{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "9e4fa299",
   "metadata": {},
   "source": [
    "## Regression With Fusion Data\n",
    "\n",
    "The aim of this notebook is to regress the time until a disruption occurs given a current state in time.\n",
    "\n",
    "\n",
    "In reality, our dataset is comprised of multiple shots, where each shot has individual time readings.\n",
    "\n",
    "| Shot Number     | Time Step            | Feature 1       | Feature 2       |\n",
    "|-----------------|----------------------|-----------------|-----------------|\n",
    "| 1               |       $t_0$          | -               | -               |\n",
    "| 1               |       $t_1$          | -               | -               |\n",
    "| 1               |       $t_2$          | -               | -               |\n",
    "| 1               |       ...            | -               | -               |\n",
    "| 1               |       $t_m$          | -               | -               |\n",
    "...\n",
    "| Shot Number     | Time Step            | Feature 1       | Feature 2       |\n",
    "|-----------------|----------------------|-----------------|-----------------|\n",
    "| N               |       $t_0$          | -               | -               |\n",
    "| N               |       $t_1$          | -               | -               |\n",
    "| N               |       $t_2$          | -               | -               |\n",
    "| N               |       ...            | -               | -               |\n",
    "| N               |       $t_m$          | -               | -               |\n",
    "\n",
    "In good approximation, we can treat individual readings in time as independent measurements and regress directly upon them.\n",
    "\n",
    "You can think of this as a sort of state estimation:\n",
    "\n",
    "Given state $S_i$, independent of $S_{j \\neq i}$, we want to regress how long the stability of our fusion process will last.\n",
    "\n",
    "Lets load our data and take a look at the features. We also need to drop some features as these will bias the regression.\n",
    "\n",
    "We will drop: \n",
    "\n",
    "__Disruptive__ - this is a binary label indicating whether or not a disruption has occured.\n",
    "\n",
    "__Shot__ - this is an indexing variable for individual shots.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 251,
   "id": "50f5c4ec",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>z_error</th>\n",
       "      <th>radiated_fraction</th>\n",
       "      <th>beta_p</th>\n",
       "      <th>lower_gap</th>\n",
       "      <th>n_e</th>\n",
       "      <th>ssep</th>\n",
       "      <th>Wmhd</th>\n",
       "      <th>p_icrf</th>\n",
       "      <th>upper_gap</th>\n",
       "      <th>beta_n</th>\n",
       "      <th>...</th>\n",
       "      <th>time_until_disrupt</th>\n",
       "      <th>n_equal_1_normalized</th>\n",
       "      <th>Greenwald_fraction</th>\n",
       "      <th>li</th>\n",
       "      <th>v_loop</th>\n",
       "      <th>time</th>\n",
       "      <th>q95</th>\n",
       "      <th>p_oh</th>\n",
       "      <th>p_rad</th>\n",
       "      <th>ip_error</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>148403</th>\n",
       "      <td>-0.000308</td>\n",
       "      <td>0.348539</td>\n",
       "      <td>0.294560</td>\n",
       "      <td>0.049893</td>\n",
       "      <td>2.203347e+20</td>\n",
       "      <td>-0.025936</td>\n",
       "      <td>21782.460938</td>\n",
       "      <td>3.840209e+01</td>\n",
       "      <td>0.093674</td>\n",
       "      <td>0.742089</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0160</td>\n",
       "      <td>0.001137</td>\n",
       "      <td>0.604698</td>\n",
       "      <td>1.259797</td>\n",
       "      <td>-2.511963</td>\n",
       "      <td>1.0899</td>\n",
       "      <td>2.940394</td>\n",
       "      <td>1.125851e+06</td>\n",
       "      <td>392415.960065</td>\n",
       "      <td>271289.0625</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>169792</th>\n",
       "      <td>0.000219</td>\n",
       "      <td>0.018243</td>\n",
       "      <td>0.257212</td>\n",
       "      <td>0.078944</td>\n",
       "      <td>1.575422e+20</td>\n",
       "      <td>-0.013643</td>\n",
       "      <td>44130.542969</td>\n",
       "      <td>3.537796e+06</td>\n",
       "      <td>0.103161</td>\n",
       "      <td>0.439488</td>\n",
       "      <td>...</td>\n",
       "      <td>0.2849</td>\n",
