667550086975098 0.3396882116794586\n",
      "261 31.502594657283623 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.515474207740625 0.44280675053596497 11.602447509765625 0.3394186794757843\n",
      "262 31.371311137461834 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.517989763999157 0.4406566917896271 11.536957740783691 0.3387492001056671\n",
      "263 31.2370505836676 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.520059904165112 0.43872156739234924 11.469935417175293 0.338398277759552\n",
      "264 31.102446255350895 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.521042433644123 0.43612420558929443 11.404091835021973 0.3370959460735321\n",
      "265 30.967043258187612 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.521828480002721 0.4337800145149231 11.337847709655762 0.3357393443584442\n",
      "266 30.834500223844294 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.523216873853449 0.43166404962539673 11.273009300231934 0.3336006999015808\n",
      "267 30.702633646575855 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.524832216350482 0.4298989176750183 11.208325386047363 0.33125174045562744\n",
      "268 30.57476291463385 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.52647772834311 0.42785900831222534 11.145639419555664 0.3291473388671875\n",
      "269 30.444267592913217 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.529186717754907 0.4257306754589081 11.081456184387207 0.3264378309249878\n",
      "270 30.317080268759838 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.530703405041807 0.4235628843307495 11.019126892089844 0.3245601952075958\n",
      "271 30.18482406782765 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.533057767626234 0.4216318428516388 10.953363418579102 0.3234076201915741\n",
      "272 30.05355723213378 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.533749982826149 0.41949835419654846 10.88952922821045 0.32125043869018555\n",
      "273 29.923203318175286 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.536746172961204 0.41695618629455566 10.824344635009766 0.3208116888999939\n",
      "274 29.791876288297182 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.538106205108171 0.41499149799346924 10.76003646850586 0.31870564818382263\n",
      "275 29.66481464247695 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.539351693200981 0.4129829406738281 10.69751262664795 0.3174547553062439\n",
      "276 29.53434686942558 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.540773919062421 0.4108797013759613 10.63323974609375 0.3162137567996979\n",
      "277 29.404261981213217 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.541641627514485 0.40880802273750305 10.569626808166504 0.3145587146282196\n",
      "278 29.283390999255126 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.54332008999342 0.4067721664905548 10.510041236877441 0.31321626901626587\n",
      "279 29.150332934985443 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.54508323205499 0.4051866829395294 10.443799018859863 0.31246498227119446\n",
      "280 29.020838234507984 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.545919421994634 0.403565913438797 10.379622459411621 0.31210798025131226\n",
      "281 28.897996943608327 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.547286909237904 0.4017215073108673 10.319402694702148 0.3101831376552582\n",
      "282 28.777740389793607 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.547962874859066 0.40004730224609375 10.260605812072754 0.30851858854293823\n",
      "283 28.654769134260494 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.549541246630032 0.3980681598186493 10.200125694274902 0.30690833926200867\n",
      "284 28.530625143227816 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.550998010812043 0.39625436067581177 10.138594627380371 0.30618351697921753\n",
      "285 28.408910898638897 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.552708504868679 0.39429882168769836 10.078624725341797 0.3046541213989258\n",
      "286 28.291114425519197 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.55333045707056 0.3925175070762634 10.020825386047363 0.3036156892776489\n",
      "287 28.165653692458996 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.555855648492701 0.3910491466522217 9.958488464355469 0.30177196860313416\n",
      "288 28.041119140496832 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.557037395349127 0.389276385307312 9.897355079650879 0.30009520053863525\n",
      "289 27.91050165745276 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.558320103612122 0.3875470459461212 9.832420349121094 0.2997938096523285\n",
      "290 27.783551711849817 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.560670245699534 0.3858082592487335 9.769707679748535 0.29765784740448\n",
      "291 27.65966574319718 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.562330166186072 0.383505642414093 9.708377838134766 0.29707425832748413\n",
      "292 27.53146157696065 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.563267271063889 0.3813764452934265 9.645630836486816 0.2955561876296997\n",
      "293 27.40491765196596 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.565584236934527 0.38007882237434387 9.582771301269531 0.29371199011802673\n",
      "294 27.27311697619147 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.568408489396045 0.3782570958137512 9.516168594360352 0.29411420226097107\n",
      "295 27.148690113790718 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.5695434007723 0.3766545355319977 9.45498275756836 0.2925266623497009\n",
      "296 27.0252485749217 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.571085291812036 0.3749677538871765 9.393543243408203 0.29210904240608215\n",
      "297 26.900359895992576 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.572205540705023 0.3734598159790039 9.331155776977539 0.29238298535346985\n",
      "298 26.774280398758556 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.573050617846155 0.37144607305526733 9.269469261169434 0.29084518551826477\n",
      "299 26.64716900117255 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.573748446773061 0.36981314420700073 9.20665454864502 0.29029831290245056\n",
      "300 26.525667952861735 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.574101673927255 0.36797645688056946 9.146994590759277 0.2896006405353546\n",
      "301 26.40552627371727 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.5740050308602935 0.36648455262184143 9.08794116973877 0.2891543507575989\n",
      "302 26.280724614969053 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.5747389496477044 0.36455008387565613 9.026775360107422 0.2878848612308502\n",
      "303 26.15778721157197 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.575913913880628 0.3627343475818634 8.966087341308594 0.2869642674922943\n",
      "304 26.038851167329987 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.575973834165774 0.3609144985675812 8.907957077026367 0.28604868054389954\n",
      "305 25.91722525879264 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.576503532375101 0.3593379259109497 8.847394943237305 0.2865939140319824\n",
      "306 25.7908656811315 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.577107957561632 0.3576390743255615 8.785340309143066 0.28543803095817566\n",
      "307 25.666217624961998 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.577897727310326 0.355739027261734 8.723671913146973 0.28523704409599304\n",
      "308 25.545794991045504 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.578591522961167 0.35396507382392883 8.664122581481934 0.2849932312965393\n",
      "309 25.427916604199147 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.578918385424354 0.3517865836620331 8.60607624053955 0.2850591540336609\n",
      "310 25.30639541268698 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.579788537028945 0.3504747450351715 8.546021461486816 0.28408920764923096\n",
      "311 25.18495255152544 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.580376172464698 0.34882792830467224 8.48648452758789 0.28277939558029175\n",
      "312 25.06500270624057 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.581804391335402 0.3470458984375 8.42692756652832 0.282297283411026\n",
      "313 24.94197285834472 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.582447595035241 0.34544864296913147 8.366294860839844 0.2814868986606598\n",
      "314 24.822147606797486 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.583885013051299 0.34398674964904785 8.307546615600586 0.2791826128959656\n",
      "315 24.703452692159463 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.584567622073937 0.3421832323074341 8.249441146850586 0.27781954407691956\n",
      "316 24.585387536042127 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.5843973240784734 0.3408834934234619 8.191596984863281 0.27691274881362915\n",
      "317 24.469301007374224 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.585889003857072 0.33928343653678894 8.134164810180664 0.2757989466190338\n",
      "318 24.35232120275433 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.585911000966379 0.3381105363368988 8.07685661315918 0.27458643913269043\n",
      "319 24.23523993447465 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.586166184693996 0.33638206124305725 8.020024299621582 0.2726430892944336\n",
      "320 24.119897619252395 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.584974304204177 0.33465397357940674 7.963961601257324 0.2723461389541626\n",
      "321 24.00589188498119 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.584860320507083 0.33313658833503723 7.908468723297119 0.2709575295448303\n",
      "322 23.893976819502832 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.584780913578989 0.3319123089313507 7.853544235229492 0.2701951265335083\n",
      "323 23.77973078950471 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.584143005452814 0.3306238651275635 7.797033309936523 0.2708972990512848\n",
      "324 23.66864213130634 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.584560451723743 0.3292485475540161 7.742847919464111 0.26913729310035706\n",
      "325 23.55719552142004 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.585398382967511 0.3279305398464203 7.687515735626221 0.2688351273536682\n",
      "326 23.44735042092376 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.584648379059367 0.3263964354991913 7.634007453918457 0.2682906985282898\n",
      "327 23.33483614285285 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.583963232830113 0.3250352144241333 7.579745292663574 0.26634711027145386\n",
      "328 23.221043752966093 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.584035741624998 0.32357800006866455 7.5235161781311035 0.2663976550102234\n",
      "329 23.112167691573923 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.584376727924171 0.3220177888870239 7.4708404541015625 0.26409226655960083\n",
      "330 23.007747050089996 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.584030445380371 0.3204912543296814 7.419710159301758 0.2638050317764282\n",
      "331 22.90002008815565 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.583662803371334 0.31882813572883606 7.367596626281738 0.2623358964920044\n",
      "332 22.79699878337437 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.582341297889295 0.31750452518463135 7.317690372467041 0.2617722153663635\n",
      "333 22.69631627061574 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.581523536231202 0.31590574979782104 7.268494129180908 0.2618987262248993\n",
      "334 22.59232517342962 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.580182366184365 0.31451719999313354 7.218736171722412 0.26015326380729675\n",
      "335 22.49152481368201 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.57946550658362 0.3128942847251892 7.170039653778076 0.2590857148170471\n",
      "336 22.390526303607736 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.577561238366876 0.3115127682685852 7.120826244354248 0.2597998082637787\n",
      "337 22.29701417456696 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.574592558251117 0.31021687388420105 7.076444625854492 0.25931549072265625\n",
