deeplift-heart_failure.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/maciej/.local/lib/python3.10/site-packages/tqdm/auto.py:22: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import torch\n",
    "from matplotlib import pyplot as plt\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn_pandas import DataFrameMapper\n",
    "from abc import ABC\n",
    "import torchtuples as tt\n",
    "from plotly.figure_factory import np\n",
    "from pycox.models import DeepHitSingle\n",
    "from sksurv.functions import StepFunction\n",
    "from survshap import SurvivalModelExplainer, PredictSurvSHAP\n",
    "import timeit\n",
    "from captum.attr import DeepLift"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Preparing data and models "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "real_data = pd.read_csv(\"data/exp3_heart_failure_dataset.csv\")\n",
    "real_data.rename(columns={\"DEATH_EVENT\": \"event\"}, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train = real_data\n",
    "df_test = df_train.sample(frac=0.2)\n",
    "df_train = df_train.drop(df_test.index)\n",
    "df_val = df_train.sample(frac=0.2)\n",
    "df_train = df_train.drop(df_val.index)\n",
    "train_column_names = [\"age\", \"creatinine_phosphokinase\", \"ejection_fraction\", \"platelets\", \"serum_creatinine\",\n",
    "                      \"serum_sodium\", \"sex\", \"smoking\"]\n",
    "max_time = real_data.time.max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "cols_standardize = [\"age\", \"creatinine_phosphokinase\", \"ejection_fraction\", \"platelets\", \"serum_creatinine\", \"serum_sodium\"]\n",
    "cols_leave = ['sex', 'smoking']\n",
    "\n",
    "standardize = [([col], StandardScaler()) for col in cols_standardize]\n",
    "leave = [(col, None) for col in cols_leave]\n",
    "\n",
    "x_mapper = DataFrameMapper(standardize + leave)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train = x_mapper.fit_transform(df_train).astype('float32')\n",
    "x_val = x_mapper.transform(df_val).astype('float32')\n",
    "x_test = x_mapper.transform(df_test).astype('float32')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train = (df_train['time'].values, df_train['event'].values)\n",
    "y_val = (df_val['time'].values, df_val['event'].values)\n",
    "\n",
    "train = (x_train, y_train)\n",
    "val = (x_val, y_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "in_features = x_train.shape[1]\n",
    "num_nodes = [32, 32]\n",
    "out_features = max_time + 1\n",
    "batch_norm = True\n",
    "dropout = 0.1\n",
    "\n",
    "network = tt.practical.MLPVanilla(in_features, num_nodes, out_features, batch_norm, dropout)\n",
    "\n",
    "class CustomDeepHitSingle(DeepHitSingle, ABC):\n",
    "    def __init__(self, net, optimizer=None, device=None, duration_index=None, alpha=0.2, sigma=0.1, loss=None):\n",
    "        super().__init__(net, optimizer=optimizer, device=device, duration_index=duration_index, alpha=alpha,\n",
    "                         sigma=sigma, loss=loss)\n",
    "\n",
    "\n",
    "    def predict_survival_function(self, X, return_array=False):\n",
    "        preds = self.predict_surv_df(X.values)\n",
    "        ret = [StepFunction(x=np.arange(0, preds.shape[0]), y=preds[column].values) for i, column in enumerate(preds)]\n",
    "        return ret\n",
