run_FashionMNIST.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "b0c3b92f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "output_dir=./outputs/demo/dataset=FashionMNIST/seed=0_epoch=5_lr=0.001_alpha=0.001_tau=10.0\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import random\n",
    "\n",
    "import torch\n",
    "import torch.nn.functional as F\n",
    "import cl_gym as cl\n",
    "\n",
    "import sys\n",
    "import os\n",
    "\n",
    "seed = 0\n",
    "\n",
    "np.random.seed(seed)\n",
    "torch.manual_seed(seed)\n",
    "np.random.seed(seed)\n",
    "random.seed(seed)\n",
    "torch.cuda.manual_seed_all(seed)\n",
    "torch.manual_seed(seed)\n",
    "torch.backends.cudnn.deterministic = True\n",
    "torch.backends.cudnn.benchmark = False\n",
    "torch.backends.cudnn.enabled = False\n",
    "torch.set_num_threads(8)\n",
    "\n",
    "\n",
    "def make_params() -> dict:\n",
    "    import os\n",
    "    from pathlib import Path\n",
    "    import uuid\n",
    "\n",
    "    params = {\n",
    "            # dataset\n",
    "            # 'dataset': \"MNIST\",\n",
    "            'dataset': \"FashionMNIST\",\n",
    "            'random_class_idx': False,\n",
    "\n",
    "            # benchmark\n",
    "            'seed': seed,\n",
    "            'num_tasks': 5,\n",
    "            'epochs_per_task': 5,\n",
    "            'per_task_examples': np.inf,\n",
    "            # 'per_task_examples': 10000,\n",
    "            'per_task_memory_examples': 64,\n",
    "            'batch_size_train': 64,\n",
    "            'batch_size_memory': 64,\n",
    "            'batch_size_validation': 256,\n",
    "            'tau': 10.,\n",
    "            # 'tau': 0.0,\n",
    "\n",
    "            # algorithm\n",
    "            'optimizer': 'sgd',\n",
    "            'learning_rate': 0.001,\n",
    "            'momentum': 0.9,\n",
    "            'learning_rate_decay': 1.0,\n",
    "            'criterion': torch.nn.CrossEntropyLoss(),\n",
    "            'device': torch.device('cuda:6' if torch.cuda.is_available() else 'cpu'),\n",
    "             \n",
    "            # sample selection\n",
    "            'alpha':0.001,\n",
    "            'lambda': .1,\n",
    "            'lambda_old': 0.0,\n",
    "              }\n",
    "\n",
    "#     trial_id = str(uuid.uuid4())\n",
    "    trial_id = f\"demo/dataset={params['dataset']}/seed={params['seed']}_epoch={params['epochs_per_task']}_lr={params['learning_rate']}_alpha={params['alpha']}_tau={params['tau']}\"\n",
    "    params['trial_id'] = trial_id\n",
    "    params['output_dir'] = os.path.join(\"./outputs/{}\".format(trial_id))\n",
    "    print(f\"output_dir={params['output_dir']}\")\n",
    "    Path(params['output_dir']).mkdir(parents=True, exist_ok=True)\n",
    "\n",
    "    return params\n",
    "\n",
    "params = make_params()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "93b43ace",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0 1 2 3 4 5 6 7 8 9]\n"
     ]
    }
   ],
   "source": [
    "from datasets import MNIST\n",
    "from datasets import FashionMNIST\n",
    "\n",
    "if params['dataset'] == 'MNIST':\n",
    "    benchmark = MNIST(num_tasks=params['num_tasks'],\n",
    "                    per_task_memory_examples=params['per_task_memory_examples'],\n",
    "                    per_task_examples = params['per_task_examples'],\n",
    "                    random_class_idx = params['random_class_idx'])\n",
    "    input_dim = (28, 28)\n",
    "elif params['dataset'] == 'FashionMNIST':\n",
    "    benchmark = FashionMNIST(num_tasks=params['num_tasks'],\n",
