Experiments_CIFAR10_IID.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "#================================= Start of importing required packages and libraries =========================================#\n",
    "from __future__ import print_function\n",
    "%matplotlib inline\n",
    "import numpy as np\n",
    "import torch\n",
    "from experiment_federated import *\n",
    "import random\n",
    "#================================== End of importing required packages and libraries ==========================================#"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "#=============================== Defining global variables ========================#\n",
    "DATASET_NAME = \"CIFAR10\"\n",
    "MODEL_NAME = \"ResNet18\"\n",
    "DD_TYPE = 'IID'\n",
    "ALPHA = 1000000\n",
    "NUM_PEERS = 20 # \"number of peers: K\" \n",
    "FRAC_PEERS = 1 #'the fraction of peers: C to bel selected in each round'\n",
    "SEED = 7 #fixed seed\n",
    "random.seed(SEED)\n",
    "CRITERION = nn.CrossEntropyLoss()\n",
    "GLOBAL_ROUNDS = 200 #\"number of rounds of federated model training\"\n",
    "LOCAL_EPOCHS = 3 #\"the number of local epochs: E for each peer\"\n",
    "TEST_BATCH_SIZE = 1000\n",
    "LOCAL_BS = 32 #\"local batch size: B for each peer\"\n",
    "LOCAL_LR =  0.01#local learning rate: lr for each peer\n",
    "LOCAL_MOMENTUM = 0.9 #local momentum for each peer\n",
    "NUM_CLASSES = 10 # number of classes in an experiment\n",
    "LABELS_DICT = {'Plane':0,\n",
    "               'Car':1, \n",
    "               'Bird':2, \n",
    "               'Cat':3, \n",
    "               'Deer':4,\n",
    "               'Dog':5, \n",
    "               'Frog':6, \n",
    "               'Horse':7, \n",
    "               'Ship':8, \n",
    "               'Truck':9}\n",
    "\n",
    "#select the device to work with cpu or gpu\n",
    "if torch.cuda.is_available():\n",
    "    DEVICE = \"cuda\"\n",
    "else:\n",
    "    DEVICE = \"cpu\"\n",
    "DEVICE = torch.device(DEVICE)\n",
    "SOURCE_CLASS = 5 # the source class\n",
    "TARGET_CLASS = 3 # the target class \n",
    "\n",
    "CLASS_PER_PEER = 1\n",
    "SAMPLES_PER_CLASS = 582\n",
    "RATE_UNBALANCE = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "--> Starting experiment...\n",
      "--> Loading of CIFAR10 dataset\n",
      "Files already downloaded and verified\n",
      "Files already downloaded and verified\n",
      "--> Dataset has been loaded!\n",
      "--> Creating ResNet18 model.....\n",
      "--> Model has been created!\n",
      "--> Distributing training data among peers\n",
      "--> Training data have been distributed among peers\n",
      "--> Creating peets instances\n",
      "Data set: CIFAR10\n",
      "Data distribution: IID\n",
      "Aggregation rule: ppfl_defender\n",
      "Attack Type: label_flipping\n",
      "Attackers Ratio: 0.0 %\n",
      "Malicious Behavior Rate: 100 %\n",
      "\n",
      "===>Simulation started...\n",
      "\n",
      "====>Global model training started...\n",
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2460ac630a5740fbac002a3880e7e235",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=200.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " | Global training round : 1/200 |\n",
      "\n",
      "training time\n",
      "127.09251427650452\n",
      "training time\n",