    "\n",
    "model = CustomDeepHitSingle(network, optimizer=tt.optim.Adam, alpha=0.2, sigma=0.1, duration_index=np.arange(0, max_time + 1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "batch_size = 256\n",
    "lr_finder = model.lr_finder(x_train, y_train, batch_size, tolerance=3)\n",
    "_ = lr_finder.plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.optimizer.set_lr(lr_finder.get_best_lr() * 0.7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0:\t[0s / 0s],\t\ttrain_loss: 0.7620,\tval_loss: 0.6347\n",
      "1:\t[0s / 0s],\t\ttrain_loss: 0.7250,\tval_loss: 0.6329\n",
      "2:\t[0s / 0s],\t\ttrain_loss: 0.7023,\tval_loss: 0.6317\n",
      "3:\t[0s / 0s],\t\ttrain_loss: 0.6747,\tval_loss: 0.6311\n",
      "4:\t[0s / 0s],\t\ttrain_loss: 0.6504,\tval_loss: 0.6309\n",
      "5:\t[0s / 0s],\t\ttrain_loss: 0.6288,\tval_loss: 0.6309\n",
      "6:\t[0s / 0s],\t\ttrain_loss: 0.6176,\tval_loss: 0.6313\n",
      "7:\t[0s / 0s],\t\ttrain_loss: 0.5942,\tval_loss: 0.6328\n",
      "8:\t[0s / 0s],\t\ttrain_loss: 0.5833,\tval_loss: 0.6364\n",
      "9:\t[0s / 0s],\t\ttrain_loss: 0.5609,\tval_loss: 0.6427\n",
      "10:\t[0s / 0s],\t\ttrain_loss: 0.5497,\tval_loss: 0.6460\n",
      "11:\t[0s / 0s],\t\ttrain_loss: 0.5348,\tval_loss: 0.6483\n",
      "12:\t[0s / 0s],\t\ttrain_loss: 0.5002,\tval_loss: 0.6477\n",
      "13:\t[0s / 0s],\t\ttrain_loss: 0.4980,\tval_loss: 0.6471\n",
      "14:\t[0s / 0s],\t\ttrain_loss: 0.4754,\tval_loss: 0.6472\n",
      "15:\t[0s / 0s],\t\ttrain_loss: 0.4697,\tval_loss: 0.6483\n"
     ]
    }
   ],
   "source": [
    "epochs = 100\n",
    "callbacks = [tt.callbacks.EarlyStopping()]\n",
    "log = model.fit(x_train, y_train, batch_size, epochs, callbacks, val_data=val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "_ = log.plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/maciej/.local/lib/python3.10/site-packages/captum/_utils/gradient.py:57: UserWarning: Input Tensor 0 did not already require gradients, required_grads has been set automatically.\n",
      "  warnings.warn(\n",
      "/home/maciej/.local/lib/python3.10/site-packages/captum/attr/_core/deep_lift.py:304: UserWarning: Setting forward, backward hooks and attributes on non-linear\n",
      "               activations. The hooks and attributes will be removed\n",
      "            after the attribution is finished\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "ename": "RuntimeError",
     "evalue": "Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu! (when checking argument for argument mat1 in method wrapper_addmm)",
     "output_type": "error",
     "traceback": [
      "\u001B[0;31m---------------------------------------------------------------------------\u001B[0m",
      "\u001B[0;31mRuntimeError\u001B[0m                              Traceback (most recent call last)",
      "Cell \u001B[0;32mIn[15], line 10\u001B[0m\n\u001B[1;32m      7\u001B[0m pred_idx \u001B[38;5;241m=\u001B[39m model\u001B[38;5;241m.\u001B[39mpredict(x_test)\u001B[38;5;241m.\u001B[39margmax(axis\u001B[38;5;241m=\u001B[39m\u001B[38;5;241m1\u001B[39m)\n\u001B[1;32m      9\u001B[0m start \u001B[38;5;241m=\u001B[39m timeit\u001B[38;5;241m.\u001B[39mtimeit()\n\u001B[0;32m---> 10\u001B[0m attributions, delta \u001B[38;5;241m=\u001B[39m \u001B[43mdl\u001B[49m\u001B[38;5;241;43m.