    "                            per_task_memory_examples=params['per_task_memory_examples'],\n",
    "                            per_task_examples = params['per_task_examples'],\n",
    "                            random_class_idx = params['random_class_idx'])\n",
    "    input_dim = (28, 28)\n",
    "else:\n",
    "    raise NotImplementedError\n",
    "class_idx = benchmark.class_idx\n",
    "num_classes = len(class_idx)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "20c7b886",
   "metadata": {},
   "outputs": [],
   "source": [
    "from algorithms.imbalance import ImbalanceAlgorithm\n",
    "from metrics import MetricCollector2\n",
    "from backbones import MLP2Layers2\n",
    "\n",
    "backbone = MLP2Layers2(\n",
    "    input_dim=input_dim, \n",
    "    hidden_dim_1=256, \n",
    "    hidden_dim_2=256, \n",
    "    output_dim=num_classes,\n",
    "    class_idx=class_idx,\n",
    "    config=params\n",
    "    ).to(params['device'])\n",
    "algorithm = ImbalanceAlgorithm(backbone, benchmark, params, requires_memory=True)\n",
    "metric_manager_callback = MetricCollector2(num_tasks=params['num_tasks'],\n",
    "                                                        eval_interval='epoch',\n",
    "                                                        epochs_per_task=params['epochs_per_task'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "d8efa3ad",
   "metadata": {},
   "outputs": [],
   "source": [
    "from trainers.imbalance_trainer import ImbalanceContinualTrainer1 as ContinualTrainer\n",
    "\n",
    "trainer = ContinualTrainer(algorithm, params, callbacks=[metric_manager_callback])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "7c1bb4fe",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "---------------------------- Task 1 -----------------------\n",
      "[1] Eval metrics for task 1 >> {'accuracy': 0.9715, 'loss': 0.00033985645696520806, 'std': 0.004500000000000004, 'EER': -1}\n",
      "[2] Eval metrics for task 1 >> {'accuracy': 0.978, 'loss': 0.000244151184335351, 'std': 0.0, 'EER': -1}\n",
      "[3] Eval metrics for task 1 >> {'accuracy': 0.9815, 'loss': 0.00020586758106946946, 'std': 0.006500000000000006, 'EER': -1}\n",
      "[4] Eval metrics for task 1 >> {'accuracy': 0.984, 'loss': 0.0001882472177967429, 'std': 0.007000000000000006, 'EER': -1}\n",
      "[5] Eval metrics for task 1 >> {'accuracy': 0.985, 'loss': 0.0001759307961910963, 'std': 0.006000000000000005, 'EER': -1}\n",
      "training_task_end\n",
      "---------------------------- Task 2 -----------------------\n",
      "Elapsed time(grad):5.733\n",
      "losses=tensor([[ 0.0198,  0.0195,  8.7685, 12.5658]])\n",
      "### Cplex absolute_and_nonabsolute_minsum LP solver ###\n",
      "Elapsed time(optim):11.283\n",
      "Fairness:[ 0.          0.79366192  0.06358251 -0.8572444 ]\n",
      "Current class expected loss:[ 0.          0.         -0.07740601 -0.26157141]\n",
      "len(updated_seq_indices)=11840\n",
      "[6] Eval metrics for task 1 >> {'accuracy': 0.8985000000000001, 'loss': 0.0014707416594028472, 'std': 0.03650000000000003, 'EER': -1}\n",
      "[6] Eval metrics for task 2 >> {'accuracy': 0.9005, 'loss': 0.0012480256855487823, 'std': 0.04349999999999998, 'EER': -1}\n",
      "Elapsed time(grad):5.429\n",
      "losses=tensor([[0.0126, 0.0032, 0.2033, 0.3676]])\n",
      "### Cplex absolute_and_nonabsolute_minsum LP solver ###\n",
      "Elapsed time(optim):11.738\n",
      "Fairness:[4.20494013e-08 0.00000000e+00 0.00000000e+00 0.00000000e+00]\n",
      "Current class expected loss:[ 0.          0.         -0.00849676 -0.00849677]\n",
      "len(updated_seq_indices)=1600\n",