      "126.76475596427917\n",
      "training time\n",
      "125.51689076423645\n",
      "training time\n",
      "124.24494528770447\n",
      "training time\n",
      "123.68282866477966\n",
      "training time\n",
      "123.73558783531189\n",
      "training time\n",
      "123.489328622818\n",
      "training time\n",
      "123.54176211357117\n",
      "training time\n",
      "123.7128381729126\n",
      "training time\n",
      "123.6275851726532\n",
      "training time\n",
      "125.58847951889038\n",
      "training time\n",
      "124.58264327049255\n",
      "training time\n",
      "144.5197286605835\n",
      "training time\n",
      "131.0119185447693\n",
      "training time\n",
      "132.2419149875641\n",
      "training time\n",
      "132.46862292289734\n",
      "training time\n",
      "130.08744955062866\n",
      "training time\n",
      "128.65416550636292\n",
      "training time\n",
      "125.25802373886108\n",
      "training time\n",
      "132.98094582557678\n",
      "Average of peers' local losses: 2.117753\n",
      "aggregate time:\n",
      "1.3838305473327637\n",
      "\n",
      "Average test loss: 2.3234, Test accuracy: 1000/10000 (10.00%)\n",
      "\n",
      "Class      - Accuracy\n",
      "Plane      - 0.0\n",
      "Car        - 0.0\n",
      "Bird       - 100.0\n",
      "Cat        - 0.0\n",
      "Deer       - 0.0\n",
      "Dog        - 0.0\n",
      "Frog       - 0.0\n",
      "Horse      - 0.0\n",
      "Ship       - 0.0\n",
      "Truck      - 0.0\n",
      "\n",
      " | Global training round : 2/200 |\n",
      "\n",
      "training time\n",
      "163.17411065101624\n",
      "training time\n",
      "133.58422827720642\n",
      "training time\n",
      "152.6824357509613\n",
      "training time\n",
      "126.98082900047302\n",
      "training time\n",
      "131.99769616127014\n",
      "training time\n",
      "134.95468974113464\n",
      "training time\n",
      "136.05385994911194\n",
      "training time\n",
      "126.43527173995972\n",
      "training time\n",
      "126.58537006378174\n",
      "training time\n",
      "124.72307515144348\n",
      "training time\n",
      "124.5506272315979\n",
      "training time\n",
      "124.76152467727661\n",
      "training time\n",
      "133.29872298240662\n",
      "training time\n",
      "132.40166211128235\n",
      "training time\n",
      "128.37853121757507\n",
      "training time\n",
      "125.28007197380066\n",
      "training time\n",
      "130.41232872009277\n",
      "training time\n",
      "125.90780258178711\n",
      "training time\n",
      "127.20479798316956\n",
      "training time\n",
      "126.02269077301025\n",
      "Average of peers' local losses: 1.919380\n",
      "aggregate time:\n",
      "2.490140676498413\n",
      "\n",
      "Average test loss: 2.2257, Test accuracy: 1255/10000 (12.55%)\n",
      "\n",
      "Class      - Accuracy\n",
      "Plane      - 0.0\n",
      "Car        - 0.0\n",
      "Bird       - 96.3\n",
      "Cat        - 0.8\n",
      "Deer       - 0.3\n",
      "Dog        - 0.0\n",
      "Frog       - 0.0\n",
      "Horse      - 0.0\n",
      "Ship       - 28.1\n",
      "Truck      - 0.0\n",
      "\n",
      " | Global training round : 3/200 |\n",
      "\n",
      "training time\n",
      "127.80422592163086\n",
      "training time\n",