\u001B[39;49m\u001B[43mattribute\u001B[49m\u001B[43m(\u001B[49m\u001B[43mx_test\u001B[49m\u001B[43m,\u001B[49m\u001B[43m \u001B[49m\u001B[43mbaseline\u001B[49m\u001B[43m,\u001B[49m\u001B[43m \u001B[49m\u001B[43mtarget\u001B[49m\u001B[38;5;241;43m=\u001B[39;49m\u001B[43mpred_idx\u001B[49m\u001B[43m,\u001B[49m\u001B[43m \u001B[49m\u001B[43mreturn_convergence_delta\u001B[49m\u001B[38;5;241;43m=\u001B[39;49m\u001B[38;5;28;43;01mTrue\u001B[39;49;00m\u001B[43m)\u001B[49m\n\u001B[1;32m     11\u001B[0m end \u001B[38;5;241m=\u001B[39m timeit\u001B[38;5;241m.\u001B[39mtimeit()\n\u001B[1;32m     12\u001B[0m \u001B[38;5;28mprint\u001B[39m(end \u001B[38;5;241m-\u001B[39m start)\n",
      "File \u001B[0;32m~/.local/lib/python3.10/site-packages/captum/log/__init__.py:42\u001B[0m, in \u001B[0;36mlog_usage.<locals>._log_usage.<locals>.wrapper\u001B[0;34m(*args, **kwargs)\u001B[0m\n\u001B[1;32m     40\u001B[0m \u001B[38;5;129m@wraps\u001B[39m(func)\n\u001B[1;32m     41\u001B[0m \u001B[38;5;28;01mdef\u001B[39;00m \u001B[38;5;21mwrapper\u001B[39m(\u001B[38;5;241m*\u001B[39margs, \u001B[38;5;241m*\u001B[39m\u001B[38;5;241m*\u001B[39mkwargs):\n\u001B[0;32m---> 42\u001B[0m     \u001B[38;5;28;01mreturn\u001B[39;00m \u001B[43mfunc\u001B[49m\u001B[43m(\u001B[49m\u001B[38;5;241;43m*\u001B[39;49m\u001B[43margs\u001B[49m\u001B[43m,\u001B[49m\u001B[43m \u001B[49m\u001B[38;5;241;43m*\u001B[39;49m\u001B[38;5;241;43m*\u001B[39;49m\u001B[43mkwargs\u001B[49m\u001B[43m)\u001B[49m\n",
      "File \u001B[0;32m~/.local/lib/python3.10/site-packages/captum/attr/_core/deep_lift.py:330\u001B[0m, in \u001B[0;36mDeepLift.attribute\u001B[0;34m(self, inputs, baselines, target, additional_forward_args, return_convergence_delta, custom_attribution_func)\u001B[0m\n\u001B[1;32m    320\u001B[0m expanded_target \u001B[38;5;241m=\u001B[39m _expand_target(\n\u001B[1;32m    321\u001B[0m     target, \u001B[38;5;241m2\u001B[39m, expansion_type\u001B[38;5;241m=\u001B[39mExpansionTypes\u001B[38;5;241m.\u001B[39mrepeat\n\u001B[1;32m    322\u001B[0m )\n\u001B[1;32m    324\u001B[0m wrapped_forward_func \u001B[38;5;241m=\u001B[39m \u001B[38;5;28mself\u001B[39m\u001B[38;5;241m.\u001B[39m_construct_forward_func(\n\u001B[1;32m    325\u001B[0m     \u001B[38;5;28mself\u001B[39m\u001B[38;5;241m.\u001B[39mmodel,\n\u001B[1;32m    326\u001B[0m     (inputs, baselines),\n\u001B[1;32m    327\u001B[0m     expanded_target,\n\u001B[1;32m    328\u001B[0m     additional_forward_args,\n\u001B[1;32m    329\u001B[0m )\n\u001B[0;32m--> 330\u001B[0m gradients \u001B[38;5;241m=\u001B[39m \u001B[38;5;28;43mself\u001B[39;49m\u001B[38;5;241;43m.\u001B[39;49m\u001B[43mgradient_func\u001B[49m\u001B[43m(\u001B[49m\u001B[43mwrapped_forward_func\u001B[49m\u001B[43m,\u001B[49m\u001B[43m \u001B[49m\u001B[43minputs\u001B[49m\u001B[43m)\u001B[49m\n\u001B[1;32m    331\u001B[0m \u001B[38;5;28;01mif\u001B[39;00m custom_attribution_func \u001B[38;5;129;01mis\u001B[39;00m \u001B[38;5;28;01mNone\u001B[39;00m:\n\u001B[1;32m    332\u001B[0m     \u001B[38;5;28;01mif\u001B[39;00m \u001B[38;5;28mself\u001B[39m\u001B[38;5;241m.\u001B[39mmultiplies_by_inputs:\n",