      "[7] Eval metrics for task 1 >> {'accuracy': 0.9245, 'loss': 0.0010876024067401886, 'std': 0.02849999999999997, 'EER': -1}\n",
      "[7] Eval metrics for task 2 >> {'accuracy': 0.813, 'loss': 0.0024259970784187316, 'std': 0.014999999999999958, 'EER': -1}\n",
      "Elapsed time(grad):5.582\n",
      "losses=tensor([[3.7318e-03, 6.5331e-04, 6.9319e-01, 4.7807e-01]])\n",
      "### Cplex absolute_and_nonabsolute_minsum LP solver ###\n",
      "Elapsed time(optim):11.436\n",
      "Fairness:[2.78397744e-08 0.00000000e+00 0.00000000e+00 0.00000000e+00]\n",
      "Current class expected loss:[ 0.          0.         -0.01090839 -0.0109084 ]\n",
      "len(updated_seq_indices)=1472\n",
      "[8] Eval metrics for task 1 >> {'accuracy': 0.9410000000000001, 'loss': 0.0008402319177985191, 'std': 0.012999999999999956, 'EER': -1}\n",
      "[8] Eval metrics for task 2 >> {'accuracy': 0.792, 'loss': 0.0027824744284152985, 'std': 0.05199999999999999, 'EER': -1}\n",
      "Elapsed time(grad):4.968\n",
      "losses=tensor([[8.0726e-04, 4.0776e-04, 5.7459e-01, 7.6856e-01]])\n",
      "### Cplex absolute_and_nonabsolute_minsum LP solver ###\n",
      "Elapsed time(optim):11.26\n",
      "Fairness:[4.79227553e-08 0.00000000e+00 0.00000000e+00 0.00000000e+00]\n",
      "Current class expected loss:[ 0.          0.         -0.01681463 -0.01681463]\n",
      "len(updated_seq_indices)=1216\n",
      "[9] Eval metrics for task 1 >> {'accuracy': 0.94, 'loss': 0.0008811754584312439, 'std': 0.012999999999999956, 'EER': -1}\n",
      "[9] Eval metrics for task 2 >> {'accuracy': 0.78, 'loss': 0.0029394803643226623, 'std': 0.017000000000000015, 'EER': -1}\n",
      "Elapsed time(grad):5.553\n",
      "losses=tensor([[7.8725e-04, 4.6445e-04, 7.0819e-01, 7.1400e-01]])\n",
      "### Cplex absolute_and_nonabsolute_minsum LP solver ###\n",
      "Elapsed time(optim):11.317\n",
      "Fairness:[3.72050184e-08 0.00000000e+00 0.00000000e+00 0.00000000e+00]\n",
      "Current class expected loss:[ 0.          0.         -0.01581688 -0.01581689]\n",
      "len(updated_seq_indices)=1408\n",
      "[10] Eval metrics for task 1 >> {'accuracy': 0.9410000000000001, 'loss': 0.0008769591003656387, 'std': 0.011999999999999955, 'EER': -1}\n",
      "[10] Eval metrics for task 2 >> {'accuracy': 0.7735000000000001, 'loss': 0.0030360363125801087, 'std': 0.020500000000000018, 'EER': -1}\n",
      "training_task_end\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "---------------------------- Task 3 -----------------------\n",
      "Elapsed time(grad):6.735\n",
      "losses=tensor([[6.1697e-04, 4.2972e-04, 1.0220e+00, 6.6581e-01, 1.0766e+01, 5.0212e+00]])\n",
      "### Cplex absolute_and_nonabsolute_minsum LP solver ###\n",
      "Elapsed time(optim):14.722\n",
      "Fairness:[ 0.33580073  0.34593127 -0.10027229  0.04606414 -0.62752403  0.        ]\n",
      "Current class expected loss:[ 0.          0.          0.          0.         -0.26665879 -0.07840162]\n",
      "len(updated_seq_indices)=9024\n",
      "[11] Eval metrics for task 1 >> {'accuracy': 0.8545, 'loss': 0.0026130593866109846, 'std': 0.0645, 'EER': -1}\n",
      "[11] Eval metrics for task 2 >> {'accuracy': 0.76, 'loss': 0.0031204813420772554, 'std': 0.039000000000000035, 'EER': -1}\n",
      "[11] Eval metrics for task 3 >> {'accuracy': 0.9139999999999999, 'loss': 0.0010427628234028817, 'std': 0.07700000000000001, 'EER': -1}\n",
      "Elapsed time(grad):5.502\n",
      "losses=tensor([[0.0048, 0.0044, 0.0438, 0.0227, 0.3994, 0.1416]])\n",
      "### Cplex absolute_and_nonabsolute_minsum LP solver ###\n",
      "Elapsed time(optim):14.942\n",
      "Fairness:[ 0.01684997  0.         -0.01684998  0.          0.          0.        ]\n",