      "124.33146929740906\n",
      "training time\n",
      "123.99117636680603\n",
      "training time\n",
      "125.67293453216553\n",
      "training time\n",
      "132.0570216178894\n",
      "training time\n",
      "166.01767110824585\n",
      "training time\n",
      "126.01841306686401\n",
      "training time\n",
      "124.6076431274414\n",
      "training time\n",
      "130.29550313949585\n",
      "training time\n",
      "131.29522109031677\n",
      "training time\n",
      "134.1926121711731\n",
      "training time\n",
      "125.82759189605713\n",
      "training time\n",
      "135.60902285575867\n",
      "training time\n",
      "124.44861936569214\n",
      "training time\n",
      "125.02011775970459\n",
      "training time\n",
      "123.73866868019104\n",
      "training time\n",
      "134.37365984916687\n",
      "training time\n",
      "128.94308519363403\n",
      "training time\n",
      "124.73362994194031\n",
      "training time\n",
      "130.69545722007751\n",
      "Average of peers' local losses: 1.695829\n",
      "aggregate time:\n",
      "1.1596684455871582\n",
      "\n",
      "Average test loss: 1.8265, Test accuracy: 3484/10000 (34.84%)\n",
      "\n",
      "Class      - Accuracy\n",
      "Plane      - 44.5\n",
      "Car        - 15.4\n",
      "Bird       - 37.6\n",
      "Cat        - 26.6\n",
      "Deer       - 58.3\n",
      "Dog        - 5.1\n",
      "Frog       - 64.7\n",
      "Horse      - 29.1\n",
      "Ship       - 51.4\n",
      "Truck      - 15.7\n",
      "\n",
      " | Global training round : 4/200 |\n",
      "\n",
      "training time\n",
      "125.02136421203613\n",
      "training time\n",
      "124.19819140434265\n",
      "training time\n",
      "124.04437947273254\n",
      "training time\n",
      "131.87213373184204\n",
      "training time\n",
      "124.12812876701355\n",
      "training time\n",
      "130.6947340965271\n",
      "training time\n",
      "137.19402408599854\n",
      "training time\n",
      "153.73380374908447\n",
      "training time\n",
      "147.2568974494934\n",
      "training time\n",
      "150.13412928581238\n",
      "training time\n",
      "142.52850127220154\n",
      "training time\n",
      "132.2867820262909\n",
      "training time\n",
      "131.19084191322327\n",
      "training time\n",
      "133.67717504501343\n",
      "training time\n",
      "166.82329511642456\n",
      "training time\n",
      "135.6666705608368\n",
      "training time\n",
      "146.33179831504822\n",
      "training time\n",
      "128.064311504364\n",
      "training time\n",
      "152.10884618759155\n",
      "training time\n",
      "165.70497250556946\n",
      "Average of peers' local losses: 1.450988\n",
      "aggregate time:\n",
      "9.075217723846436\n",
      "\n",
      "Average test loss: 1.3419, Test accuracy: 5179/10000 (51.79%)\n",
      "\n",
      "Class      - Accuracy\n",
      "Plane      - 44.8\n",
      "Car        - 54.6\n",
      "Bird       - 34.6\n",
      "Cat        - 42.6\n",
      "Deer       - 57.2\n",
      "Dog        - 29.7\n",
      "Frog       - 71.6\n",
      "Horse      - 55.4\n",
      "Ship       - 71.6\n",
      "Truck      - 55.8\n",
      "\n",
      " | Global training round : 5/200 |\n",
      "\n",
      "training time\n",
      "175.97275567054749\n",