      "File \u001B[0;32m~/.local/lib/python3.10/site-packages/captum/_utils/gradient.py:112\u001B[0m, in \u001B[0;36mcompute_gradients\u001B[0;34m(forward_fn, inputs, target_ind, additional_forward_args)\u001B[0m\n\u001B[1;32m     94\u001B[0m \u001B[38;5;124mr\u001B[39m\u001B[38;5;124;03m\"\"\"\u001B[39;00m\n\u001B[1;32m     95\u001B[0m \u001B[38;5;124;03mComputes gradients of the output with respect to inputs for an\u001B[39;00m\n\u001B[1;32m     96\u001B[0m \u001B[38;5;124;03marbitrary forward function.\u001B[39;00m\n\u001B[0;32m   (...)\u001B[0m\n\u001B[1;32m    108\u001B[0m \u001B[38;5;124;03m                arguments) if no additional arguments are required\u001B[39;00m\n\u001B[1;32m    109\u001B[0m \u001B[38;5;124;03m\"\"\"\u001B[39;00m\n\u001B[1;32m    110\u001B[0m \u001B[38;5;28;01mwith\u001B[39;00m torch\u001B[38;5;241m.\u001B[39mautograd\u001B[38;5;241m.\u001B[39mset_grad_enabled(\u001B[38;5;28;01mTrue\u001B[39;00m):\n\u001B[1;32m    111\u001B[0m     \u001B[38;5;66;03m# runs forward pass\u001B[39;00m\n\u001B[0;32m--> 112\u001B[0m     outputs \u001B[38;5;241m=\u001B[39m \u001B[43m_run_forward\u001B[49m\u001B[43m(\u001B[49m\u001B[43mforward_fn\u001B[49m\u001B[43m,\u001B[49m\u001B[43m \u001B[49m\u001B[43minputs\u001B[49m\u001B[43m,\u001B[49m\u001B[43m \u001B[49m\u001B[43mtarget_ind\u001B[49m\u001B[43m,\u001B[49m\u001B[43m \u001B[49m\u001B[43madditional_forward_args\u001B[49m\u001B[43m)\u001B[49m\n\u001B[1;32m    113\u001B[0m     \u001B[38;5;28;01massert\u001B[39;00m outputs[\u001B[38;5;241m0\u001B[39m]\u001B[38;5;241m.\u001B[39mnumel() \u001B[38;5;241m==\u001B[39m \u001B[38;5;241m1\u001B[39m, (\n\u001B[1;32m    114\u001B[0m         \u001B[38;5;124m\"\u001B[39m\u001B[38;5;124mTarget not provided when necessary, cannot\u001B[39m\u001B[38;5;124m\"\u001B[39m\n\u001B[1;32m    115\u001B[0m         \u001B[38;5;124m\"\u001B[39m\u001B[38;5;124m take gradient with respect to multiple outputs.\u001B[39m\u001B[38;5;124m\"\u001B[39m\n\u001B[1;32m    116\u001B[0m     )\n\u001B[1;32m    117\u001B[0m     \u001B[38;5;66;03m# torch.unbind(forward_out) is a list of scalar tensor tuples and\u001B[39;00m\n\u001B[1;32m    118\u001B[0m     \u001B[38;5;66;03m# contains batch_size * #steps elements\u001B[39;00m\n",
      "File \u001B[0;32m~/.local/lib/python3.10/site-packages/captum/_utils/common.py:474\u001B[0m, in \u001B[0;36m_run_forward\u001B[0;34m(forward_func, inputs, target, additional_forward_args)\u001B[0m\n\u001B[1;32m    472\u001B[0m forward_func_args \u001B[38;5;241m=\u001B[39m signature(forward_func)\u001B[38;5;241m.\u001B[39mparameters\n\u001B[1;32m    473\u001B[0m \u001B[38;5;28;01mif\u001B[39;00m \u001B[38;5;28mlen\u001B[39m(forward_func_args) \u001B[38;5;241m==\u001B[39m \u001B[38;5;241m0\u001B[39m:\n\u001B[0;32m--> 474\u001B[0m     output \u001B[38;5;241m=\u001B[39m \u001B[43mforward_func\u001B[49m\u001B[43m(\u001B[49m\u001B[43m)\u001B[49m\n\u001B[1;32m    475\u001B[0m     \u001B[38;5;28;01mreturn\u001B[39;00m output \u001B[38;5;28;01mif\u001B[39;00m target \u001B[38;5;129;01mis\u001B[39;00m \u001B[38;5;28;01mNone\u001B[39;00m \u001B[38;5;28;01melse\u001B[39;00m _select_targets(output, target)\n\u001B[1;32m    477\u001B[0m \u001B[38;5;66;03m# make everything a tuple so that it is easy to unpack without\u001B[39;00m\n\u001B[1;32m    478\u001B[0m \u001B[38;5;66;03m# using if-statements\u001B[39;00m\n",