      "338 22.19818564178813 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.572407592557555 0.30877697467803955 7.030126571655273 0.25674793124198914\n",
      "339 22.106387722160903 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.57121058073291 0.30718672275543213 6.98603630065918 0.25591781735420227\n",
      "340 22.016698446012807 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.569194313026739 0.3058052361011505 6.944194793701172 0.25330930948257446\n",
      "341 21.929691142771652 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.5673802265862715 0.30446943640708923 6.902621269226074 0.25259894132614136\n",
      "342 21.838344362359937 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.564910617452557 0.30349549651145935 6.859152793884277 0.2516326606273651\n",
      "343 21.75098119941701 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.562754738006485 0.30215850472450256 6.81734561920166 0.2513767182826996\n",
      "344 21.66345077821599 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.560561627834902 0.30039551854133606 6.775860786437988 0.25077205896377563\n",
      "345 21.582054874174503 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.558083882801442 0.29884546995162964 6.737839221954346 0.2494470775127411\n",
      "346 21.4970784668507 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.556454036075055 0.2975021302700043 6.6976165771484375 0.24788914620876312\n",
      "347 21.417569667440645 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.554204658013573 0.29598236083984375 6.661088943481445 0.24520476162433624\n",
      "348 21.331397277688595 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.551787701820944 0.2945418655872345 6.620954513549805 0.24315868318080902\n",
      "349 21.2477556889477 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.549276543182396 0.29310768842697144 6.58206844329834 0.24123457074165344\n",
      "350 21.167624164638013 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.547707621273487 0.29207807779312134 6.543906211853027 0.24002604186534882\n",
      "351 21.08805664285894 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.546665732114275 0.2902788519859314 6.506427764892578 0.2382565289735794\n",
      "352 21.00873649306691 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.544290946797492 0.28868600726127625 6.469754219055176 0.236251100897789\n",
      "353 20.92829912084737 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.542296118322996 0.28719598054885864 6.432206630706787 0.23439376056194305\n",
      "354 20.85039632080883 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.539540082801285 0.2855771780014038 6.39591121673584 0.23345662653446198\n",
      "355 20.773406372503683 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.5375268469614225 0.28428220748901367 6.359344005584717 0.23290930688381195\n",
      "356 20.696601861219527 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.53483974287999 0.2830406129360199 6.3239216804504395 0.23087814450263977\n",
      "357 20.618138138977297 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.5335442021763415 0.2818650007247925 6.2870564460754395 0.22861604392528534\n",
      "358 20.54063611163683 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.53187664284296 0.2806873023509979 6.2507524490356445 0.22656726837158203\n",
      "359 20.465917947743158 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.528891774866802 0.27925845980644226 6.216352462768555 0.2250627875328064\n",
      "360 20.391431930911573 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.528059321892291 0.2780076861381531 6.180700302124023 0.22396431863307953\n",
      "361 20.31599070405247 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.52571222459557 0.2763177156448364 6.145895481109619 0.22216980159282684\n",
      "362 20.238287791713645 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.524615913733413 0.2748536765575409 6.108892440795898 0.22103331983089447\n",
      "363 20.15819003239704 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.5239314045055 0.27342528104782104 6.070086479187012 0.22066038846969604\n",
      "364 20.085542435905566 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.522229503056635 0.2723359167575836 6.035439968109131 0.22009707987308502\n",
      "365 20.00947790556585 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.519842676455131 0.2715303301811218 5.999337673187256 0.21942955255508423\n",
      "366 19.935343441406403 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.5180646151696795 0.27037742733955383 5.964043140411377 0.21881511807441711\n",
      "367 19.860497421687114 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.51551643676089 0.26943618059158325 5.928433418273926 0.21867796778678894\n",
      "368 19.785784941418076 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.515191894276047 0.26833125948905945 5.891845226287842 0.21857133507728577\n",
      "369 19.713123542849154 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.5132351495662615 0.2675692141056061 5.857186317443848 0.21794654428958893\n",
      "370 19.638255643435613 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.511871042438642 0.26665616035461426 5.821087837219238 0.2175527662038803\n",
      "371 19.562539712424382 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.5101185747338395 0.2654450535774231 5.785500526428223 0.21597503125667572\n",
      "372 19.48993026490429 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.509099580600053 0.2639746069908142 5.750584602355957 0.2156868726015091\n",
      "373 19.416279801236065 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.50842020760623 0.26315081119537354 5.7151312828063965 0.21444621682167053\n",
      "374 19.340888274662774 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.507568387143893 0.2621130347251892 5.678761005401611 0.2136848419904709\n",
      "375 19.268030711143766 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.506120585530555 0.2610655128955841 5.643756866455078 0.21333087980747223\n",
      "376 19.195854272737066 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.504724826707401 0.2600845992565155 5.609325408935547 0.21239402890205383\n",
      "377 19.122480157465297 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.503492552489595 0.2589924931526184 5.57363748550415 0.2127201408147812\n",
      "378 19.055414425287946 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.502485605035527 0.2580864727497101 5.541464328765869 0.21191368997097015\n",
      "379 18.98222157442529 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.500716448664258 0.25703033804893494 5.506392955780029 0.21168887615203857\n",
      "380 18.908545935319612 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.500448876784989 0.2560102939605713 5.471062660217285 0.2099614441394806\n",
      "381 18.834907567415684 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.499482831273526 0.25470206141471863 5.435801982879639 0.20911870896816254\n",
      "382 18.765868675410793 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4981383648515205 0.253614604473114 5.402891635894775 0.20833243429660797\n",
      "383 18.696821943624126 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.497149751288997 0.25305110216140747 5.369448661804199 0.20772376656532288\n",
      "384 18.62328346488467 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.496546883559137 0.251790314912796 5.334076404571533 0.20679345726966858\n",
      "385 18.550370504369205 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.495536496152348 0.25081267952919006 5.298832416534424 0.2063564956188202\n",
      "386 18.47777552116157 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.494920618417277 0.24983817338943481 5.263781547546387 0.20545363426208496\n",
      "387 18.405082559231012 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.493803134206024 0.2487516552209854 5.228762149810791 0.20500347018241882\n",
      "388 18.331150965116002 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.492794738433338 0.247653529047966 5.193589687347412 0.20352332293987274\n",
      "389 18.258589490813705 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.491996302766297 0.24663563072681427 5.158041954040527 0.2038736492395401\n",
      "390 18.186884983585028 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.491242691085485 0.24552571773529053 5.123292446136475 0.2035316824913025\n",
      "391 18.11538966892316 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.490263828616878 0.2448793202638626 5.0885820388793945 0.20308244228363037\n",
      "392 18.043994789119655 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4891573238334 0.24351705610752106 5.054175853729248 0.20296870172023773\n",
      "393 17.9766985194053 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.488536866125042 0.24243636429309845 5.021479606628418 0.2027660757303238\n",
      "394 17.907343421493017 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.487176258478604 0.2413756400346756 4.988251686096191 0.20228815078735352\n",
      "395 17.836800842205164 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.486775337377666 0.24058766663074493 4.9538893699646 0.20165909826755524\n",
      "396 17.766425091113344 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.486605542983309 0.2397860735654831 4.919692516326904 0.2006484419107437\n",
      "397 17.695789742645402 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.485769438919204 0.23871545493602753 4.88517951965332 0.20094580948352814\n",
      "398 17.626050196191308 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.485086135169505 0.23792649805545807 4.851072788238525 0.20089198648929596\n",
      "399 17.557422018782507 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.484027690426719 0.23694749176502228 4.818075180053711 0.20029647648334503\n",
      "400 17.489163255079365 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4839289390879635 0.23605482280254364 4.784621715545654 0.19993606209754944\n",
      "401 17.41795999280636 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.483871464650077 0.23510709404945374 4.749980926513672 0.19901958107948303\n",
      "402 17.348330049053857 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.482705114618964 0.23406089842319489 4.716139793395996 0.19928444921970367\n",
      "403 17.276839119728525 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.482560170229395 0.23300625383853912 4.680913925170898 0.19944484531879425\n",
      "404 17.209613466401635 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4813443364101015 0.23201268911361694 4.648940563201904 0.19837531447410583\n",
      "405 17.13975131252858 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.481002819930048 0.2309727519750595 4.614904880523682 0.19796597957611084\n",
      "406 17.06902707168182 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.479821642723881 0.23010894656181335 4.580376625061035 0.19834323227405548\n",
      "407 17.000995936357967 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.478968875968448 0.22915717959403992 4.547366142272949 0.19813759624958038\n",
      "408 16.93727809973253 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.478617184238567 0.2282901257276535 4.516208648681641 0.1979534924030304\n",
      "409 16.868804589738158 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.478487924565582 0.22748063504695892 4.482852935791016 0.19713015854358673\n",
      "410 16.80142061670214 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.477514791087209 0.2266472578048706 4.450591564178467 0.196075439453125\n",
      "411 16.732214431609282 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.477166752661835 0.2256084382534027 4.417041301727295 0.19535663723945618\n",
      "412 16.664075395855207 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.477066672238605 0.22478975355625153 4.38368034362793 0.19485828280448914\n",
      "413 16.59941301915112 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.477136862010392 0.22403626143932343 4.351936340332031 0.1943672150373459\n",
      "414 16.52929943778004 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.47725939728703 0.2232624888420105 4.317320346832275 0.1941368579864502\n",