      "training time\n",
      "153.55581378936768\n",
      "training time\n",
      "135.20151209831238\n",
      "training time\n",
      "124.51229047775269\n",
      "training time\n",
      "123.35790634155273\n"
     ]
    }
   ],
   "source": [
    "RULE = 'ppfl_defender'\n",
    "ATTACK_TYPE='label_flipping'\n",
    "MALICIOUS_BEHAVIOR_RATE = 1\n",
    "for atr in [0.0, 0.1, 0.2, 0.3, 0.4, 0.5]:\n",
    "    run_exp(dataset_name = DATASET_NAME, model_name = MODEL_NAME, dd_type = DD_TYPE, num_peers = NUM_PEERS, \n",
    "            frac_peers = FRAC_PEERS, seed = SEED, test_batch_size = TEST_BATCH_SIZE,\n",
    "                criterion = CRITERION, global_rounds = GLOBAL_ROUNDS, local_epochs = LOCAL_EPOCHS, local_bs = LOCAL_BS, \n",
    "                 local_lr = LOCAL_LR, local_momentum = LOCAL_MOMENTUM, labels_dict = LABELS_DICT, device = DEVICE,\n",
    "                attackers_ratio = atr, attack_type=ATTACK_TYPE, \n",
    "                 malicious_behavior_rate = MALICIOUS_BEHAVIOR_RATE, rule = RULE,\n",
    "                source_class = SOURCE_CLASS, target_class = TARGET_CLASS,\n",
    "               class_per_peer = CLASS_PER_PEER, samples_per_class = SAMPLES_PER_CLASS, \n",
    "               rate_unbalance = RATE_UNBALANCE, alpha = ALPHA, resume = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "--> Starting experiment...\n",
      "--> Loading of CIFAR10 dataset\n",
      "Using downloaded and verified file: data/cifar/cifar-10-python.tar.gz\n",
      "Extracting data/cifar/cifar-10-python.tar.gz to data/cifar/\n",
      "Files already downloaded and verified\n",
      "--> Dataset has been loaded!\n",
      "--> Creating ResNet18 model.....\n",
      "--> Model has been created!\n",
      "--> Distributing training data among peers\n",
      "--> Training data have been distributed among peers\n",
      "--> Creating peets instances\n",
      "Data set: CIFAR10\n",
      "Data distribution: IID\n",
      "Aggregation rule: fedavg\n",
      "Attack Type: label_flipping\n",
      "Attackers Ratio: 0.0 %\n",
      "Malicious Behavior Rate: 100 %\n",
      "\n",
      "===>Simulation started...\n",
      "\n",
      "====>Global model training started...\n",
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "145d253536f44caa81f44882fe539be6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " | Global training round : 1/100 |\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Baseline|: FedAvg-no attacks (FL)\n",
    "RULE = 'fedavg'\n",
    "ATTACK_TYPE='label_flipping'\n",
    "MALICIOUS_BEHAVIOR_RATE = 1\n",
    "for atr in [0.0, 0.1, 0.2, 0.3, 0.4, 0.5]:\n",
    "    run_exp(dataset_name = DATASET_NAME, model_name = MODEL_NAME, dd_type = DD_TYPE, num_peers = NUM_PEERS, \n",
    "            frac_peers = FRAC_PEERS, seed = SEED, test_batch_size = TEST_BATCH_SIZE,\n",
    "                criterion = CRITERION, global_rounds = GLOBAL_ROUNDS, local_epochs = LOCAL_EPOCHS, local_bs = LOCAL_BS, \n",
    "                 local_lr = LOCAL_LR, local_momentum = LOCAL_MOMENTUM, labels_dict = LABELS_DICT, device = DEVICE,\n",
    "                attackers_ratio = atr, attack_type=ATTACK_TYPE, \n",
    "                 malicious_behavior_rate = MALICIOUS_BEHAVIOR_RATE, rule = RULE,\n",
    "                source_class = SOURCE_CLASS, target_class = TARGET_CLASS,\n",