      "File \u001B[0;32m~/.local/lib/python3.10/site-packages/captum/attr/_core/deep_lift.py:369\u001B[0m, in \u001B[0;36mDeepLift._construct_forward_func.<locals>.forward_fn\u001B[0;34m()\u001B[0m\n\u001B[1;32m    368\u001B[0m \u001B[38;5;28;01mdef\u001B[39;00m \u001B[38;5;21mforward_fn\u001B[39m():\n\u001B[0;32m--> 369\u001B[0m     model_out \u001B[38;5;241m=\u001B[39m \u001B[43m_run_forward\u001B[49m\u001B[43m(\u001B[49m\n\u001B[1;32m    370\u001B[0m \u001B[43m        \u001B[49m\u001B[43mforward_func\u001B[49m\u001B[43m,\u001B[49m\u001B[43m \u001B[49m\u001B[43minputs\u001B[49m\u001B[43m,\u001B[49m\u001B[43m \u001B[49m\u001B[38;5;28;43;01mNone\u001B[39;49;00m\u001B[43m,\u001B[49m\u001B[43m \u001B[49m\u001B[43madditional_forward_args\u001B[49m\n\u001B[1;32m    371\u001B[0m \u001B[43m    \u001B[49m\u001B[43m)\u001B[49m\n\u001B[1;32m    372\u001B[0m     \u001B[38;5;28;01mreturn\u001B[39;00m _select_targets(\n\u001B[1;32m    373\u001B[0m         torch\u001B[38;5;241m.\u001B[39mcat((model_out[:, \u001B[38;5;241m0\u001B[39m], model_out[:, \u001B[38;5;241m1\u001B[39m])), target\n\u001B[1;32m    374\u001B[0m     )\n",
      "File \u001B[0;32m~/.local/lib/python3.10/site-packages/captum/_utils/common.py:482\u001B[0m, in \u001B[0;36m_run_forward\u001B[0;34m(forward_func, inputs, target, additional_forward_args)\u001B[0m\n\u001B[1;32m    479\u001B[0m inputs \u001B[38;5;241m=\u001B[39m _format_inputs(inputs)\n\u001B[1;32m    480\u001B[0m additional_forward_args \u001B[38;5;241m=\u001B[39m _format_additional_forward_args(additional_forward_args)\n\u001B[0;32m--> 482\u001B[0m output \u001B[38;5;241m=\u001B[39m \u001B[43mforward_func\u001B[49m\u001B[43m(\u001B[49m\n\u001B[1;32m    483\u001B[0m \u001B[43m    \u001B[49m\u001B[38;5;241;43m*\u001B[39;49m\u001B[43m(\u001B[49m\u001B[38;5;241;43m*\u001B[39;49m\u001B[43minputs\u001B[49m\u001B[43m,\u001B[49m\u001B[43m \u001B[49m\u001B[38;5;241;43m*\u001B[39;49m\u001B[43madditional_forward_args\u001B[49m\u001B[43m)\u001B[49m\n\u001B[1;32m    484\u001B[0m \u001B[43m    \u001B[49m\u001B[38;5;28;43;01mif\u001B[39;49;00m\u001B[43m \u001B[49m\u001B[43madditional_forward_args\u001B[49m\u001B[43m \u001B[49m\u001B[38;5;129;43;01mis\u001B[39;49;00m\u001B[43m \u001B[49m\u001B[38;5;129;43;01mnot\u001B[39;49;00m\u001B[43m \u001B[49m\u001B[38;5;28;43;01mNone\u001B[39;49;00m\n\u001B[1;32m    485\u001B[0m \u001B[43m    \u001B[49m\u001B[38;5;28;43;01melse\u001B[39;49;00m\u001B[43m \u001B[49m\u001B[43minputs\u001B[49m\n\u001B[1;32m    486\u001B[0m \u001B[43m\u001B[49m\u001B[43m)\u001B[49m\n\u001B[1;32m    487\u001B[0m \u001B[38;5;28;01mreturn\u001B[39;00m _select_targets(output, target)\n",
      "File \u001B[0;32m~/.local/lib/python3.10/site-packages/torch/nn/modules/module.py:1212\u001B[0m, in \u001B[0;36mModule._call_impl\u001B[0;34m(self, *input, **kwargs)\u001B[0m\n\u001B[1;32m   1209\u001B[0m     bw_hook \u001B[38;5;241m=\u001B[39m hooks\u001B[38;5;241m.\u001B[39mBackwardHook(\u001B[38;5;28mself\u001B[39m, full_backward_hooks)\n\u001B[1;32m   1210\u001B[0m     \u001B[38;5;28minput\u001B[39m \u001B[38;5;241m=\u001B[39m bw_hook\u001B[38;5;241m.\u001B[39msetup_input_hook(\u001B[38;5;28minput\u001B[39m)\n\u001B[0;32m-> 1212\u001B[0m result \u001B[38;5;241m=\u001B[39m \u001B[43mforward_call\u001B[49m\u001B[43m(\u001B[49m\u001B[38;5;241;43m*\u001B[39;49m\u001B[38;5;28;43minput\u001B[39;49m\u001B[43m,\u001B[49m\u001B[43m \u001B[49m\u001B[38;5;241;43m*\u001B[39;49m\u001B[38;5;241;43m*\u001B[39;49m\u001B[43mkwargs\u001B[49m\u001B[43m)\u001B[49m\n\u001B[1;32m   1213\u001B[0m \u001B[38;5;28;01mif\u001B[39;00m _global_forward_hooks \u001B[38;5;129;01mor\u001B[39;00m \u001B[38;5;28mself\u001B[39m\u001B[38;5;241m.\u001B[39m_forward_hooks:\n\u001B[1;32m   1214\u001B[0m     \u001B[38;5;28;01mfor\u001B[39;00m hook \u001B[38;5;129;01min\u001B[39;00m (\u001B[38;5;241m*\u001B[39m_global_forward_hooks\u001B[38;5;241m.\u001B[39mvalues(), \u001B[38;5;241m*\u001B[39m\u001B[38;5;28mself\u001B[39m\u001B[38;5;241m.\u001B[39m_forward_hooks\u001B[38;5;241m.\u001B[39mvalues()):\n",