      "415 16.462483217751164 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.476351981794016 0.2224561870098114 4.284791469573975 0.19409210979938507\n",
      "416 16.39647419391584 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.475514855370041 0.22153392434120178 4.2528181076049805 0.19378919899463654\n",
      "417 16.33036240560744 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.475368212887977 0.22094638645648956 4.220243453979492 0.1935608983039856\n",
      "418 16.26498346088568 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.475138679725326 0.22020073235034943 4.188233375549316 0.19317729771137238\n",
      "419 16.19677455617696 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.474328978793917 0.2195463329553604 4.155075550079346 0.19274814426898956\n",
      "420 16.129157822385356 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.473730798616931 0.21878917515277863 4.122226715087891 0.1921844184398651\n",
      "421 16.05982710716763 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4737736165288995 0.21804630756378174 4.0878777503967285 0.19225168228149414\n",
      "422 15.99116379653966 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.473093694801682 0.2169974446296692 4.054412841796875 0.19224697351455688\n",
      "423 15.927225931795011 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4724512540415615 0.2162799835205078 4.023541450500488 0.1914117932319641\n",
      "424 15.86187393708563 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.471643877461911 0.215439572930336 3.9915709495544434 0.19164858758449554\n",
      "425 15.794181027296723 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.471586757186593 0.21452470123767853 3.9584603309631348 0.19114890694618225\n",
      "426 15.730397997274217 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.470864383354006 0.21379314363002777 3.92779541015625 0.19014964997768402\n",
      "427 15.663323140136693 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4704807907266355 0.21294420957565308 3.894814968109131 0.19026820361614227\n",
      "428 15.598041775107927 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.469973477006501 0.2120295763015747 3.8627798557281494 0.19047901034355164\n",
      "429 15.532172455568716 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.469416930456565 0.21116755902767181 3.8305304050445557 0.19052715599536896\n",
      "430 15.467967981312134 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.469284931037285 0.21027614176273346 3.7990965843200684 0.19021373987197876\n",
      "431 15.402912509034758 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4687620963560395 0.20959407091140747 3.7672128677368164 0.19013060629367828\n",
      "432 15.3386485955576 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.467736371503264 0.20893169939517975 3.7361679077148438 0.1896447092294693\n",
      "433 15.276851915529924 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4677620462691365 0.20810814201831818 3.7060458660125732 0.1888899952173233\n",
      "434 15.214617233120624 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.467335383736316 0.20764493942260742 3.675501823425293 0.1886332631111145\n",
      "435 15.151240706312171 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.467751979696265 0.2069559097290039 3.64424991607666 0.18803298473358154\n",
      "436 15.090471238504591 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.467116818438711 0.20619063079357147 3.6141860485076904 0.1887916922569275\n",
      "437 15.028391985679935 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4664287404072995 0.20550920069217682 3.583265781402588 0.1899224817752838\n",
      "438 14.96715006142389 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4658600828005 0.20486421883106232 3.5535194873809814 0.189386785030365\n",
      "439 14.90870423868646 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.464666056189435 0.20421625673770905 3.525195598602295 0.18943072855472565\n",
      "440 14.848462733116556 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.464353191342759 0.20363669097423553 3.4955031871795654 0.1894664764404297\n",
      "441 14.78800311896141 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.464290308087424 0.20324674248695374 3.4658565521240234 0.18875296413898468\n",
      "442 14.726985224854047 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.463917189609105 0.2026415765285492 3.436115026473999 0.18819640576839447\n",
      "443 14.6681255874765 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.463177430404433 0.20201797783374786 3.4075212478637695 0.18788768351078033\n",
      "444 14.61006097406125 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.461328416168451 0.20135225355625153 3.3804144859313965 0.18655133247375488\n",
      "445 14.551868119790981 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4604442357568255 0.20052705705165863 3.3524413108825684 0.18601420521736145\n",
      "446 14.495552471373005 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4587675542568395 0.19963331520557404 3.3258309364318848 0.1854897290468216\n",
      "447 14.438275899460951 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.45777655737702 0.19916027784347534 3.2981345653533936 0.18506993353366852\n",
      "448 14.38196205775579 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.456957739399275 0.1985146850347519 3.2707858085632324 0.18491801619529724\n",
      "449 14.324660329568063 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.456379695284044 0.1981041580438614 3.2429447174072266 0.18428704142570496\n",
      "450 14.26988028167383 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.455240951763551 0.19711117446422577 3.217014789581299 0.1834985762834549\n",
      "451 14.215191536471428 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4541531883927 0.19647862017154694 3.1906638145446777 0.18323209881782532\n",
      "452 14.16072735941432 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.452882982850294 0.19602732360363007 3.164902687072754 0.182011678814888\n",
      "453 14.107375822592733 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.45165038817024 0.19529736042022705 3.1397275924682617 0.1809728890657425\n",
      "454 14.053706071162852 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.45070230104986 0.1947575956583023 3.1144540309906006 0.17933811247348785\n",
      "455 14.001955926595258 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4501870723348125 0.1941804140806198 3.089245557785034 0.17909732460975647\n",
      "456 13.951188309523577 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.449657796356195 0.19370359182357788 3.0647382736206055 0.17835037410259247\n",
      "457 13.900383723591071 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.44901786527465 0.1932205706834793 3.040700674057007 0.17674393951892853\n",
      "458 13.852048492398849 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.448170489040008 0.19248756766319275 3.0174710750579834 0.1764482855796814\n",
      "459 13.803170951692332 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.446869569508303 0.1917235553264618 2.994217872619629 0.17614208161830902\n",
      "460 13.754512092771261 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.446763402523725 0.19121798872947693 2.9709372520446777 0.1746561974287033\n",
      "461 13.7066792509905 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.446122478639237 0.19072356820106506 2.9481306076049805 0.17357198894023895\n",
      "462 13.65940657862287 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.444970300875715 0.19034649431705475 2.9256300926208496 0.17282959818840027\n",
      "463 13.613308680845723 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.444064331842884 0.1897858828306198 2.903710126876831 0.1720382124185562\n",
      "464 13.566297243837326 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.443130310181588 0.18921595811843872 2.8810009956359863 0.17194898426532745\n",
      "465 13.523405662942867 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.442368633318882 0.18865059316158295 2.8604772090911865 0.1714320182800293\n",
      "466 13.47868023224343 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.440962070302856 0.18816056847572327 2.8396737575531006 0.1702100783586502\n",
      "467 13.433699542079957 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.439591699634583 0.18767879903316498 2.8188259601593018 0.16877712309360504\n",
      "468 13.392847999078079 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4386514118731455 0.18728917837142944 2.799165964126587 0.1685754805803299\n",
      "469 13.350268166951825 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.437528800658872 0.18667072057724 2.778604507446289 0.16885963082313538\n",
      "470 13.3078413436766 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.436377284919772 0.18626396358013153 2.75900936126709 0.1671813726425171\n",
      "471 13.265304483996967 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4354228606096315 0.18575479090213776 2.7393784523010254 0.1653699278831482\n",
      "472 13.22352766378074 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.434123882485918 0.18522575497627258 2.719914197921753 0.16434963047504425\n",
      "473 13.181252112254896 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.43342608736495 0.18461348116397858 2.7003700733184814 0.16247239708900452\n",
      "474 13.141696876043362 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.43243375992207 0.18421047925949097 2.681788444519043 0.1614757478237152\n",
      "475 13.101259843293521 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.431860015574786 0.18364758789539337 2.6624257564544678 0.16090072691440582\n",
      "476 13.060713837560204 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.431168476756601 0.1830599159002304 2.643245220184326 0.15999500453472137\n",
      "477 13.021287454586696 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.430674178820324 0.18247289955615997 2.624812602996826 0.15851517021656036\n",
      "478 12.981779456915618 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.429946065726043 0.1819847673177719 2.6062090396881104 0.15743054449558258\n",
      "479 12.943446049480176 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4288540873811 0.1815052628517151 2.588101387023926 0.15688392519950867\n",
      "480 12.902129911226185 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.427957542580517 0.18108420073986053 2.568240165710449 0.15660783648490906\n",
      "481 12.864793696526894 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.426743933085808 0.180597722530365 2.550572156906128 0.15630772709846497\n",
      "482 12.826592970490747 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.425510365367227 0.1801736056804657 2.5322775840759277 0.15635383129119873\n",
      "483 12.78886999573638 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4246019610398015 0.17964757978916168 2.5140905380249023 0.1564393788576126\n",
      "484 12.750740428104214 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.424203727974705 0.17935475707054138 2.4956889152526855 0.1558041125535965\n",
      "485 12.711650005630265 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.423239103606949 0.17883335053920746 2.477384090423584 0.15480937063694\n",
      "486 12.67316056411731 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.422680851034996 0.17843963205814362 2.4587562084198 0.1545276641845703\n",
      "487 12.635383239457186 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.421584194967325 0.17790700495243073 2.441150188446045 0.15359166264533997\n",
      "488 12.597758668027506 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.421225103102313 0.17735162377357483 2.4233968257904053 0.1523882895708084\n",
      "489 12.559459657199357 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.420571476943467 0.17680665850639343 2.404917001724243 0.15224751830101013\n",
      "490 12.52324168587694 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.420177789578531 0.17644339799880981 2.3873131275177 0.1519942432641983\n",
      "491 12.48717265787047 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.41975876154822 0.1759604662656784 2.370274066925049 0.1509052962064743\n",
