Decoding Attention Level with Braindecode smaller Windows

AttentionExp.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(2272, 8, 128)\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "data = pd.read_csv('./user1_test10_attention.csv')\n",
    "\n",
    "import numpy as np\n",
    "import torch\n",
    "torch.backends.cudnn.benchmark = True\n",
    "\n",
    "window_size_samples = 128\n",
    "\n",
    "X = [np.array(data.iloc[i:i+window_size_samples,:-1]) \n",
    "     for i in range(data.shape[0] - window_size_samples)]\n",
    "X = np.array(X).transpose(0,2,1).astype(np.float32)\n",
    "y = np.array(data.iloc[window_size_samples:,-1]).astype(np.int64) - 1\n",
    "\n",
    "print(X.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "from braindecode.datautil import create_from_X_y\n",
    "# or in prev versions: from braindecode.datasets.xy import create_from_X_y\n",
    "\n",
    "\n",
    "n_split = 1600\n",
    "train_X, train_y = X[:n_split], y[:n_split]\n",
    "valid_X, valid_y = X[n_split:], y[n_split:]\n",
    "means = train_X.mean(axis=(0,2), keepdims=True)\n",
    "stds = train_X.std(axis=(0,2), keepdims=True)\n",
    " \n",
    "# standardize per channel\n",
    "train_X = (train_X - means) / (stds)\n",
    "valid_X = (valid_X - means) / (stds)\n",
    "\n",
    "\n",
    "train_set = create_from_X_y(train_X, train_y, drop_last_window=False)\n",
    "\n",
    "valid_set = create_from_X_y(valid_X, valid_y, drop_last_window=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "from braindecode.util import set_random_seeds\n",
    "from braindecode.models import ShallowFBCSPNet\n",
    "\n",
    "cuda = torch.cuda.is_available()  # check if GPU is available, if True chooses to use it\n",
    "device = 'cuda' if cuda else 'cpu'\n",
    "if cuda:\n",
    "    torch.backends.cudnn.benchmark = True\n",
    "seed = 20200220  # random seed to make results reproducible\n",
    "# Set random seed to be able to reproduce results\n",
    "set_random_seeds(seed=seed, cuda=cuda)\n",
    "\n",
    "n_classes=5\n",
    "# Extract number of chans and time steps from dataset\n",
    "n_chans = train_set[0][0].shape[0]\n",
    "input_window_samples = train_set[0][0].shape[1]\n",
    "\n",
    "model = ShallowFBCSPNet(\n",
    "    n_chans,\n",
    "    n_classes,\n",
    "    input_window_samples=input_window_samples,\n",
    "    final_conv_length='auto',\n",
    ")\n",
    "\n",
    "# Send model to GPU\n",
    "if cuda:\n",
    "    model.cuda()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  epoch    train_accuracy    train_loss    valid_accuracy    valid_loss     dur\n",
      "-------  ----------------  ------------  ----------------  ------------  ------\n",
      "      1            \u001b[36m0.5519\u001b[0m        \u001b[32m1.1509\u001b[0m            \u001b[35m0.0000\u001b[0m        \u001b[31m2.2398\u001b[0m  1.3503\n",
      "      2            \u001b[36m0.6606\u001b[0m        \u001b[32m0.7179\u001b[0m            \u001b[35m0.1726\u001b[0m        4.3311  1.3391\n",
      "      3            \u001b[36m0.9113\u001b[0m        \u001b[32m0.4896\u001b[0m            0.1726        2.9743  1.3597\n",
      "      4            \u001b[36m0.9581\u001b[0m        \u001b[32m0.3229\u001b[0m            0.1726        4.4045  1.3775\n",
      "      5            0.9300        \u001b[32m0.2481\u001b[0m            0.1726        4.0017  1.3307\n",
      "      6            \u001b[36m0.9750\u001b[0m        \u001b[32m0.2348\u001b[0m            0.1726        5.6199  1.3246\n",
      "      7            \u001b[36m0.9881\u001b[0m        \u001b[32m0.1922\u001b[0m            0.1726        6.0158  1.3421\n",
      "      8            0.9400        \u001b[32m0.1526\u001b[0m            0.1726        7.7374  1.2859\n",
      "      9            0.9738        \u001b[32m0.1486\u001b[0m            0.1726        4.7569  1.1396\n",
      "     10            0.9675        \u001b[32m0.1331\u001b[0m            0.1726        5.1696  1.1235\n",
      "     11            0.8912        \u001b[32m0.1180\u001b[0m            0.1726        7.8588  1.1291\n",
      "     12            0.9487        0.1250            \u001b[35m0.3839\u001b[0m        \u001b[31m1.3711\u001b[0m  1.0218\n",
      "     13            0.9775        0.1919            0.1726        3.1013  1.0330\n",
      "     14            \u001b[36m0.9975\u001b[0m        \u001b[32m0.1103\u001b[0m            0.1726        4.3662  1.3610\n",
      "     15            0.9956        \u001b[32m0.1028\u001b[0m            0.1726        4.7099  1.3303\n",
      "     16            0.9794        \u001b[32m0.0814\u001b[0m            0.1726        6.6700  1.3671\n",