      "File \u001B[0;32m~/.local/lib/python3.10/site-packages/torchtuples/practical.py:84\u001B[0m, in \u001B[0;36mMLPVanilla.forward\u001B[0;34m(self, input)\u001B[0m\n\u001B[1;32m     83\u001B[0m \u001B[38;5;28;01mdef\u001B[39;00m \u001B[38;5;21mforward\u001B[39m(\u001B[38;5;28mself\u001B[39m, \u001B[38;5;28minput\u001B[39m):\n\u001B[0;32m---> 84\u001B[0m     \u001B[38;5;28;01mreturn\u001B[39;00m \u001B[38;5;28;43mself\u001B[39;49m\u001B[38;5;241;43m.\u001B[39;49m\u001B[43mnet\u001B[49m\u001B[43m(\u001B[49m\u001B[38;5;28;43minput\u001B[39;49m\u001B[43m)\u001B[49m\n",
      "File \u001B[0;32m~/.local/lib/python3.10/site-packages/torch/nn/modules/module.py:1194\u001B[0m, in \u001B[0;36mModule._call_impl\u001B[0;34m(self, *input, **kwargs)\u001B[0m\n\u001B[1;32m   1190\u001B[0m \u001B[38;5;66;03m# If we don't have any hooks, we want to skip the rest of the logic in\u001B[39;00m\n\u001B[1;32m   1191\u001B[0m \u001B[38;5;66;03m# this function, and just call forward.\u001B[39;00m\n\u001B[1;32m   1192\u001B[0m \u001B[38;5;28;01mif\u001B[39;00m \u001B[38;5;129;01mnot\u001B[39;00m (\u001B[38;5;28mself\u001B[39m\u001B[38;5;241m.\u001B[39m_backward_hooks \u001B[38;5;129;01mor\u001B[39;00m \u001B[38;5;28mself\u001B[39m\u001B[38;5;241m.\u001B[39m_forward_hooks \u001B[38;5;129;01mor\u001B[39;00m \u001B[38;5;28mself\u001B[39m\u001B[38;5;241m.\u001B[39m_forward_pre_hooks \u001B[38;5;129;01mor\u001B[39;00m _global_backward_hooks\n\u001B[1;32m   1193\u001B[0m         \u001B[38;5;129;01mor\u001B[39;00m _global_forward_hooks \u001B[38;5;129;01mor\u001B[39;00m _global_forward_pre_hooks):\n\u001B[0;32m-> 1194\u001B[0m     \u001B[38;5;28;01mreturn\u001B[39;00m \u001B[43mforward_call\u001B[49m\u001B[43m(\u001B[49m\u001B[38;5;241;43m*\u001B[39;49m\u001B[38;5;28;43minput\u001B[39;49m\u001B[43m,\u001B[49m\u001B[43m \u001B[49m\u001B[38;5;241;43m*\u001B[39;49m\u001B[38;5;241;43m*\u001B[39;49m\u001B[43mkwargs\u001B[49m\u001B[43m)\u001B[49m\n\u001B[1;32m   1195\u001B[0m \u001B[38;5;66;03m# Do not call functions when jit is used\u001B[39;00m\n\u001B[1;32m   1196\u001B[0m full_backward_hooks, non_full_backward_hooks \u001B[38;5;241m=\u001B[39m [], []\n",
      "File \u001B[0;32m~/.local/lib/python3.10/site-packages/torch/nn/modules/container.py:204\u001B[0m, in \u001B[0;36mSequential.forward\u001B[0;34m(self, input)\u001B[0m\n\u001B[1;32m    202\u001B[0m \u001B[38;5;28;01mdef\u001B[39;00m \u001B[38;5;21mforward\u001B[39m(\u001B[38;5;28mself\u001B[39m, \u001B[38;5;28minput\u001B[39m):\n\u001B[1;32m    203\u001B[0m     \u001B[38;5;28;01mfor\u001B[39;00m module \u001B[38;5;129;01min\u001B[39;00m \u001B[38;5;28mself\u001B[39m:\n\u001B[0;32m--> 204\u001B[0m         \u001B[38;5;28minput\u001B[39m \u001B[38;5;241m=\u001B[39m \u001B[43mmodule\u001B[49m\u001B[43m(\u001B[49m\u001B[38;5;28;43minput\u001B[39;49m\u001B[43m)\u001B[49m\n\u001B[1;32m    205\u001B[0m     \u001B[38;5;28;01mreturn\u001B[39;00m \u001B[38;5;28minput\u001B[39m\n",