      "Current class expected loss:[ 0.          0.          0.          0.         -0.00758622 -0.00758622]\n",
      "len(updated_seq_indices)=576\n",
      "[12] Eval metrics for task 1 >> {'accuracy': 0.8105, 'loss': 0.003692596584558487, 'std': 0.09750000000000003, 'EER': -1}\n",
      "[12] Eval metrics for task 2 >> {'accuracy': 0.829, 'loss': 0.0023936653435230257, 'std': 0.0030000000000000027, 'EER': -1}\n",
      "[12] Eval metrics for task 3 >> {'accuracy': 0.8385, 'loss': 0.0017119548618793488, 'std': 0.15449999999999997, 'EER': -1}\n",
      "Elapsed time(grad):4.73\n",
      "losses=tensor([[0.0287, 0.0081, 0.0107, 0.0209, 0.7380, 0.1118]])\n",
      "### Cplex absolute_and_nonabsolute_minsum LP solver ###\n",
      "Elapsed time(optim):15.3\n",
      "Fairness:[ 0.03773563  0.00429523 -0.04203087  0.          0.          0.        ]\n",
      "Current class expected loss:[ 0.          0.          0.          0.         -0.01127922 -0.01127922]\n",
      "len(updated_seq_indices)=832\n",
      "[13] Eval metrics for task 1 >> {'accuracy': 0.8260000000000001, 'loss': 0.0033273731768131256, 'std': 0.07500000000000001, 'EER': -1}\n",
      "[13] Eval metrics for task 2 >> {'accuracy': 0.819, 'loss': 0.0025316413938999177, 'std': 0.022999999999999965, 'EER': -1}\n",
      "[13] Eval metrics for task 3 >> {'accuracy': 0.8465, 'loss': 0.0015707600861787796, 'std': 0.14750000000000002, 'EER': -1}\n",
      "Elapsed time(grad):5.348\n",
      "losses=tensor([[0.0126, 0.0070, 0.0146, 0.0146, 0.6808, 0.0999]])\n",
      "### Cplex absolute_and_nonabsolute_minsum LP solver ###\n",
      "Elapsed time(optim):15.583\n",
      "Fairness:[ 0.02846479  0.01278051 -0.0412453   0.          0.          0.        ]\n",
      "Current class expected loss:[ 0.          0.          0.          0.         -0.00997084 -0.00997084]\n",
      "len(updated_seq_indices)=896\n",
      "[14] Eval metrics for task 1 >> {'accuracy': 0.833, 'loss': 0.003303080439567566, 'std': 0.07, 'EER': -1}\n",
      "[14] Eval metrics for task 2 >> {'accuracy': 0.8180000000000001, 'loss': 0.0025839677155017854, 'std': 0.00599999999999995, 'EER': -1}\n",
      "[14] Eval metrics for task 3 >> {'accuracy': 0.8335, 'loss': 0.001808586522936821, 'std': 0.16049999999999998, 'EER': -1}\n",
      "Elapsed time(grad):5.117\n",
      "losses=tensor([[0.0071, 0.0064, 0.0086, 0.0120, 0.7948, 0.1056]])\n",
      "### Cplex absolute_and_nonabsolute_minsum LP solver ###\n",
      "Elapsed time(optim):15.383\n",
      "Fairness:[ 0.03343788  0.02148125 -0.05491914  0.          0.          0.        ]\n",
      "Current class expected loss:[ 0.          0.          0.          0.         -0.01058457 -0.01058457]\n",
      "len(updated_seq_indices)=960\n",
      "[15] Eval metrics for task 1 >> {'accuracy': 0.8245, 'loss': 0.003478297472000122, 'std': 0.07850000000000001, 'EER': -1}\n",
      "[15] Eval metrics for task 2 >> {'accuracy': 0.8205, 'loss': 0.0026436534523963926, 'std': 0.01749999999999996, 'EER': -1}\n",
      "[15] Eval metrics for task 3 >> {'accuracy': 0.847, 'loss': 0.0016512087881565094, 'std': 0.14800000000000002, 'EER': -1}\n",
      "training_task_end\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "---------------------------- Task 4 -----------------------\n",
      "Elapsed time(grad):5.059\n",
      "losses=tensor([[9.2899e-03, 4.1465e-03, 8.8448e-03, 9.5310e-03, 7.3148e-01, 7.8498e-02,\n",
      "         1.1693e+01, 8.8563e+00]])\n",
      "### Cplex absolute_and_nonabsolute_minsum LP solver ###\n",
      "Elapsed time(optim):18.649\n",
      "Fairness:[ 0.17198677  0.2227017   0.20766674  0.28182848  0.05449505 -0.09969759\n",