      "492 12.450813511146936 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.418935129179392 0.17546501755714417 2.353095769882202 0.1502218246459961\n",
      "493 12.414091412991343 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.418619309634027 0.17492380738258362 2.3354971408843994 0.14955401420593262\n",
      "494 12.379810967157821 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.417987026880722 0.17443794012069702 2.3192732334136963 0.14883953332901\n",
      "495 12.346656694081013 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.417280553844157 0.174148291349411 2.30354380607605 0.14814023673534393\n",
      "496 12.311965459577129 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.416526728860424 0.17382824420928955 2.2872278690338135 0.14715474843978882\n",
      "497 12.277029350714264 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.416140243725357 0.17345701158046722 2.2703866958618164 0.14665870368480682\n",
      "498 12.24372885674961 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.41556515843399 0.1730569303035736 2.2549774646759033 0.14515183866024017\n",
      "499 12.20865762947069 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.415364347081049 0.1727619767189026 2.238067388534546 0.14439652860164642\n",
      "500 12.175512661918631 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.414664169534674 0.17260460555553436 2.2214930057525635 0.1452578753232956\n",
      "501 12.143206118391905 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.41377989833251 0.17218373715877533 2.2062206268310547 0.14480122923851013\n",
      "502 12.109209250450066 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.413704018831185 0.1717689484357834 2.189656972885132 0.14442233741283417\n",
      "503 12.07660744903058 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.413300124978724 0.1714363694190979 2.173736810684204 0.1443973332643509\n",
      "504 12.042714069478876 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.412492225759393 0.17108550667762756 2.157217502593994 0.14470133185386658\n",
      "505 12.009299377477575 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.412121454751897 0.17086391150951385 2.141369104385376 0.14357580244541168\n",
      "506 11.975693132683379 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4121282360998215 0.17051458358764648 2.1251041889190674 0.14284193515777588\n",
      "507 11.943258829438768 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.411899082386576 0.17006716132164001 2.109476327896118 0.14233992993831635\n",
      "508 11.910104990063086 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.411610573587789 0.16965927183628082 2.093252658843994 0.14232982695102692\n",
      "509 11.877911796134008 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4112041492872365 0.16948916018009186 2.0778074264526367 0.14160363376140594\n",
      "510 11.845125661224678 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4109624931519775 0.1688595414161682 2.0622665882110596 0.14077045023441315\n",
      "511 11.812236043689365 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.410566958663578 0.16842631995677948 2.046614646911621 0.14001347124576569\n",
      "512 11.779461331761686 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.410251252449362 0.1678502857685089 2.0305216312408447 0.14031653106212616\n",
      "513 11.748581277666784 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.409914928017353 0.16755585372447968 2.0155327320098877 0.14004503190517426\n",
      "514 11.716944518776499 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.409533861370646 0.16720739006996155 2.0001413822174072 0.13992050290107727\n",
      "515 11.68485708333502 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.409408999457456 0.1668534129858017 1.9845350980758667 0.13952447474002838\n",
      "516 11.652881975132992 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.409070263464024 0.1666197031736374 1.9691694974899292 0.1388530135154724\n",
      "517 11.621903271663582 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.408501164305603 0.16631989181041718 1.9543051719665527 0.13847187161445618\n",
      "518 11.589328692417178 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.408271149854693 0.16587406396865845 1.9383583068847656 0.13846686482429504\n",
      "519 11.5603589217428 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.407709876752261 0.16558048129081726 1.924272894859314 0.13852277398109436\n",
      "520 11.529685573084466 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.407446445329301 0.16540692746639252 1.9099394083023071 0.13695338368415833\n",
      "521 11.49813651676748 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.407188337403275 0.16506093740463257 1.8942586183547974 0.1373700052499771\n",
      "522 11.467198765621294 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.407088733062852 0.16472500562667847 1.879511833190918 0.13636136054992676\n",
      "523 11.437465685708613 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.407124656541438 0.1643076092004776 1.8648972511291504 0.1362389177083969\n",
      "524 11.409122360914047 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.407001404135521 0.1640584021806717 1.85102117061615 0.1360202133655548\n",
      "525 11.380080759850358 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.406857565370416 0.163910910487175 1.8367323875427246 0.13584750890731812\n",
      "526 11.348948022894788 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.406245833806921 0.16357974708080292 1.8218252658843994 0.135471910238266\n",
      "527 11.32095201233697 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.406141429341556 0.16344580054283142 1.8083152770996094 0.13473422825336456\n",
      "528 11.291731506344691 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.40572456991376 0.16302955150604248 1.7942016124725342 0.13457415997982025\n",
      "529 11.263126435310763 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.405296849639339 0.1629197895526886 1.780336618423462 0.13423655927181244\n",
      "530 11.23226104173155 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.40491739901991 0.16256381571292877 1.7657113075256348 0.1333572119474411\n",
      "531 11.204231022970909 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4046600750853955 0.16236743330955505 1.7521361112594604 0.13293129205703735\n",
      "532 11.17668081204187 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.404169821541419 0.1621646285057068 1.7388516664505005 0.13264302909374237\n",
      "533 11.148402329489787 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.404158305332263 0.16181960701942444 1.725307822227478 0.13180877268314362\n",
      "534 11.120294756288432 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.404162935013675 0.16157925128936768 1.7116572856903076 0.13123799860477448\n",
      "535 11.091566181627666 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.404034129229939 0.1613924205303192 1.6975730657577515 0.13099350035190582\n",
      "536 11.06412873389224 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.404018796701231 0.16119955480098724 1.6841660737991333 0.13057823479175568\n",
      "537 11.034811411809482 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.403541778039494 0.16099698841571808 1.6698198318481445 0.13063298165798187\n",
      "538 11.006036956772597 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.403213341698439 0.1607840210199356 1.6559230089187622 0.1301935762166977\n",
      "539 10.978465095956533 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.403216437299459 0.16051428020000458 1.6420114040374756 0.13071157038211823\n",
      "540 10.950384675830588 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.403159483641372 0.16024376451969147 1.6285176277160645 0.12994617223739624\n",
      "541 10.921882112303837 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.402623151222333 0.15996785461902618 1.6146320104599 0.13002708554267883\n",
      "542 10.894804893084057 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.402703760214337 0.15968173742294312 1.6014043092727661 0.12961077690124512\n",
      "543 10.86840826127691 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.402659563757608 0.1593238115310669 1.5885626077651978 0.12929967045783997\n",
      "544 10.8417423955147 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.402259989732954 0.15909571945667267 1.5756311416625977 0.12912440299987793\n",
      "545 10.813645149951132 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.40218181479469 0.15879659354686737 1.5623823404312134 0.1279020607471466\n",
      "546 10.784607168036324 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.402147482829911 0.15834128856658936 1.5480886697769165 0.1279410570859909\n",
      "547 10.758511588236775 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.401927396914448 0.15794628858566284 1.5354228019714355 0.12779229879379272\n",
      "548 10.732355561343176 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.402108740297302 0.1578616499900818 1.5227000713348389 0.12698502838611603\n",
      "549 10.705489760178967 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4019556696600155 0.15766490995883942 1.5095138549804688 0.1268414705991745\n",
      "550 10.678987689849793 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4017450144261705 0.15731094777584076 1.4967113733291626 0.12650898098945618\n",
      "551 10.65198722513691 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.401555344782029 0.1570642590522766 1.483524203300476 0.12631921470165253\n",
      "552 10.625705336763616 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.4015845857403235 0.15685895085334778 1.470337986946106 0.12658582627773285\n",
      "553 10.601183436868205 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.401018790957942 0.15643425285816193 1.4588931798934937 0.12594403326511383\n",
      "554 10.57511025721094 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.400640413704126 0.15619350969791412 1.4466842412948608 0.12490785121917725\n",
      "555 10.54888338840893 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.400429087349684 0.15587493777275085 1.4339938163757324 0.12459173053503036\n",
      "556 10.524648129741728 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.400491975049555 0.15562628209590912 1.4220985174179077 0.12433283776044846\n",
      "557 10.49925350848431 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.400321249699539 0.155352383852005 1.4099825620651245 0.12361475080251694\n",
      "558 10.473654279113568 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.400049747527397 0.15498624742031097 1.3977148532867432 0.1231885775923729\n",
      "559 10.448402110948644 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.399935108139596 0.1547147035598755 1.3855862617492676 0.12257977575063705\n",
      "560 10.423690061716407 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.399788360742896 0.15451133251190186 1.3733688592910767 0.12265264987945557\n",
      "561 10.3982334840765 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.399714108156204 0.1543634533882141 1.3609645366668701 0.12222684919834137\n",
      "562 10.37366047847927 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.399615514761809 0.15390029549598694 1.3489587306976318 0.12222720682621002\n",
      "563 10.349037192169108 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.399407460692801 0.15358878672122955 1.3366540670394897 0.12273281067609787\n",
      "564 10.324414825162494 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.3994864282222625 0.15327271819114685 1.3247185945510864 0.12221848964691162\n",
      "565 10.298932181739339 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.399364585124978 0.15315133333206177 1.312632441520691 0.1211513802409172\n",
      "566 10.275902454915602 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.3989273879557445 0.15280742943286896 1.3013683557510376 0.12143092602491379\n",
      "567 10.252053677795466 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.3986406696420275 0.15258005261421204 1.2899316549301147 0.12096964567899704\n",