      "File \u001B[0;32m~/.local/lib/python3.10/site-packages/torch/nn/modules/module.py:1194\u001B[0m, in \u001B[0;36mModule._call_impl\u001B[0;34m(self, *input, **kwargs)\u001B[0m\n\u001B[1;32m   1190\u001B[0m \u001B[38;5;66;03m# If we don't have any hooks, we want to skip the rest of the logic in\u001B[39;00m\n\u001B[1;32m   1191\u001B[0m \u001B[38;5;66;03m# this function, and just call forward.\u001B[39;00m\n\u001B[1;32m   1192\u001B[0m \u001B[38;5;28;01mif\u001B[39;00m \u001B[38;5;129;01mnot\u001B[39;00m (\u001B[38;5;28mself\u001B[39m\u001B[38;5;241m.\u001B[39m_backward_hooks \u001B[38;5;129;01mor\u001B[39;00m \u001B[38;5;28mself\u001B[39m\u001B[38;5;241m.\u001B[39m_forward_hooks \u001B[38;5;129;01mor\u001B[39;00m \u001B[38;5;28mself\u001B[39m\u001B[38;5;241m.\u001B[39m_forward_pre_hooks \u001B[38;5;129;01mor\u001B[39;00m _global_backward_hooks\n\u001B[1;32m   1193\u001B[0m         \u001B[38;5;129;01mor\u001B[39;00m _global_forward_hooks \u001B[38;5;129;01mor\u001B[39;00m _global_forward_pre_hooks):\n\u001B[0;32m-> 1194\u001B[0m     \u001B[38;5;28;01mreturn\u001B[39;00m \u001B[43mforward_call\u001B[49m\u001B[43m(\u001B[49m\u001B[38;5;241;43m*\u001B[39;49m\u001B[38;5;28;43minput\u001B[39;49m\u001B[43m,\u001B[49m\u001B[43m \u001B[49m\u001B[38;5;241;43m*\u001B[39;49m\u001B[38;5;241;43m*\u001B[39;49m\u001B[43mkwargs\u001B[49m\u001B[43m)\u001B[49m\n\u001B[1;32m   1195\u001B[0m \u001B[38;5;66;03m# Do not call functions when jit is used\u001B[39;00m\n\u001B[1;32m   1196\u001B[0m full_backward_hooks, non_full_backward_hooks \u001B[38;5;241m=\u001B[39m [], []\n",
      "File \u001B[0;32m~/.local/lib/python3.10/site-packages/torchtuples/practical.py:59\u001B[0m, in \u001B[0;36mDenseVanillaBlock.forward\u001B[0;34m(self, input)\u001B[0m\n\u001B[1;32m     58\u001B[0m \u001B[38;5;28;01mdef\u001B[39;00m \u001B[38;5;21mforward\u001B[39m(\u001B[38;5;28mself\u001B[39m, \u001B[38;5;28minput\u001B[39m):\n\u001B[0;32m---> 59\u001B[0m     \u001B[38;5;28minput\u001B[39m \u001B[38;5;241m=\u001B[39m \u001B[38;5;28mself\u001B[39m\u001B[38;5;241m.\u001B[39mactivation(\u001B[38;5;28;43mself\u001B[39;49m\u001B[38;5;241;43m.\u001B[39;49m\u001B[43mlinear\u001B[49m\u001B[43m(\u001B[49m\u001B[38;5;28;43minput\u001B[39;49m\u001B[43m)\u001B[49m)\n\u001B[1;32m     60\u001B[0m     \u001B[38;5;28;01mif\u001B[39;00m \u001B[38;5;28mself\u001B[39m\u001B[38;5;241m.\u001B[39mbatch_norm:\n\u001B[1;32m     61\u001B[0m         \u001B[38;5;28minput\u001B[39m \u001B[38;5;241m=\u001B[39m \u001B[38;5;28mself\u001B[39m\u001B[38;5;241m.\u001B[39mbatch_norm(\u001B[38;5;28minput\u001B[39m)\n",
      "File \u001B[0;32m~/.local/lib/python3.10/site-packages/torch/nn/modules/module.py:1194\u001B[0m, in \u001B[0;36mModule._call_impl\u001B[0;34m(self, *input, **kwargs)\u001B[0m\n\u001B[1;32m   1190\u001B[0m \u001B[38;5;66;03m# If we don't have any hooks, we want to skip the rest of the logic in\u001B[39;00m\n\u001B[1;32m   1191\u001B[0m \u001B[38;5;66;03m# this function, and just call forward.\u001B[39;00m\n\u001B[1;32m   1192\u001B[0m \u001B[38;5;28;01mif\u001B[39;00m \u001B[38;5;129;01mnot\u001B[39;00m (\u001B[38;5;28mself\u001B[39m\u001B[38;5;241m.\u001B[39m_backward_hooks \u001B[38;5;129;01mor\u001B[39;00m \u001B[38;5;28mself\u001B[39m\u001B[38;5;241m.