      "     17            0.9888        0.0982            0.2827        4.3589  1.3759\n",
      "     18            0.9862        \u001b[32m0.0740\u001b[0m            0.1726        6.2744  1.3433\n",
      "     19            0.9831        0.0955            0.1726        6.5324  1.3806\n",
      "     20            0.9788        \u001b[32m0.0713\u001b[0m            0.1726        7.6627  1.3642\n",
      "     21            0.9956        \u001b[32m0.0658\u001b[0m            0.2783        4.2936  1.3517\n",
      "     22            0.9806        \u001b[32m0.0624\u001b[0m            0.1726        7.2627  1.3614\n",
      "     23            0.9925        \u001b[32m0.0518\u001b[0m            0.1726        5.6753  1.3646\n",
      "     24            0.9875        0.0534            0.3363        4.2460  1.3827\n",
      "     25            0.9800        0.0572            0.1726        7.7427  0.9306\n",
      "     26            0.9862        0.0651            0.2500        6.2331  0.9330\n",
      "     27            0.9962        0.0600            0.2768        4.6223  0.9309\n",
      "     28            0.9744        0.0559            0.2976        2.9818  0.9232\n",
      "     29            0.9894        0.0612            0.2619        6.2250  0.9244\n",
      "     30            0.9900        0.0623            0.2158        6.8295  0.9265\n",
      "     31            0.9881        0.0578            0.3333        4.4365  1.0223\n",
      "     32            0.9956        \u001b[32m0.0495\u001b[0m            0.2798        4.9459  0.9264\n",
      "     33            0.9956        0.0525            0.2902        6.0256  1.0238\n",
      "     34            0.9975        0.0507            0.3110        4.6143  0.9232\n",
      "     35            0.9900        \u001b[32m0.0462\u001b[0m            0.2857        6.0690  0.9318\n",
      "     36            0.9931        0.0610            0.2991        5.7356  0.9115\n",
      "     37            0.9938        0.0516            0.2976        5.8834  0.9155\n",
      "     38            0.9944        0.0504            0.2976        5.8595  0.9123\n",
      "     39            0.9931        0.0680            0.2946        5.9841  0.9106\n",
      "     40            \u001b[36m0.9981\u001b[0m        0.0526            0.3140        5.2729  0.9140\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<class 'braindecode.classifier.EEGClassifier'>[initialized](\n",
       "  module_=ShallowFBCSPNet(\n",
       "    (ensuredims): Ensure4d()\n",
       "    (dimshuffle): Expression(expression=transpose_time_to_spat) \n",
       "    (conv_time): Conv2d(1, 40, kernel_size=(25, 1), stride=(1, 1))\n",
       "    (conv_spat): Conv2d(40, 40, kernel_size=(1, 8), stride=(1, 1), bias=False)\n",
       "    (bnorm): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    (conv_nonlin_exp): Expression(expression=square) \n",
       "    (pool): AvgPool2d(kernel_size=(75, 1), stride=(15, 1), padding=0)\n",
       "    (pool_nonlin_exp): Expression(expression=safe_log) \n",
       "    (drop): Dropout(p=0.5, inplace=False)\n",
       "    (conv_classifier): Conv2d(40, 5, kernel_size=(2, 1), stride=(1, 1))\n",
       "    (softmax): LogSoftmax()\n",
       "    (squeeze): Expression(expression=squeeze_final_output) \n",
       "  ),\n",
       ")"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from skorch.callbacks import LRScheduler\n",
    "from skorch.helper import predefined_split\n",
    "\n",
    "from braindecode import EEGClassifier\n",
    "# These values we found good for shallow network:\n",
    "lr = 0.0625 * 0.01\n",
    "weight_decay = 0\n",
    "\n",
    "# For deep4 they should be:\n",
    "# lr = 1 * 0.01\n",
    "# weight_decay = 0.5 * 0.001\n",
    "\n",
    "batch_size = 32\n",
    "n_epochs = 40\n",
    "\n",
    "clf = EEGClassifier(\n",
    "    model,\n",
    "    criterion=torch.nn.NLLLoss,\n",
    "    optimizer=torch.optim.AdamW,\n",
    "    train_split=predefined_split(valid_set),  # using valid_set for validation\n",
    "    optimizer__lr=lr,\n",
    "    optimizer__weight_decay=weight_decay,\n",
    "    batch_size=batch_size,\n",
    "    callbacks=[\n",
    "        \"accuracy\", (\"lr_scheduler\", LRScheduler('CosineAnnealingLR', T_max=n_epochs - 1)),\n",
    "    ],\n",
    "    device=device,\n",
    ")\n",
    "# Model training for a specified number of epochs. `y` is None as it is already supplied\n",
    "# in the dataset.\n",
    "clf.fit(train_set, y=None, epochs=n_epochs)\n"
   ]
  },
  {
   "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": "3.7.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}