      " -0.66834691 -0.17063428]\n",
      "Current class expected loss:[ 0.          0.          0.          0.          0.          0.\n",
      " -0.34138961 -0.14230459]\n",
      "len(updated_seq_indices)=12000\n",
      "[16] Eval metrics for task 1 >> {'accuracy': 0.7484999999999999, 'loss': 0.004442325592041016, 'std': 0.16950000000000004, 'EER': -1}\n",
      "[16] Eval metrics for task 2 >> {'accuracy': 0.614, 'loss': 0.006746992468833923, 'std': 0.129, 'EER': -1}\n",
      "[16] Eval metrics for task 3 >> {'accuracy': 0.7130000000000001, 'loss': 0.0032867903709411623, 'std': 0.04799999999999999, 'EER': -1}\n",
      "[16] Eval metrics for task 4 >> {'accuracy': 0.8200000000000001, 'loss': 0.002080927923321724, 'std': 0.14499999999999996, 'EER': -1}\n",
      "Elapsed time(grad):6.465\n",
      "losses=tensor([[0.0361, 0.0016, 0.0185, 0.0101, 0.0265, 0.0271, 0.8095, 0.1189]])\n",
      "### Cplex absolute_and_nonabsolute_minsum LP solver ###\n",
      "Elapsed time(optim):19.014\n",
      "Fairness:[-0.00031796  0.0255905   0.          0.          0.          0.00232192\n",
      " -0.04995611  0.02236165]\n",
      "Current class expected loss:[ 0.          0.          0.          0.          0.          0.\n",
      " -0.02892711  0.        ]\n",
      "len(updated_seq_indices)=512\n",
      "[17] Eval metrics for task 1 >> {'accuracy': 0.7965, 'loss': 0.0036212439239025115, 'std': 0.10750000000000004, 'EER': -1}\n",
      "[17] Eval metrics for task 2 >> {'accuracy': 0.5185, 'loss': 0.008924689650535583, 'std': 0.06249999999999997, 'EER': -1}\n",
      "[17] Eval metrics for task 3 >> {'accuracy': 0.618, 'loss': 0.0047480337619781494, 'std': 0.03300000000000003, 'EER': -1}\n",
      "[17] Eval metrics for task 4 >> {'accuracy': 0.8260000000000001, 'loss': 0.0020600121170282364, 'std': 0.16699999999999998, 'EER': -1}\n",
      "Elapsed time(grad):5.115\n",
      "losses=tensor([[0.0102, 0.0065, 0.0267, 0.1087, 0.0361, 0.1300, 0.8862, 0.0424]])\n",
      "### Cplex absolute_and_nonabsolute_minsum LP solver ###\n",
      "Elapsed time(optim):19.01\n",
      "Fairness:[ 0.00000000e+00  2.71875339e-02  0.00000000e+00  0.00000000e+00\n",
      "  1.73472348e-18  2.64857475e-05 -5.32159869e-02  2.60019755e-02]\n",
      "Current class expected loss:[ 0.          0.          0.          0.          0.          0.\n",
      " -0.03168719  0.        ]\n",
      "len(updated_seq_indices)=448\n",
      "[18] Eval metrics for task 1 >> {'accuracy': 0.7935000000000001, 'loss': 0.003864931732416153, 'std': 0.11349999999999999, 'EER': -1}\n",
      "[18] Eval metrics for task 2 >> {'accuracy': 0.587, 'loss': 0.007364213347434998, 'std': 0.184, 'EER': -1}\n",
      "[18] Eval metrics for task 3 >> {'accuracy': 0.6955, 'loss': 0.0035259064733982087, 'std': 0.004500000000000004, 'EER': -1}\n",
      "[18] Eval metrics for task 4 >> {'accuracy': 0.8005, 'loss': 0.0022678533494472504, 'std': 0.1775, 'EER': -1}\n",
      "Elapsed time(grad):5.562\n",
      "losses=tensor([[0.0091, 0.0055, 0.0473, 0.0060, 0.0279, 0.0475, 0.9483, 0.0913]])\n",
      "### Cplex absolute_and_nonabsolute_minsum LP solver ###\n",
      "Elapsed time(optim):19.609\n",
      "Fairness:[-0.00753515  0.03040073  0.          0.          0.          0.00113705\n",
      " -0.05131686  0.02731423]\n",
      "Current class expected loss:[ 0.          0.          0.          0.          0.          0.\n",
      " -0.03145244  0.        ]\n",
      "len(updated_seq_indices)=576\n",
      "[19] Eval metrics for task 1 >> {'accuracy': 0.763, 'loss': 0.004639093607664109, 'std': 0.16000000000000003, 'EER': -1}\n",