    "               class_per_peer = CLASS_PER_PEER, samples_per_class = SAMPLES_PER_CLASS, \n",
    "               rate_unbalance = RATE_UNBALANCE, alpha = ALPHA, resume = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "RULE = 'median'\n",
    "ATTACK_TYPE='label_flipping'\n",
    "MALICIOUS_BEHAVIOR_RATE = 1\n",
    "for atr in [0.0, 0.1, 0.2, 0.3, 0.4, 0.5]:\n",
    "    run_exp(dataset_name = DATASET_NAME, model_name = MODEL_NAME, dd_type = DD_TYPE, num_peers = NUM_PEERS, \n",
    "            frac_peers = FRAC_PEERS, seed = SEED, test_batch_size = TEST_BATCH_SIZE,\n",
    "                criterion = CRITERION, global_rounds = GLOBAL_ROUNDS, local_epochs = LOCAL_EPOCHS, local_bs = LOCAL_BS, \n",
    "                 local_lr = LOCAL_LR, local_momentum = LOCAL_MOMENTUM, labels_dict = LABELS_DICT, device = DEVICE,\n",
    "                attackers_ratio = atr, attack_type=ATTACK_TYPE, \n",
    "                 malicious_behavior_rate = MALICIOUS_BEHAVIOR_RATE, rule = RULE,\n",
    "                source_class = SOURCE_CLASS, target_class = TARGET_CLASS,\n",
    "               class_per_peer = CLASS_PER_PEER, samples_per_class = SAMPLES_PER_CLASS, \n",
    "               rate_unbalance = RATE_UNBALANCE, alpha = ALPHA, resume = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "RULE = 'rmedian'\n",
    "ATTACK_TYPE='label_flipping'\n",
    "MALICIOUS_BEHAVIOR_RATE = 1\n",
    "for atr in [0.0, 0.1, 0.2, 0.3, 0.4, 0.5]:\n",
    "    run_exp(dataset_name = DATASET_NAME, model_name = MODEL_NAME, dd_type = DD_TYPE, num_peers = NUM_PEERS, \n",
    "            frac_peers = FRAC_PEERS, seed = SEED, test_batch_size = TEST_BATCH_SIZE,\n",
    "                criterion = CRITERION, global_rounds = GLOBAL_ROUNDS, local_epochs = LOCAL_EPOCHS, local_bs = LOCAL_BS, \n",
    "                 local_lr = LOCAL_LR, local_momentum = LOCAL_MOMENTUM, labels_dict = LABELS_DICT, device = DEVICE,\n",
    "                attackers_ratio = atr, attack_type=ATTACK_TYPE, \n",
    "                 malicious_behavior_rate = MALICIOUS_BEHAVIOR_RATE, rule = RULE,\n",
    "                source_class = SOURCE_CLASS, target_class = TARGET_CLASS,\n",
    "               class_per_peer = CLASS_PER_PEER, samples_per_class = SAMPLES_PER_CLASS, \n",
    "               rate_unbalance = RATE_UNBALANCE, alpha = ALPHA, resume = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "RULE = 'tmean'\n",
    "ATTACK_TYPE='label_flipping'\n",
    "MALICIOUS_BEHAVIOR_RATE = 1\n",
    "for atr in [0.0, 0.1, 0.2, 0.3, 0.4]:\n",
    "    run_exp(dataset_name = DATASET_NAME, model_name = MODEL_NAME, dd_type = DD_TYPE, num_peers = NUM_PEERS, \n",
    "            frac_peers = FRAC_PEERS, seed = SEED, test_batch_size = TEST_BATCH_SIZE,\n",
    "                criterion = CRITERION, global_rounds = GLOBAL_ROUNDS, local_epochs = LOCAL_EPOCHS, local_bs = LOCAL_BS, \n",
    "                 local_lr = LOCAL_LR, local_momentum = LOCAL_MOMENTUM, labels_dict = LABELS_DICT, device = DEVICE,\n",
    "                attackers_ratio = atr, attack_type=ATTACK_TYPE, \n",
    "                 malicious_behavior_rate = MALICIOUS_BEHAVIOR_RATE, rule = RULE,\n",
    "                source_class = SOURCE_CLASS, target_class = TARGET_CLASS,\n",
    "               class_per_peer = CLASS_PER_PEER, samples_per_class = SAMPLES_PER_CLASS, \n",