\u001B[39m_forward_hooks \u001B[38;5;129;01mor\u001B[39;00m \u001B[38;5;28mself\u001B[39m\u001B[38;5;241m.\u001B[39m_forward_pre_hooks \u001B[38;5;129;01mor\u001B[39;00m _global_backward_hooks\n\u001B[1;32m   1193\u001B[0m         \u001B[38;5;129;01mor\u001B[39;00m _global_forward_hooks \u001B[38;5;129;01mor\u001B[39;00m _global_forward_pre_hooks):\n\u001B[0;32m-> 1194\u001B[0m     \u001B[38;5;28;01mreturn\u001B[39;00m \u001B[43mforward_call\u001B[49m\u001B[43m(\u001B[49m\u001B[38;5;241;43m*\u001B[39;49m\u001B[38;5;28;43minput\u001B[39;49m\u001B[43m,\u001B[49m\u001B[43m \u001B[49m\u001B[38;5;241;43m*\u001B[39;49m\u001B[38;5;241;43m*\u001B[39;49m\u001B[43mkwargs\u001B[49m\u001B[43m)\u001B[49m\n\u001B[1;32m   1195\u001B[0m \u001B[38;5;66;03m# Do not call functions when jit is used\u001B[39;00m\n\u001B[1;32m   1196\u001B[0m full_backward_hooks, non_full_backward_hooks \u001B[38;5;241m=\u001B[39m [], []\n",
      "File \u001B[0;32m~/.local/lib/python3.10/site-packages/torch/nn/modules/linear.py:114\u001B[0m, in \u001B[0;36mLinear.forward\u001B[0;34m(self, input)\u001B[0m\n\u001B[1;32m    113\u001B[0m \u001B[38;5;28;01mdef\u001B[39;00m \u001B[38;5;21mforward\u001B[39m(\u001B[38;5;28mself\u001B[39m, \u001B[38;5;28minput\u001B[39m: Tensor) \u001B[38;5;241m-\u001B[39m\u001B[38;5;241m>\u001B[39m Tensor:\n\u001B[0;32m--> 114\u001B[0m     \u001B[38;5;28;01mreturn\u001B[39;00m \u001B[43mF\u001B[49m\u001B[38;5;241;43m.\u001B[39;49m\u001B[43mlinear\u001B[49m\u001B[43m(\u001B[49m\u001B[38;5;28;43minput\u001B[39;49m\u001B[43m,\u001B[49m\u001B[43m \u001B[49m\u001B[38;5;28;43mself\u001B[39;49m\u001B[38;5;241;43m.\u001B[39;49m\u001B[43mweight\u001B[49m\u001B[43m,\u001B[49m\u001B[43m \u001B[49m\u001B[38;5;28;43mself\u001B[39;49m\u001B[38;5;241;43m.\u001B[39;49m\u001B[43mbias\u001B[49m\u001B[43m)\u001B[49m\n",
      "\u001B[0;31mRuntimeError\u001B[0m: Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu! (when checking argument for argument mat1 in method wrapper_addmm)"
     ]
    }
   ],
   "source": [
    "y_dtype = np.dtype({'names':['event', 'time'], 'formats':[np.uint8, np.uint8]})\n",
    "y = np.array([tuple(row) for row in np.stack((y_train[1], y_train[0])).transpose(1, 0)], dtype=y_dtype)\n",
    "dl = DeepLift(model.net)\n",
    "time_array = []\n",
    "x_test = torch.Tensor(x_test)\n",
    "baseline = torch.zeros_like(x_test)\n",
    "pred_idx = model.predict(x_test).argmax(axis=1)\n",
    "\n",
    "start = timeit.timeit()\n",
    "attributions, delta = dl.attribute(x_test, baseline, target=pred_idx, return_convergence_delta=True)\n",
    "end = timeit.timeit()\n",
    "print(end - start)\n",
    "time_array.append(end - start)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([-0.3553, -0.0614, -0.0621,  0.3124, -0.2414, -0.1846,  0.2679, -0.0991],\n",
      "       grad_fn=<SelectBackward0>)\n",
      "tensor(-1.2893)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax= plt.subplot()\n",
    "plt.plot(train_column_names, attributions[0].detach().numpy(), 'gd')\n",
    "plt.setp(ax.get_xticklabels(), rotation=30, ha='right')\n",
    "print(attributions[0])\n",
    "print(delta[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.10.7"
  },
  "vscode": {
   "interpreter": {
    "hash": "2e89d17bf3f53615b213f4c00662e1677a8885f31ece09e136535e9b43ddada0"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}