       "      <td>0.001021</td>\n",
       "      <td>0.309502</td>\n",
       "      <td>1.414968</td>\n",
       "      <td>-0.851196</td>\n",
       "      <td>1.4200</td>\n",
       "      <td>4.553412</td>\n",
       "      <td>4.030026e+05</td>\n",
       "      <td>71890.617371</td>\n",
       "      <td>104532.6250</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>275818</th>\n",
       "      <td>0.000992</td>\n",
       "      <td>0.415253</td>\n",
       "      <td>0.135165</td>\n",
       "      <td>0.044478</td>\n",
       "      <td>7.909377e+19</td>\n",
       "      <td>-0.018565</td>\n",
       "      <td>23948.123047</td>\n",
       "      <td>3.787166e+01</td>\n",
       "      <td>0.097886</td>\n",
       "      <td>0.223583</td>\n",
       "      <td>...</td>\n",
       "      <td>1.2241</td>\n",
       "      <td>0.000308</td>\n",
       "      <td>0.148950</td>\n",
       "      <td>1.505030</td>\n",
       "      <td>-0.808014</td>\n",
       "      <td>0.5200</td>\n",
       "      <td>4.437774</td>\n",
       "      <td>8.496335e+05</td>\n",
       "      <td>352828.804445</td>\n",
       "      <td>-13796.6875</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44013</th>\n",
       "      <td>0.000420</td>\n",
       "      <td>-0.329507</td>\n",
       "      <td>0.092925</td>\n",
       "      <td>0.052791</td>\n",
       "      <td>7.200679e+19</td>\n",
       "      <td>0.010766</td>\n",
       "      <td>26898.298828</td>\n",
       "      <td>1.013809e+02</td>\n",
       "      <td>0.087245</td>\n",
       "      <td>0.192169</td>\n",
       "      <td>...</td>\n",
       "      <td>0.1745</td>\n",
       "      <td>0.000827</td>\n",
       "      <td>0.106209</td>\n",
       "      <td>1.382588</td>\n",
       "      <td>0.796661</td>\n",
       "      <td>0.6400</td>\n",
       "      <td>3.791784</td>\n",
       "      <td>-1.451691e+06</td>\n",
       "      <td>478308.292191</td>\n",
       "      <td>-234166.7500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22454</th>\n",
       "      <td>0.001223</td>\n",
       "      <td>0.350599</td>\n",
       "      <td>0.189295</td>\n",
       "      <td>0.055474</td>\n",
       "      <td>1.547137e+20</td>\n",
       "      <td>-0.007201</td>\n",
       "      <td>53719.265625</td>\n",
       "      <td>7.274901e+05</td>\n",
       "      <td>0.108699</td>\n",
       "      <td>0.408583</td>\n",
       "      <td>...</td>\n",
       "      <td>0.8579</td>\n",
       "      <td>0.000525</td>\n",
       "      <td>0.233198</td>\n",
       "      <td>1.368598</td>\n",
       "      <td>-1.063354</td>\n",
       "      <td>0.6800</td>\n",
       "      <td>3.532992</td>\n",
       "      <td>1.178322e+06</td>\n",
       "      <td>668175.674777</td>\n",
       "      <td>-209515.4375</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 29 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         z_error  radiated_fraction    beta_p  lower_gap           n_e  \\\n",
       "148403 -0.000308           0.348539  0.294560   0.049893  2.203347e+20   \n",
       "169792  0.000219           0.018243  0.257212   0.078944  1.575422e+20   \n",
       "275818  0.000992           0.415253  0.135165   0.044478  7.909377e+19   \n",
       "44013   0.000420          -0.329507  0.092925   0.052791  7.200679e+19   \n",
       "22454   0.001223           0.350599  0.189295   0.055474  1.547137e+20   \n",
       "\n",
       "            ssep          Wmhd        p_icrf  upper_gap    beta_n  ...  \\\n",
       "148403 -0.025936  21782.460938  3.840209e+01   0.093674  0.742089  ...   \n",
       "169792 -0.013643  44130.542969  3.537796e+06   0.103161  0.439488  ...   \n",
       "275818 -0.018565  23948.123047  3.787166e+01   0.097886  0.223583  ...   \n",
       "44013   0.010766  26898.298828  1.013809e+02   0.087245  0.192169  ...   \n",
       "22454  -0.007201  53719.265625  7.274901e+05   0.108699  0.408583  ...   \n",
       "\n",
       "        time_until_disrupt  n_equal_1_normalized  Greenwald_fraction  \\\n",
       "148403              0.0160              0.001137            0.604698   \n",
       "169792              0.2849              0.001021            0.309502   \n",
       "275818              1.2241              0.000308            0.148950   \n",
       "44013               0.1745              0.000827            0.106209   \n",
       "22454               0.8579              0.000525            0.233198   \n",
       "\n",
       "              li    v_loop    time       q95          p_oh          p_rad  \\\n",