      "[19] Eval metrics for task 2 >> {'accuracy': 0.661, 'loss': 0.006280809998512268, 'std': 0.08500000000000002, 'EER': -1}\n",
      "[19] Eval metrics for task 3 >> {'accuracy': 0.651, 'loss': 0.004200104802846908, 'std': 0.03400000000000003, 'EER': -1}\n",
      "[19] Eval metrics for task 4 >> {'accuracy': 0.806, 'loss': 0.0023621878027915953, 'std': 0.184, 'EER': -1}\n",
      "Elapsed time(grad):5.174\n",
      "losses=tensor([[0.0254, 0.0013, 0.0101, 0.0152, 0.0243, 0.0816, 1.0197, 0.0520]])\n",
      "### Cplex absolute_and_nonabsolute_minsum LP solver ###\n",
      "Elapsed time(optim):18.829\n",
      "Fairness:[ 0.          0.03290372  0.          0.          0.          0.00181056\n",
      " -0.06190682  0.02719254]\n",
      "Current class expected loss:[ 0.          0.          0.          0.          0.          0.\n",
      " -0.03563975  0.        ]\n",
      "len(updated_seq_indices)=576\n",
      "[20] Eval metrics for task 1 >> {'accuracy': 0.794, 'loss': 0.0039762196540832516, 'std': 0.10799999999999998, 'EER': -1}\n",
      "[20] Eval metrics for task 2 >> {'accuracy': 0.563, 'loss': 0.00831496912240982, 'std': 0.04499999999999999, 'EER': -1}\n",
      "[20] Eval metrics for task 3 >> {'accuracy': 0.6930000000000001, 'loss': 0.0036173990070819855, 'std': 0.025999999999999968, 'EER': -1}\n",
      "[20] Eval metrics for task 4 >> {'accuracy': 0.7935, 'loss': 0.0025485284924507143, 'std': 0.1945, 'EER': -1}\n",
      "training_task_end\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "---------------------------- Task 5 -----------------------\n",
      "Elapsed time(grad):6.628\n",
      "losses=tensor([[8.7837e-03, 5.5662e-03, 1.2797e-02, 6.6567e-02, 2.0752e-02, 4.6943e-02,\n",
      "         1.2180e+00, 1.3534e-01, 8.1631e+00, 1.0387e+01]])\n",
      "### Cplex absolute_and_nonabsolute_minsum LP solver ###\n",
      "Elapsed time(optim):22.747\n",
      "Fairness:[ 0.14577574  0.11983806  0.16066068  0.14986013  0.09223527  0.210518\n",
      " -0.12869993 -0.12563444 -0.25013431 -0.37441947]\n",
      "Current class expected loss:[ 0.          0.          0.          0.          0.          0.\n",
      "  0.          0.         -0.16152417 -0.22366673]\n",
      "len(updated_seq_indices)=11264\n",
      "[21] Eval metrics for task 1 >> {'accuracy': 0.768, 'loss': 0.005037185102701187, 'std': 0.14600000000000002, 'EER': -1}\n",
      "[21] Eval metrics for task 2 >> {'accuracy': 0.6225, 'loss': 0.008337077021598817, 'std': 0.1355, 'EER': -1}\n",
      "[21] Eval metrics for task 3 >> {'accuracy': 0.622, 'loss': 0.006127363979816437, 'std': 0.04600000000000004, 'EER': -1}\n",
      "[21] Eval metrics for task 4 >> {'accuracy': 0.7795000000000001, 'loss': 0.0029113339483737947, 'std': 0.09249999999999997, 'EER': -1}\n",
      "[21] Eval metrics for task 5 >> {'accuracy': 0.9259999999999999, 'loss': 0.0009058945402503014, 'std': 0.025999999999999968, 'EER': -1}\n",
      "Elapsed time(grad):5.25\n",
      "losses=tensor([[0.0070, 0.0020, 0.0050, 0.0063, 0.0120, 0.0088, 0.0261, 0.0609, 0.3147,\n",
      "         0.1399]])\n",
      "### Cplex absolute_and_nonabsolute_minsum LP solver ###\n",
      "Elapsed time(optim):22.051\n",
      "Fairness:[ 4.24150794e-03  6.06308830e-03  5.73385621e-03  0.00000000e+00\n",
      " -8.67361738e-19  0.00000000e+00 -3.73691005e-03 -1.23015403e-02\n",
      "  0.00000000e+00  0.00000000e+00]\n",
      "Current class expected loss:[ 0.          0.          0.          0.          0.          0.\n",
      "  0.          0.         -0.00311831 -0.00311831]\n",
      "len(updated_seq_indices)=960\n",
      "[22] Eval metrics for task 1 >> {'accuracy': 0.757, 'loss': 0.0055060192346572875, 'std': 0.15900000000000003, 'EER': -1}\n",