    "               rate_unbalance = RATE_UNBALANCE, alpha = ALPHA, resume = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "RULE = 'mkrum'\n",
    "ATTACK_TYPE='label_flipping'\n",
    "MALICIOUS_BEHAVIOR_RATE = 1\n",
    "for atr in [0.0, 0.1, 0.2, 0.3, 0.4, 0.5]:\n",
    "    run_exp(dataset_name = DATASET_NAME, model_name = MODEL_NAME, dd_type = DD_TYPE, num_peers = NUM_PEERS, \n",
    "            frac_peers = FRAC_PEERS, seed = SEED, test_batch_size = TEST_BATCH_SIZE,\n",
    "                criterion = CRITERION, global_rounds = GLOBAL_ROUNDS, local_epochs = LOCAL_EPOCHS, local_bs = LOCAL_BS, \n",
    "                 local_lr = LOCAL_LR, local_momentum = LOCAL_MOMENTUM, labels_dict = LABELS_DICT, device = DEVICE,\n",
    "                attackers_ratio = atr, attack_type=ATTACK_TYPE, \n",
    "                 malicious_behavior_rate = MALICIOUS_BEHAVIOR_RATE, rule = RULE,\n",
    "                source_class = SOURCE_CLASS, target_class = TARGET_CLASS,\n",
    "               class_per_peer = CLASS_PER_PEER, samples_per_class = SAMPLES_PER_CLASS, \n",
    "               rate_unbalance = RATE_UNBALANCE, alpha = ALPHA, resume = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "RULE = 'foolsgold'\n",
    "ATTACK_TYPE='label_flipping'\n",
    "MALICIOUS_BEHAVIOR_RATE = 1\n",
    "for atr in [0.0, 0.1, 0.2, 0.3, 0.4, 0.5]:\n",
    "    run_exp(dataset_name = DATASET_NAME, model_name = MODEL_NAME, dd_type = DD_TYPE, num_peers = NUM_PEERS, \n",
    "            frac_peers = FRAC_PEERS, seed = SEED, test_batch_size = TEST_BATCH_SIZE,\n",
    "                criterion = CRITERION, global_rounds = GLOBAL_ROUNDS, local_epochs = LOCAL_EPOCHS, local_bs = LOCAL_BS, \n",
    "                 local_lr = LOCAL_LR, local_momentum = LOCAL_MOMENTUM, labels_dict = LABELS_DICT, device = DEVICE,\n",
    "                attackers_ratio = atr, attack_type=ATTACK_TYPE, \n",
    "                 malicious_behavior_rate = MALICIOUS_BEHAVIOR_RATE, rule = RULE,\n",
    "                source_class = SOURCE_CLASS, target_class = TARGET_CLASS,\n",
    "               class_per_peer = CLASS_PER_PEER, samples_per_class = SAMPLES_PER_CLASS, \n",
    "               rate_unbalance = RATE_UNBALANCE, alpha = ALPHA, resume = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "RULE = 'lfa_defender'\n",
    "ATTACK_TYPE='label_flipping'\n",
    "MALICIOUS_BEHAVIOR_RATE = 1\n",
    "for atr in [0.0, 0.1, 0.2, 0.3, 0.4, 0.5]:\n",
    "    run_exp(dataset_name = DATASET_NAME, model_name = MODEL_NAME, dd_type = DD_TYPE, num_peers = NUM_PEERS, \n",
    "            frac_peers = FRAC_PEERS, seed = SEED, test_batch_size = TEST_BATCH_SIZE,\n",
    "                criterion = CRITERION, global_rounds = GLOBAL_ROUNDS, local_epochs = LOCAL_EPOCHS, local_bs = LOCAL_BS, \n",
    "                 local_lr = LOCAL_LR, local_momentum = LOCAL_MOMENTUM, labels_dict = LABELS_DICT, device = DEVICE,\n",
    "                attackers_ratio = atr, attack_type=ATTACK_TYPE, \n",
    "                 malicious_behavior_rate = MALICIOUS_BEHAVIOR_RATE, rule = RULE,\n",
    "                source_class = SOURCE_CLASS, target_class = TARGET_CLASS,\n",
    "               class_per_peer = CLASS_PER_PEER, samples_per_class = SAMPLES_PER_CLASS, \n",
    "               rate_unbalance = RATE_UNBALANCE, alpha = ALPHA, resume = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}