      "568 10.229242451135944 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.3982710983799205 0.15235286951065063 1.2789301872253418 0.12075810879468918\n",
      "569 10.204390536974419 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.398122120629298 0.15214373171329498 1.2668930292129517 0.12033862620592117\n",
      "570 10.181108846325952 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.397890790362436 0.15195833146572113 1.2553744316101074 0.12051086127758026\n",
      "571 10.158209701612545 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.39793477727468 0.15168018639087677 1.2440333366394043 0.12052806466817856\n",
      "572 10.135758600792116 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.397447101434892 0.15133605897426605 1.233331561088562 0.12031231820583344\n",
      "573 10.111443942893263 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.397176517654178 0.1512119323015213 1.221760869026184 0.11953375488519669\n",
      "574 10.088450083766062 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.397115024063666 0.1509668231010437 1.2104713916778564 0.1194254532456398\n",
      "575 10.064988088624984 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.397188094275505 0.15076032280921936 1.1992998123168945 0.11844004690647125\n",
      "576 10.043310908291229 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.396899154100308 0.15071038901805878 1.1885309219360352 0.1186395213007927\n",
      "577 10.021601183968507 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.39675649877163 0.15029799938201904 1.1778727769851685 0.11880113184452057\n",
      "578 10.000918413884154 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.396663668937197 0.14995041489601135 1.1677119731903076 0.11888038367033005\n",
      "579 9.978981879592407 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.396447223663795 0.14967641234397888 1.1572362184524536 0.11838580667972565\n",
      "580 9.957748883423067 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.3963043773739106 0.14954815804958344 1.1467928886413574 0.11831057071685791\n",
      "581 9.935460987132121 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.3962688842826365 0.14930938184261322 1.1360951662063599 0.11769238859415054\n",
      "582 9.914164953390792 tensor(2, device='cuda:4') tensor(0, device='cuda:4') 7.396057449738218 0.14890024065971375 1.1259479522705078 0.11731135845184326\n",
      "583 9.892914533533768 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.395929776050762 0.14865918457508087 1.1158721446990967 0.11658128350973129\n",
      "584 9.871456496438148 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.395645499250056 0.14828069508075714 1.1053764820098877 0.11677733808755875\n",
      "585 9.851182027335254 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.395443102891532 0.14784157276153564 1.0956443548202515 0.1166086420416832\n",
      "586 9.829404086089724 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.395172283268565 0.147671177983284 1.0850389003753662 0.11648282408714294\n",
      "587 9.807951653830997 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.39515156530618 0.14742876589298248 1.0743424892425537 0.11668634414672852\n",
      "588 9.787091641061345 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.395041158728638 0.14724627137184143 1.0644110441207886 0.11598212271928787\n",
      "589 9.767276032487292 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.394822201351064 0.14691706001758575 1.0550817251205444 0.11537332087755203\n",
      "590 9.74585358558255 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.395063338261421 0.1466461420059204 1.0446261167526245 0.11489187180995941\n",
      "591 9.726720506701492 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.394833038005851 0.14637984335422516 1.035642385482788 0.11422285437583923\n",
      "592 9.706390169144868 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.394545976699589 0.14619463682174683 1.0260612964630127 0.11352696269750595\n",
      "593 9.685602505003185 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.394493695591659 0.14584781229496002 1.016019582748413 0.11322183161973953\n",
      "594 9.665705439908542 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.394314130110778 0.14547036588191986 1.0065045356750488 0.11291187256574631\n",
      "595 9.645665105860873 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.394196492117091 0.14521515369415283 0.9967520236968994 0.11274941265583038\n",
      "596 9.626498989779547 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.393886253256396 0.1448691338300705 0.9876031279563904 0.11253734678030014\n",
      "597 9.608285253334097 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.393496421563677 0.1445528119802475 0.9787417054176331 0.11275260895490646\n",
      "598 9.588926784938101 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.393147059028868 0.14410486817359924 0.9698390960693359 0.11199666559696198\n",
      "599 9.570104701557556 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.393138041832843 0.14397233724594116 0.960874080657959 0.11124616116285324\n",
      "600 9.550367472219204 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.393014220046734 0.1436203122138977 0.9515615701675415 0.11060979962348938\n",
      "601 9.532420275780485 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.392867511185931 0.1434219926595688 0.9427064657211304 0.11071784049272537\n",
      "602 9.514521140503874 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.392946313965312 0.1431656926870346 0.9341220259666443 0.1101650819182396\n",
      "603 9.496913354162103 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.392729769710429 0.1428438425064087 0.9259605407714844 0.10941866040229797\n",
      "604 9.479292511780848 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.392581224282375 0.14249402284622192 0.9176741242408752 0.10886901617050171\n",
      "605 9.462220510816955 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.392407028353595 0.14224767684936523 0.9096811413764954 0.10820352286100388\n",
      "606 9.44428465762973 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.392253908010722 0.14204438030719757 0.9011024236679077 0.10778152197599411\n",
      "607 9.426943888939718 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.392137413896421 0.14166472852230072 0.8931384086608887 0.10686492919921875\n",
      "608 9.410929699726184 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3918298085404235 0.14145150780677795 0.8857775330543518 0.10609331727027893\n",
      "609 9.394554864944617 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.391580175937811 0.14116698503494263 0.8781154751777649 0.10557675361633301\n",
      "610 9.377938471536147 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3917043356103775 0.14088664948940277 0.8698945641517639 0.1055583581328392\n",
      "611 9.361371989078535 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3916248033714425 0.14067193865776062 0.862216591835022 0.10464206337928772\n",
      "612 9.34530399539099 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.391394456704715 0.14027711749076843 0.8548373579978943 0.10395770519971848\n",
      "613 9.329467138093829 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.391297226590038 0.1400303691625595 0.8472406268119812 0.1036582887172699\n",
      "614 9.313490041633512 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.391002723117735 0.13973698019981384 0.8398795127868652 0.10299131274223328\n",
      "615 9.297507566827601 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.390735348182983 0.13938699662685394 0.8324686288833618 0.10244796425104141\n",
      "616 9.28237875531605 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.390830464467316 0.138993039727211 0.8253480792045593 0.10185909271240234\n",
      "617 9.267149633249288 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.390772907516007 0.1387564241886139 0.8185550570487976 0.10051018744707108\n",
      "618 9.252962668977855 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3906658939944485 0.1384911984205246 0.8116845488548279 0.10043647885322571\n",
      "619 9.23790337049277 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.390600643091513 0.13827703893184662 0.8041194677352905 0.10078675299882889\n",
      "620 9.223730426874178 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.390518133189696 0.13802726566791534 0.7972152829170227 0.10075446218252182\n",
      "621 9.209419719486892 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.390279315024078 0.13778841495513916 0.7903816103935242 0.10058876872062683\n",
      "622 9.194549879595337 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.390053740307389 0.1374359279870987 0.7835136651992798 0.10003288090229034\n",
      "623 9.180552699902425 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.389909082391628 0.13717924058437347 0.7768657803535461 0.09973281621932983\n",
      "624 9.166850492945464 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3895797511847325 0.13698653876781464 0.7705637216567993 0.09915675967931747\n",
      "625 9.153684603688546 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.389422381934949 0.13666017353534698 0.7643187642097473 0.09896451979875565\n",
      "626 9.140354225606849 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.389170447559288 0.13643981516361237 0.7582641839981079 0.09821559488773346\n",
      "627 9.127532924166257 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.388834948292311 0.13614019751548767 0.752274751663208 0.09800827503204346\n",
      "628 9.114427318004559 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.388673235801648 0.13586528599262238 0.7459099292755127 0.09806893765926361\n",
      "629 9.101718215000103 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.388597598027657 0.13579319417476654 0.7399306297302246 0.09746616333723068\n",
      "630 9.089193730161488 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.388465120003522 0.13551118969917297 0.7340355515480042 0.09714631736278534\n",
      "631 9.076089843115497 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.388527295491387 0.13513177633285522 0.7279258966445923 0.09657897800207138\n",
      "632 9.063424833138965 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3884713053903806 0.1349080502986908 0.7220385074615479 0.09596846252679825\n",
      "633 9.05043684707002 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.388402357855601 0.13456811010837555 0.7164313793182373 0.09460362046957016\n",
      "634 9.038496393640076 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.388485301215683 0.13434390723705292 0.7110121250152588 0.0936429351568222\n",
      "635 9.026695256376705 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.388408889317951 0.13406144082546234 0.7056726813316345 0.09287956357002258\n",
      "636 9.015481358268119 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.388299463190891 0.13379459083080292 0.7004425525665283 0.09250219911336899\n",
      "637 9.004154778757286 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.388154141345216 0.1334393471479416 0.6953953504562378 0.09177058935165405\n",
      "638 8.992867454128568 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.387915081517999 0.13316041231155396 0.6902857422828674 0.09122047573328018\n",
      "639 8.981955807729621 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.387447875066657 0.13276483118534088 0.6855523586273193 0.09063838422298431\n",
      "640 8.970903849822713 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3873454617683425 0.13251571357250214 0.6802751421928406 0.09049239009618759\n",
      "641 8.960335699846809 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.387084087362354 0.13220369815826416 0.6755070686340332 0.09003377705812454\n",
      "642 8.950057026863503 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.387195800364422 0.13192161917686462 0.6705847978591919 0.08977001160383224\n",
      "643 8.93886161479561 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.387023157211229 0.13160353899002075 0.6653867959976196 0.0894613265991211\n",
      "644 8.928456512050639 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.387012545721541 0.13135264813899994 0.6605066657066345 0.08907798677682877\n",