      "[22] Eval metrics for task 2 >> {'accuracy': 0.6205, 'loss': 0.00859028595685959, 'std': 0.14350000000000002, 'EER': -1}\n",
      "[22] Eval metrics for task 3 >> {'accuracy': 0.582, 'loss': 0.006987734138965607, 'std': 0.019000000000000017, 'EER': -1}\n",
      "[22] Eval metrics for task 4 >> {'accuracy': 0.83, 'loss': 0.002397051274776459, 'std': 0.133, 'EER': -1}\n",
      "[22] Eval metrics for task 5 >> {'accuracy': 0.863, 'loss': 0.0015582866817712784, 'std': 0.050000000000000044, 'EER': -1}\n",
      "Elapsed time(grad):5.933\n",
      "losses=tensor([[0.0121, 0.0017, 0.0056, 0.0073, 0.0114, 0.0101, 0.0257, 0.0060, 0.2706,\n",
      "         0.5076]])\n",
      "### Cplex absolute_and_nonabsolute_minsum LP solver ###\n",
      "Elapsed time(optim):22.514\n",
      "Fairness:[ 6.14261158e-03  1.01841312e-02  8.50154624e-03  0.00000000e+00\n",
      " -8.67361738e-19  0.00000000e+00 -2.04914600e-03 -2.27791368e-02\n",
      "  0.00000000e+00  0.00000000e+00]\n",
      "Current class expected loss:[ 0.          0.          0.          0.          0.          0.\n",
      "  0.          0.         -0.00402287 -0.00402287]\n",
      "len(updated_seq_indices)=1344\n",
      "[23] Eval metrics for task 1 >> {'accuracy': 0.7635000000000001, 'loss': 0.0053127301931381226, 'std': 0.15050000000000002, 'EER': -1}\n",
      "[23] Eval metrics for task 2 >> {'accuracy': 0.6015, 'loss': 0.009220065414905549, 'std': 0.1415, 'EER': -1}\n",
      "[23] Eval metrics for task 3 >> {'accuracy': 0.63, 'loss': 0.006104864180088043, 'std': 0.02400000000000002, 'EER': -1}\n",
      "[23] Eval metrics for task 4 >> {'accuracy': 0.8275, 'loss': 0.0024523099064826964, 'std': 0.14049999999999996, 'EER': -1}\n",
      "[23] Eval metrics for task 5 >> {'accuracy': 0.845, 'loss': 0.0017762907147407532, 'std': 0.064, 'EER': -1}\n",
      "Elapsed time(grad):5.198\n",
      "losses=tensor([[0.0091, 0.0021, 0.0071, 0.0122, 0.0081, 0.0042, 0.0234, 0.0049, 0.2815,\n",
      "         0.6031]])\n",
      "### Cplex absolute_and_nonabsolute_minsum LP solver ###\n",
      "Elapsed time(optim):23.263\n",
      "Fairness:[ 0.006182    0.00994004  0.00848641  0.          0.          0.\n",
      " -0.00068514 -0.02392331  0.          0.        ]\n",
      "Current class expected loss:[ 0.          0.          0.          0.          0.          0.\n",
      "  0.          0.         -0.00411653 -0.00411653]\n",
      "len(updated_seq_indices)=1280\n",
      "[24] Eval metrics for task 1 >> {'accuracy': 0.769, 'loss': 0.005175349950790406, 'std': 0.14200000000000002, 'EER': -1}\n",
      "[24] Eval metrics for task 2 >> {'accuracy': 0.6194999999999999, 'loss': 0.008772728025913239, 'std': 0.1405, 'EER': -1}\n",
      "[24] Eval metrics for task 3 >> {'accuracy': 0.6214999999999999, 'loss': 0.006380520820617676, 'std': 0.04150000000000004, 'EER': -1}\n",
      "[24] Eval metrics for task 4 >> {'accuracy': 0.825, 'loss': 0.002541595607995987, 'std': 0.14099999999999996, 'EER': -1}\n",
      "[24] Eval metrics for task 5 >> {'accuracy': 0.8540000000000001, 'loss': 0.0016878833919763566, 'std': 0.04499999999999999, 'EER': -1}\n",
      "Elapsed time(grad):5.148\n",
      "losses=tensor([[0.0063, 0.0025, 0.0050, 0.0055, 0.0084, 0.0033, 0.0199, 0.0044, 0.3104,\n",
      "         0.5264]])\n",
      "### Cplex absolute_and_nonabsolute_minsum LP solver ###\n",
      "Elapsed time(optim):22.24\n",
      "Fairness:[ 0.00666584  0.00918433  0.00835461  0.          0.          0.\n",
      "  0.         -0.02420478  0.          0.        ]\n",
      "Current class expected loss:[ 0.          0.          0.          0.          0.          0.\n",