      "645 8.91795377553432 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.386877307007003 0.13100464642047882 0.6558892130851746 0.0882933959364891\n",
      "646 8.907921977631679 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3867281976267005 0.13079789280891418 0.6513097286224365 0.08777642995119095\n",
      "647 8.897077757034074 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.38672319344021 0.1303955763578415 0.6464464068412781 0.0870661735534668\n",
      "648 8.887350270323415 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.386570798031945 0.13013224303722382 0.6417917013168335 0.08706382662057877\n",
      "649 8.877283647798594 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.386545077108439 0.1297847330570221 0.637398362159729 0.08615711331367493\n",
      "650 8.86773440600663 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.386476490509803 0.12938664853572845 0.6332758665084839 0.08531953394412994\n",
      "651 8.858138896605768 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.386408597311773 0.12905608117580414 0.6288784146308899 0.08491738885641098\n",
      "652 8.848526362237626 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.386439155695134 0.12889347970485687 0.6242343187332153 0.08472508937120438\n",
      "653 8.838619085878324 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.38638023558278 0.12862572073936462 0.6198018193244934 0.08400949090719223\n",
      "654 8.828828554278168 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3862878565247385 0.12837934494018555 0.6153603792190552 0.08344059437513351\n",
      "655 8.819185761188116 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.38618918267354 0.1281062364578247 0.6111910343170166 0.08250827342271805\n",
      "656 8.8093714174165 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3861253541244745 0.12785007059574127 0.6067445278167725 0.08190693706274033\n",
      "657 8.800242625602802 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.386036254891475 0.12752829492092133 0.6025534868240356 0.08157110214233398\n",
      "658 8.791602874739226 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3860497796965205 0.12729960680007935 0.5984551310539246 0.08134322613477707\n",
      "659 8.78335006137576 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.385885942477846 0.12696802616119385 0.5946422815322876 0.08121152967214584\n",
      "660 8.774204356196158 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.385598456087345 0.12679320573806763 0.5905231237411499 0.08076644688844681\n",
      "661 8.766047773452756 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.385308598669527 0.12647956609725952 0.5869150757789612 0.08042945712804794\n",
      "662 8.757549460934241 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.385137300895293 0.12609533965587616 0.5831181406974792 0.0800805389881134\n",
      "663 8.749879683975262 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.385080847565217 0.12588860094547272 0.5795339941978455 0.07984224706888199\n",
      "664 8.741637479067226 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3851244258884385 0.12573647499084473 0.5755806565284729 0.0796152651309967\n",
      "665 8.733267489337175 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.385104763769357 0.12532979249954224 0.5717626810073853 0.07930757105350494\n",
      "666 8.72519543053319 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.385099612490436 0.1250203549861908 0.5680670738220215 0.07894131541252136\n",
      "667 8.716714544913634 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.385005376062258 0.12476032972335815 0.5643281936645508 0.07829245179891586\n",
      "668 8.70919857398288 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.385037305212185 0.12464909255504608 0.5605878233909607 0.07833652943372726\n",
      "669 8.701684049905971 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.385083548368649 0.12430320680141449 0.5570710301399231 0.07815523445606232\n",
      "670 8.693555850111817 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.384985673748826 0.1240309625864029 0.553383469581604 0.07777227461338043\n",
      "671 8.685739409903608 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.384917233864389 0.12361544370651245 0.5499597787857056 0.07728717476129532\n",
      "672 8.67803188323873 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.384800530075011 0.12343358248472214 0.5468165874481201 0.0761645957827568\n",
      "673 8.670366447137571 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.384691875146605 0.1231335997581482 0.5433661937713623 0.075808584690094\n",
      "674 8.663023534497388 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.384685489615566 0.12278345227241516 0.5399041771888733 0.07574623823165894\n",
      "675 8.655924342275943 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.384789988101806 0.12254058569669724 0.5365654826164246 0.07546280324459076\n",
      "676 8.648108734185577 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.384762971098304 0.12227603048086166 0.5333536267280579 0.07436247915029526\n",
      "677 8.641345769919369 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.384725333012554 0.12194047123193741 0.5301986932754517 0.07428257912397385\n",
      "678 8.634095306308673 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3846979241685125 0.12178798764944077 0.5270704030990601 0.07346858829259872\n",
      "679 8.626504930746142 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.384546886276786 0.12143483757972717 0.5238003730773926 0.07292246073484421\n",
      "680 8.619926087554568 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.384500853713625 0.12115254253149033 0.5206704139709473 0.07293186336755753\n",
      "681 8.612827364524213 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.384334448893872 0.12100245803594589 0.517499566078186 0.07249132543802261\n",
      "682 8.6057117226132 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.384329834819659 0.12066449970006943 0.5143864154815674 0.07194455713033676\n",
      "683 8.598710213408992 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.384226482914016 0.12033458054065704 0.5112077593803406 0.07173363119363785\n",
      "684 8.59280136245021 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.384282477374163 0.12012899667024612 0.5081813931465149 0.07202710211277008\n",
      "685 8.58601626344751 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.384284313402878 0.11980165541172028 0.5051610469818115 0.07160820066928864\n",
      "686 8.579140252942597 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3841778631754265 0.11966709792613983 0.5021154880523682 0.07106431573629379\n",
      "687 8.572664891503788 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.384103302978016 0.11928673833608627 0.49913734197616577 0.07100016623735428\n",
      "688 8.566669817188094 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.384040604451011 0.11902408301830292 0.4962812066078186 0.07104271650314331\n",
      "689 8.560411445411894 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.384002975734923 0.11878496408462524 0.4934888482093811 0.07064580917358398\n",
      "690 8.554351826893535 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.384002005445209 0.11854587495326996 0.4907354712486267 0.07033300399780273\n",
      "691 8.548322959516323 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383993877742566 0.1183062493801117 0.4881229102611542 0.06977701187133789\n",
      "692 8.54175428059364 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3839601850178305 0.11803241074085236 0.48538580536842346 0.0689900740981102\n",
      "693 8.536131467554664 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383931379649734 0.1178741380572319 0.48259127140045166 0.06914340704679489\n",
      "694 8.53075739329558 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.38400092309218 0.11761802434921265 0.4799184799194336 0.06930148601531982\n",
      "695 8.52477489740899 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.38402981640151 0.11735762655735016 0.4771868586540222 0.06901373714208603\n",
      "696 8.518785195059033 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.384000101691933 0.11705867946147919 0.4744611084461212 0.06880419701337814\n",
      "697 8.512578897450567 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383891187880636 0.11674347519874573 0.47180357575416565 0.068337082862854\n",
      "698 8.505921785388805 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383832950745441 0.11650854349136353 0.4690675735473633 0.06744514405727386\n",
      "699 8.500337560042142 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383845907016992 0.1162622943520546 0.46647608280181885 0.06727719306945801\n",
      "700 8.494845526586897 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383789466987974 0.11597615480422974 0.464020699262619 0.06703850626945496\n",
      "701 8.489026006980485 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3837106942372985 0.11563027650117874 0.4617139399051666 0.06625715643167496\n",
      "702 8.484151665726007 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3836538611438876 0.11552491784095764 0.45918333530426025 0.06660621613264084\n",
      "703 8.478382531489213 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383636597479662 0.11518582701683044 0.45670005679130554 0.06615999341011047\n",
      "704 8.473256425583655 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3835709408166945 0.11493363976478577 0.45421162247657776 0.06632860004901886\n",
      "705 8.467513098763353 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383530094670183 0.11462285369634628 0.45165207982063293 0.06605599075555801\n",
      "706 8.462121692542919 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383568954710849 0.11447378247976303 0.44903799891471863 0.06600295752286911\n",
      "707 8.456683799210984 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383601553324658 0.11432395130395889 0.44668564200401306 0.06538701057434082\n",
      "708 8.450960112197375 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383659166676973 0.11411160230636597 0.4440503418445587 0.0650886595249176\n",
      "709 8.446229448020926 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383604211211672 0.11386847496032715 0.4415464699268341 0.06566382199525833\n",
      "710 8.440567520558563 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383591895103184 0.11364960670471191 0.43914416432380676 0.06503769010305405\n",
      "711 8.435083297891893 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383497623606004 0.11343716830015182 0.43693628907203674 0.06427592784166336\n",
      "712 8.430143349964336 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383391619820312 0.11325114965438843 0.43460777401924133 0.0642850324511528\n",
      "713 8.424802043093358 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383335665417825 0.1131705641746521 0.4323537349700928 0.06358834356069565\n",
      "714 8.41999863898514 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383214444575293 0.11294367164373398 0.43011239171028137 0.06361573934555054\n",
      "715 8.415533835936918 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383184741188421 0.11269446462392807 0.42811593413352966 0.06342276185750961\n",
      "716 8.410661222398234 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383237222194625 0.11250965297222137 0.4258354604244232 0.06324342638254166\n",
      "717 8.4058629490451 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383221044392981 0.11227061599493027 0.42355281114578247 0.06326566636562347\n",
      "718 8.400804723462498 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3832904499382135 0.11204686015844345 0.4212247133255005 0.06301798671483994\n",
      "719 8.395595559680263 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3833885954854885 0.11186980456113815 0.41870632767677307 0.06292450428009033\n",
      "720 8.390728952847827 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383333797119964 0.11168611794710159 0.4164944887161255 0.06272006034851074\n",
      "721 8.3861509814793 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383234861486566 0.1114128902554512 0.414357453584671 0.06278832256793976\n",
      "722 8.381530436684436 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383421215226001 0.11117792874574661 0.41219645738601685 0.06253837794065475\n",
      "723 8.376357167331797 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383463322488886 0.11100783199071884 0.4098612070083618 0.06216359883546829\n",
      "724 8.372176144159294 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383388083315349 0.11070296913385391 0.40797096490859985 0.06214316189289093\n",
      "725 8.36784440560574 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.38332209645147 0.11041875183582306 0.4060956835746765 0.06191219016909599\n",
      "726 8.36309318123776 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383214733287874 0.11017227172851562 0.4040347933769226 0.06163658946752548\n",
      "727 8.358957204722095 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383256584160257 0.11002370715141296 0.4020123779773712 0.061652157455682755\n",
      "728 8.35444224282514 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383293159598221 0.10975487530231476 0.399783730506897 0.06182674691081047\n",
      "729 8.34974430956742 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383265868656841 0.10947536677122116 0.3979265093803406 0.06115005537867546\n",
      "730 8.345614568702217 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383099113962885 0.10922056436538696 0.3961210548877716 0.06105278059840202\n",
      "731 8.3410508257469 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383074380191212 0.10896892845630646 0.3941936194896698 0.060620278120040894\n",
      "732 8.336954996976912 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383131560925305 0.10889540612697601 0.3920246958732605 0.06087863817811012\n",
      "733 8.332858860891335 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383164726596348 0.10871629416942596 0.39000603556632996 0.06096576899290085\n",
      "734 8.328566025956674 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383206073208852 0.10850752145051956 0.3879531919956207 0.06094604730606079\n",
      "735 8.324495984079421 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383078506054462 0.108278788626194 0.3860946297645569 0.060949429869651794\n",
      "736 8.320098594346714 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.383006468930913 0.10801100730895996 0.38428282737731934 0.06051546335220337\n",
      "737 8.315350744760027 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382982093727579 0.10774822533130646 0.38233277201652527 0.05995488166809082\n",
      "738 8.311111133717832 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382866341792878 0.10752864181995392 0.38050785660743713 0.059700436890125275\n",
      "739 8.307029546633833 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.38280796392309 0.10726568847894669 0.3785167932510376 0.05992230772972107\n",
      "740 8.302886448142011 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3828537196786 0.10717269033193588 0.37651413679122925 0.05983176454901695\n",
      "741 8.298553945089251 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3828303418704575 0.10690552741289139 0.37471237778663635 0.059393320232629776\n",
      "742 8.29464314381104 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3829385774813225 0.1067160964012146 0.372862845659256 0.059262778609991074\n",
      "743 8.290667330978785 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3828956802770245 0.10649633407592773 0.37099823355674744 0.05927884951233864\n",
      "744 8.28641919911015 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3829495099197615 0.10626604408025742 0.3692632019519806 0.05867724120616913\n",
      "745 8.282758486115574 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382805776440739 0.10599946975708008 0.3676762878894806 0.058600664138793945\n",
      "746 8.279091050247732 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382826493452374 0.10579854249954224 0.36592403054237366 0.05861795321106911\n",
      "747 8.275502304592464 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382833852478599 0.10553539544343948 0.36432603001594543 0.05848099663853645\n",
      "748 8.271551499489943 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382871693346898 0.10526439547538757 0.3625817894935608 0.05825183168053627\n",
      "749 8.267025395111137 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382886043564373 0.1049410030245781 0.3607827425003052 0.05763286352157593\n",
      "750 8.263178028857912 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382875331437791 0.10471416264772415 0.35910171270370483 0.05738510936498642\n",
      "751 8.259534838070246 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382886818070981 0.10453929007053375 0.3573707938194275 0.05736714228987694\n",
      "752 8.256064028798827 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382900712399968 0.1042906641960144 0.3556947708129883 0.05748311057686806\n",
      "753 8.252172187278326 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382876620004232 0.10406073927879333 0.35403671860694885 0.05716139078140259\n",
      "754 8.248339738695798 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.38289097323912 0.10398118942975998 0.352310448884964 0.056846678256988525\n",
      "755 8.244640563552325 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382910956566279 0.10374366492033005 0.35062283277511597 0.056740276515483856\n",
      "756 8.241296213117788 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382872842249821 0.10344834625720978 0.34933361411094666 0.05630779638886452\n",
      "757 8.237986454296404 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382871700007016 0.10317537933588028 0.347514271736145 0.05691083148121834\n",
      "758 8.233946791783326 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382828242078775 0.10294727981090546 0.34598231315612793 0.05620664358139038\n",
      "759 8.230247700195605 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382837107013757 0.10276250541210175 0.3443758487701416 0.05589639022946358\n",
      "760 8.22712752558342 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382903802767906 0.10261471569538116 0.3426930606365204 0.056222885847091675\n",
      "761 8.223679888247549 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.38279443904668 0.10237857699394226 0.3410731852054596 0.056360501796007156\n",
      "762 8.22006666419017 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382729870648262 0.1021629050374031 0.33966565132141113 0.05584258586168289\n",
      "763 8.216619453146397 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382712105973897 0.10194040089845657 0.3382106423377991 0.055545661598443985\n",
      "764 8.21340865961147 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382657925997503 0.10160348564386368 0.33673298358917236 0.05568128079175949\n",
      "765 8.209631082800712 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3826512184120086 0.10140848904848099 0.3352663516998291 0.05503867194056511\n",
      "766 8.206242310300514 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3826698683649745 0.10117299854755402 0.33374524116516113 0.054908961057662964\n",
      "767 8.202966172359142 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382639534882459 0.10094733536243439 0.33219245076179504 0.05499440059065819\n",
      "768 8.199804192930532 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.38264752045281 0.10073593258857727 0.3305511474609375 0.0553184449672699\n",
      "769 8.196614862113455 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.38264421436117 0.10048583894968033 0.32910943031311035 0.05526594817638397\n",
      "770 8.193111554202048 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382653765794245 0.10038859397172928 0.32759734988212585 0.054874494671821594\n",
      "771 8.189408629736196 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382634605815898 0.10011138021945953 0.32610052824020386 0.054461587220430374\n",
      "772 8.18592655152591 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382661866550762 0.09988659620285034 0.3243798315525055 0.05461842566728592\n",
      "773 8.18273688130789 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382591881088545 0.09957241266965866 0.3229700028896332 0.054632581770420074\n",
      "774 8.1791161648929 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382636760920274 0.09938281029462814 0.32143378257751465 0.05422902852296829\n",
      "775 8.176093967088129 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3825492760321145 0.09923385828733444 0.32012543082237244 0.05405997112393379\n",
      "776 8.172618597106045 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382545802102392 0.09905514866113663 0.3189350962638855 0.05314745381474495\n",
      "777 8.16993574279982 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.3824729755862375 0.09895100444555283 0.31766295433044434 0.053185854107141495\n",
      "778 8.167015301869121 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382535870061126 0.09883774816989899 0.31621861457824707 0.053204454481601715\n",
      "779 8.163955398376494 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382570367779046 0.0985066220164299 0.3148667812347412 0.053144846111536026\n",
      "780 8.160863855520674 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382565604189821 0.09836019575595856 0.31348851323127747 0.05296102911233902\n",
      "781 8.158150744680666 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382691812757754 0.09825211763381958 0.3120177686214447 0.05317127704620361\n",
      "782 8.154839074849189 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382719876854718 0.0980873852968216 0.31061458587646484 0.052802640944719315\n",
      "783 8.152085377341248 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382657064562776 0.09781774878501892 0.30946141481399536 0.05268773436546326\n",
      "784 8.14874769716459 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382608078112179 0.09751973301172256 0.30812567472457886 0.052368536591529846\n",
      "785 8.145858095561055 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382565276299504 0.0974116250872612 0.30671948194503784 0.05244223028421402\n",
      "786 8.143010155517228 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382528071511157 0.09722430258989334 0.30531707406044006 0.05262363329529762\n",
      "787 8.140684693616954 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382549687785236 0.0970953106880188 0.3040335178375244 0.05297265946865082\n",
      "788 8.137565836457027 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382585149405015 0.09689892083406448 0.3027953803539276 0.05249100551009178\n",
      "789 8.13522816754882 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382660451605961 0.09663577377796173 0.3016570508480072 0.052617840468883514\n",
      "790 8.132430876273997 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382668385882265 0.09642167389392853 0.3003929853439331 0.05255484580993652\n",
      "791 8.129398749887367 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382577800869365 0.09616763144731522 0.29921379685401917 0.05222572386264801\n",
      "792 8.126966792577416 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.38248160057655 0.09611289948225021 0.29796046018600464 0.052451372146606445\n",
      "793 8.124003263348222 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382537873619675 0.09591799974441528 0.29668015241622925 0.05218708515167236\n",
      "794 8.121179905227514 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382473964682909 0.09579338133335114 0.2954672873020172 0.051977984607219696\n",
      "795 8.118519818515058 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382487135112757 0.09546348452568054 0.2942025065422058 0.05216418579220772\n",
      "796 8.115737467964593 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382479812910261 0.09527140855789185 0.293000191450119 0.05198586359620094\n",
      "797 8.113012769322241 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382488736014212 0.09494233131408691 0.29193800687789917 0.05170568823814392\n",
      "798 8.109868756108098 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382441966407113 0.09479989111423492 0.29071658849716187 0.0511937215924263\n",