      "  0.          0.         -0.00406402 -0.00406402]\n",
      "len(updated_seq_indices)=1152\n",
      "[25] Eval metrics for task 1 >> {'accuracy': 0.7725, 'loss': 0.005193375647068023, 'std': 0.1405, 'EER': -1}\n",
      "[25] Eval metrics for task 2 >> {'accuracy': 0.6075, 'loss': 0.009171107172966004, 'std': 0.1445, 'EER': -1}\n",
      "[25] Eval metrics for task 3 >> {'accuracy': 0.6234999999999999, 'loss': 0.006446783840656281, 'std': 0.04750000000000004, 'EER': -1}\n",
      "[25] Eval metrics for task 4 >> {'accuracy': 0.8274999999999999, 'loss': 0.0025062357783317566, 'std': 0.1365, 'EER': -1}\n",
      "[25] Eval metrics for task 5 >> {'accuracy': 0.851, 'loss': 0.0017564636170864106, 'std': 0.056999999999999995, 'EER': -1}\n",
      "training_task_end\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "final avg-acc 0.7363999999999999\n",
      "final avg-forget 0.16225\n"
     ]
    }
   ],
   "source": [
    "trainer.run()\n",
    "print(\"final avg-acc\", metric_manager_callback.meters['accuracy'].compute_final())\n",
    "print(\"final avg-forget\", metric_manager_callback.meters['forgetting'].compute_final())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "beaf952b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.985, 0.   , 0.   , 0.   , 0.   ],\n",
       "       [0.941, 0.774, 0.   , 0.   , 0.   ],\n",
       "       [0.824, 0.82 , 0.847, 0.   , 0.   ],\n",
       "       [0.794, 0.563, 0.693, 0.794, 0.   ],\n",
       "       [0.772, 0.608, 0.624, 0.827, 0.851]])"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metric_manager_callback.meters['accuracy'].get_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "26972f73",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8240383333333332"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(metric_manager_callback.meters['accuracy'].compute_overall())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "e950fb61",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.0, 0.084, 0.011, 0.083, 0.097]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[np.round(x, 3) for x in metric_manager_callback.meters['EER'].compute_overall()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "2ebfb58a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.054846777777777765"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(metric_manager_callback.meters['EER'].compute_overall())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "06af72df",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.006, 0.085, 0.098, 0.148, 0.153]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[np.round(x, 3) for x in metric_manager_callback.meters['std'].compute_overall()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "7039ed8b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.09813281356742032"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(metric_manager_callback.meters['std'].compute_overall())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "706ad34c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy:0.8240383333333332\n",
      "EER:0.054846777777777765\n",
      "std:0.09813281356742032\n"
     ]
    }
   ],
   "source": [
    "print(f\"accuracy:{np.mean(metric_manager_callback.meters['accuracy'].compute_overall())}\")\n",
    "print(f\"EER:{np.mean(metric_manager_callback.meters['EER'].compute_overall())}\")\n",
    "print(f\"std:{np.mean(metric_manager_callback.meters['std'].compute_overall())}\")\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "cil",
   "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": "3.9.18"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}