      "799 8.107140129348068 tensor(2, device='cuda:4') tensor(5, device='cuda:4') 7.382475389083652 0.09448892623186111 0.2896248698234558 0.050926074385643005\n"
     ]
    }
   ],
   "source": [
    "# Get real, random, and fake datasets\n",
    "N_SAMPLES = 50\n",
    "\n",
    "#[0'AFR' 1'AHG' 2'AMR' 3'EAS' 4'EUR' 5'OCE' 6'SAS' 7'WAS']\n",
    "target_label_pca = 4\n",
    "target_label_lainet = 2\n",
    "target_label_nadm = 5\n",
    "target_knn_label = 1\n",
    "num_iters = 800\n",
    "fake_sequence = generate_adversarial_population(N_SAMPLES, snps.shape[1], snps, labels, (W,b), target_label_pca, LN, target_label_lainet, knn_W, target_knn_label, NADM, target_label_nadm, device, num_iters = num_iters)\n",
    "\n",
    "random_sequence = torch.bernoulli(torch.mean(torch.tensor(snps).float(), dim=0, keepdims=True).repeat(N_SAMPLES,1)).float().to(device)\n",
    "real_sequence = torch.tensor(snps_test[labels_test==4,:]).float()[0:N_SAMPLES,:].to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Store datasets\n",
    "dict_to_save = {\n",
    "'name' : 'Manipulated',\n",
    "'real_seqs' : snps,\n",
    "'fake_seqs' : fake_sequence.cpu().numpy()\n",
    "}\n",
    "\n",
    "with open(output_folder_prefix+f'dataset_manipulated.pkl', 'wb') as handle:\n",
    "    pkl.dump(dict_to_save, handle, protocol=pkl.HIGHEST_PROTOCOL)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Perform inference with PCA, KNN, LAI-Net, and Neural ADMIXTURE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "###############################################################\n",
    "## Perform Inference ##########################################\n",
    "###############################################################\n",
    "\n",
    "## Real sequence ##############################################\n",
    "## PCA\n",
    "pcs_real, _, _, _ = compute_pca(real_sequence.cpu(), W, b)\n",
    "\n",
    "## KNN\n",
    "pred_real_knn = KNN.predict(np.matmul(real_sequence.cpu().numpy(), knn_W.cpu().numpy()))\n",
    "unique, count = np.unique(pred_real_knn, return_counts=True)\n",
    "real_eur_knn = unique_to_percentage(unique, count, 8)\n",
    "\n",
    "## LAI-Net\n",
    "_, pred_real = forward_haploid(LN.eval(), real_sequence)\n",
    "unique, count = torch.unique(pred_real, return_counts=True)\n",
    "real_eur_lainet = unique_to_percentage(unique.cpu().numpy(), count.cpu().numpy(), 8)\n",
    "\n",
    "## Neural ADMIXTURE\n",
    "_, pred_nadm_real = forward_nadm(NADM, real_sequence)\n",
    "unique, count = torch.unique(pred_nadm_real, return_counts=True)\n",
    "real_eur_nadm = unique_to_percentage(unique.cpu().numpy(), count.cpu().numpy(), 8)\n",
    "\n",
    "\n",
    "## Random sequence ###########################################\n",
    "## PCA\n",
    "pcs_random, _, _, _ = compute_pca(random_sequence.cpu(), W, b)\n",
    "\n",
    "## KNN\n",
    "pred_random_knn = KNN.predict(np.matmul(random_sequence.cpu().numpy(), knn_W.cpu().numpy()))\n",
    "unique, count = np.unique(pred_random_knn, return_counts=True)\n",
    "random_knn = unique_to_percentage(unique, count, 8)\n",
    "\n",
    "## LAI-Net\n",
    "_, pred_random = forward_haploid(LN.eval(), random_sequence)\n",
    "unique, count = torch.unique(pred_random, return_counts=True)\n",
    "random_lainet = unique_to_percentage(unique.cpu().numpy(), count.cpu().numpy(), 8)\n",
    "\n",
    "## Neural ADMIXTURE\n",
    "_, pred_nadm_random = forward_nadm(NADM, random_sequence)\n",
    "unique, count = torch.unique(pred_nadm_random, return_counts=True)\n",
    "random_nadm = unique_to_percentage(unique.cpu().numpy(), count.cpu().numpy(), 8)\n",
    "\n",
    "\n",
    "## Adversarial sequence ###########################################\n",
    "## PCA\n",
    "pcs_fake, _, _, _ = compute_pca(fake_sequence.cpu(), W, b)\n",
    "\n",
    "## KNN\n",
    "pred_random_knn = KNN.predict(np.matmul(fake_sequence.cpu().numpy(), knn_W.cpu().numpy()))\n",
    "unique, count = np.unique(pred_random_knn, return_counts=True)\n",
    "fake_knn = unique_to_percentage(unique, count, 8)\n",
    "\n",
    "## LAI-Net\n",
    "_, pred_random = forward_haploid(LN.eval(), fake_sequence)\n",
    "unique, count = torch.unique(pred_random, return_counts=True)\n",
    "fake_lainet = unique_to_percentage(unique.cpu().numpy(), count.cpu().numpy(), 8)\n",
    "\n",
    "## Neural ADMIXTURE\n",
    "_, pred_nadm_random = forward_nadm(NADM, fake_sequence)\n",
    "unique, count = torch.unique(pred_nadm_random, return_counts=True)\n",
    "fake_nadm = unique_to_percentage(unique.cpu().numpy(), count.cpu().numpy(), 8)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Make Final Plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "<ipython-input-16-b99ceb9ae421>:99: UserWarning: FixedFormatter should only be used together with FixedLocator\n",
      "  axd[k].set_xticklabels(ancestry_names, rotation = 45)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 10 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from matplotlib import cm\n",
    "from matplotlib.patches import Patch\n",
    "from matplotlib.lines import Line2D\n",
    "\n",
    "\n",
    "cmap = cm.get_cmap('Dark2', 8) #'magma' #'plasma' #'cividis'\n",
    "color_list = cmap([j for j in range(8)])\n",
    "\n",
    "axd = plt.figure(constrained_layout=True, figsize=(15, 6)).subplot_mosaic(\n",
    "    \"\"\"\n",
    "    AAABCD\n",
    "    AAAEFG\n",
    "    AAAHIJ\n",
    "    \"\"\"\n",
    ")\n",
    "\n",
    "axd['A'].scatter(pcat[:,0], pcat[:,1], c=labels, marker='.', s=2.0, cmap=cmap)\n",
    "axd['A'].scatter(pcs_fake.numpy()[:,0], pcs_fake.numpy()[:,1], c='red', marker='x', s=2.5, cmap=cmap)\n",
    "axd['A'].scatter(pcs_random.numpy()[:,0], pcs_random.numpy()[:,1], c='black', marker='x', s=2.5, cmap=cmap)\n",
    "axd['A'].scatter(pcs_real.numpy()[:,0], pcs_real.numpy()[:,1], c='green', marker='x', s=2.5, cmap=cmap)\n",
    "\n",
    "\n",
    "real_edge_color = 'g'\n",
    "random_edge_color = 'black'\n",
    "adversarial_edge_color = 'r'\n",
    "\n",
    "\n",
    "linewidth=3\n",
    "\n",
    "axd['B'].bar(ancestry_names, real_eur_knn, edgecolor=real_edge_color, color=color_list, linewidth=linewidth)\n",
    "axd['C'].bar(ancestry_names, real_eur_lainet, edgecolor=real_edge_color, color=color_list, linewidth=linewidth)\n",
    "axd['D'].bar(ancestry_names, real_eur_nadm, edgecolor=real_edge_color, color=color_list, linewidth=linewidth)\n",
    "\n",
    "axd['E'].bar(ancestry_names, random_knn, edgecolor=random_edge_color, color=color_list, linewidth=linewidth)\n",
    "axd['F'].bar(ancestry_names, random_lainet, edgecolor=random_edge_color, color=color_list, linewidth=linewidth)\n",
    "axd['G'].bar(ancestry_names, random_nadm, edgecolor=random_edge_color, color=color_list, linewidth=linewidth)\n",
    "\n",
    "\n",
    "axd['H'].bar(ancestry_names, fake_knn, edgecolor=adversarial_edge_color, color=color_list, linewidth=linewidth)\n",
    "axd['I'].bar(ancestry_names, fake_lainet, edgecolor=adversarial_edge_color, color=color_list, linewidth=linewidth)\n",
    "axd['J'].bar(ancestry_names, fake_nadm, edgecolor=adversarial_edge_color, color=color_list, linewidth=linewidth)\n",
    "\n",
    "\n",
    "\n",
    "\"\"\"\n",
    "AAABCD\n",
    "AAAEFG\n",
    "AAAHIJ\n",
    "\"\"\"\n",
    "\n",
    "fsize_label_axis = 12\n",
    "\n",
    "axd['A'].set_xlabel(f'Adv. Target: {ancestry_names[target_label_pca]}', fontsize=fsize_label_axis)\n",
    "axd['A'].xaxis.label.set_color(color_list[target_label_pca])\n",
    "axd['H'].set_xlabel(f'Adv. Target: {ancestry_names[target_knn_label]}', fontsize=fsize_label_axis)\n",
    "axd['H'].xaxis.label.set_color(color_list[target_knn_label])\n",
    "axd['I'].set_xlabel(f'Adv. Target: {ancestry_names[target_label_lainet]}', fontsize=fsize_label_axis)\n",
    "axd['I'].xaxis.label.set_color(color_list[target_label_lainet])\n",
    "axd['J'].set_xlabel(f'Adv. Target: {ancestry_names[target_label_nadm]}', fontsize=fsize_label_axis)\n",
    "axd['J'].xaxis.label.set_color(color_list[target_label_nadm])\n",
    "    \n",
    "\n",
    "    \n",
    "axd['G'].set_ylabel('Histogram of predictions', fontsize=fsize_label_axis)\n",
    "axd['G'].yaxis.set_label_position('right') \n",
    "#axd['G'].yaxis.label_right()\n",
    "    \n",
    "for k in ['B', 'E', 'H', 'C', 'F', 'I']:\n",
    "    axd[k].spines['right'].set_visible(False)\n",
    "    axd[k].spines['top'].set_visible(False)\n",
    "    axd[k].spines['left'].set_visible(False)\n",
    "    axd[k].set_yticks([])\n",
    "    \n",
    "for k in ['D',  'G', 'J']:\n",
    "    #axd[k].spines['right'].set_visible(False)\n",
    "    axd[k].spines['top'].set_visible(False)\n",
    "    axd[k].spines['left'].set_visible(False)\n",
    "    #axd[k].set_yticks([])\n",
    "    \n",
    "for k in ['A']:\n",
    "    axd[k].spines['right'].set_visible(False)\n",
    "    axd[k].spines['top'].set_visible(False)\n",
    "    \n",
    "    \n",
    "for k in ['B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J']:\n",
    "    axd[k].yaxis.tick_right()\n",
    "\n",
    "\n",
    "axd['B'].set_ylabel('EUR Sequence', fontsize=fsize_label_axis)\n",
    "axd['B'].yaxis.label.set_color('green')\n",
    "axd['E'].set_ylabel('Random Sequence', fontsize=fsize_label_axis)\n",
    "axd['H'].set_ylabel('Adversarial Sequence', fontsize=fsize_label_axis)\n",
    "axd['H'].yaxis.label.set_color('red')\n",
    "\n",
    "\n",
    "\n",
    "for k in ['B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J']:\n",
    "    axd[k].set_ylim([0.0,1.0])\n",
    "    axd[k].set_xticklabels(ancestry_names, rotation = 45)\n",
    "    \n",
    "for k, t in zip(['A', 'B', 'C', 'D'], ['PCA', 'Fast k-NN', 'LAI-Net', 'Neural ADMIXTURE']):\n",
    "    axd[k].set_title(t, fontsize=14)\n",
    "\n",
    "    \n",
    "\n",
    "\n",
    "    \n",
    "legend_elements = [Line2D([0], [0], color='b', lw=4, label='Line'),\n",
    "                   Line2D([0], [0], marker='o', color='w', label='Scatter',\n",
    "                          markerfacecolor='g', markersize=15),\n",
    "                   Patch(facecolor='orange', edgecolor='r',\n",
    "                         label='Color Patch')]\n",
    "\n",
    "\n",
    "#[0'AFR' 1'AHG' 2'AMR' 3'EAS' 4'EUR' 5'OCE' 6'SAS' 7'WAS']\n",
    "markersize = 7\n",
    "legend_elements = [\n",
    "                    Line2D([0], [0], marker='o', color='w', label='African (AFR)',markerfacecolor=color_list[0], markersize=markersize),\n",
    "                    Line2D([0], [0], marker='o', color='w', label='African Hunter-Gatherer (AHG)',markerfacecolor=color_list[1], markersize=markersize),\n",
    "                    Line2D([0], [0], marker='o', color='w', label='Native American (AMR)',markerfacecolor=color_list[2], markersize=markersize),\n",
    "                    Line2D([0], [0], marker='o', color='w', label='East Asian (EAS)',markerfacecolor=color_list[3], markersize=markersize),\n",
    "                    Line2D([0], [0], marker='o', color='w', label='European (EUR)',markerfacecolor=color_list[4], markersize=markersize),\n",
    "                    Line2D([0], [0], marker='o', color='w', label='Oceanian (OCE)',markerfacecolor=color_list[5], markersize=markersize),\n",
    "                    Line2D([0], [0], marker='o', color='w', label='South Asian (SAS)',markerfacecolor=color_list[6], markersize=markersize),\n",
    "                    Line2D([0], [0], marker='o', color='w', label='West Asian (WAS)',markerfacecolor=color_list[7], markersize=markersize),\n",
    "                    Line2D([0], [0], marker='s', color='w', label='Random Sequences',markerfacecolor='black', markersize=markersize),\n",
    "                    Line2D([0], [0], marker='s', color='w', label='Adversarial Sequences',markerfacecolor='r', markersize=markersize),\n",
    "                    Line2D([0], [0], marker='s', color='w', label='Real Sequences (Test EUR)',markerfacecolor='green', markersize=markersize),\n",
    "]\n",
    "\n",
    "\n",
    "axd['A'].legend(handles=legend_elements, loc='upper right', fontsize=12, frameon=False)\n",
    "\n",
    "\n",
    "plt.savefig('figs/fake_dataset.png')\n",
    "plt.show()"
   ]
  }
 ],
 "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}