       "148403  1.259797 -2.511963  1.0899  2.940394  1.125851e+06  392415.960065   \n",
       "169792  1.414968 -0.851196  1.4200  4.553412  4.030026e+05   71890.617371   \n",
       "275818  1.505030 -0.808014  0.5200  4.437774  8.496335e+05  352828.804445   \n",
       "44013   1.382588  0.796661  0.6400  3.791784 -1.451691e+06  478308.292191   \n",
       "22454   1.368598 -1.063354  0.6800  3.532992  1.178322e+06  668175.674777   \n",
       "\n",
       "           ip_error  \n",
       "148403  271289.0625  \n",
       "169792  104532.6250  \n",
       "275818  -13796.6875  \n",
       "44013  -234166.7500  \n",
       "22454  -209515.4375  \n",
       "\n",
       "[5 rows x 29 columns]"
      ]
     },
     "execution_count": 251,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "df = pd.read_csv(\"full_db-complete_classified_0.1s.csv\",sep=',',index_col=None)\n",
    "df_regression = df.drop(columns=['disruptive','shot']) # Drop disruptive - binary la\n",
    "df_regression = df_regression[df.time_until_disrupt  > 0]\n",
    "df_regression = df_regression.sample(frac=1.0) # Random shuffle shots\n",
    "df_regression.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3ecc6410",
   "metadata": {},
   "source": [
    "Lets visualize our data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 203,
   "id": "5392974b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[<AxesSubplot:title={'center':'z_error'}>,\n",
       "        <AxesSubplot:title={'center':'radiated_fraction'}>,\n",
       "        <AxesSubplot:title={'center':'beta_p'}>,\n",
       "        <AxesSubplot:title={'center':'lower_gap'}>,\n",
       "        <AxesSubplot:title={'center':'n_e'}>],\n",
       "       [<AxesSubplot:title={'center':'ssep'}>,\n",
       "        <AxesSubplot:title={'center':'Wmhd'}>,\n",
       "        <AxesSubplot:title={'center':'p_icrf'}>,\n",
       "        <AxesSubplot:title={'center':'upper_gap'}>,\n",
       "        <AxesSubplot:title={'center':'beta_n'}>],\n",
       "       [<AxesSubplot:title={'center':'zcur'}>,\n",
       "        <AxesSubplot:title={'center':'q0'}>,\n",
       "        <AxesSubplot:title={'center':'qstar'}>,\n",
       "        <AxesSubplot:title={'center':'n_over_ncrit'}>,\n",
       "        <AxesSubplot:title={'center':'ip'}>],\n",
       "       [<AxesSubplot:title={'center':'kappa'}>,\n",
       "        <AxesSubplot:title={'center':'dipprog_dt'}>,\n",
       "        <AxesSubplot:title={'center':'p_lh'}>,\n",
       "        <AxesSubplot:title={'center':'n_equal_1_mode'}>,\n",
       "        <AxesSubplot:title={'center':'time_until_disrupt'}>],\n",
       "       [<AxesSubplot:title={'center':'n_equal_1_normalized'}>,\n",
       "        <AxesSubplot:title={'center':'Greenwald_fraction'}>,\n",
       "        <AxesSubplot:title={'center':'li'}>,\n",
       "        <AxesSubplot:title={'center':'v_loop'}>,\n",
       "        <AxesSubplot:title={'center':'time'}>],\n",
       "       [<AxesSubplot:title={'center':'q95'}>,\n",
       "        <AxesSubplot:title={'center':'p_oh'}>,\n",
       "        <AxesSubplot:title={'center':'p_rad'}>,\n",
       "        <AxesSubplot:title={'center':'ip_error'}>, <AxesSubplot:>]],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 203,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1500x1500 with 30 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_regression.hist(figsize=(15,15),bins=50,density=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b647f87b",
   "metadata": {},
   "source": [
    "### Preprocessing\n",
    "\n",
    "We will create a train test split for this dataset, using the traditional 70/15/15% split.\n",
    "\n",
    "We are also going to add a flag for taking the logarithm of time."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 252,
   "id": "5f40d9ea",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total dataset size:  63501\n",
      "Number of Training data points:  44450\n",
      "Number of Validation data points:  9525\n",
      "Number of Testing data points:  9526\n"
     ]
    }
   ],
   "source": [
    "from sklearn.preprocessing import MinMaxScaler\n",
    "\n",
    "############# Important Flag ##############\n",
    "log_time = False\n",
    "###########################################\n",
    "\n",
    "def create_train_test_split(dataset,log_time=False):\n",
    "    \n",
    "    def return_x_y(dataframe,log_time=False):\n",
    "        x = dataframe.drop(columns=['time_until_disrupt']).to_numpy()\n",
    "        if log_time:\n",
    "            y = np.log(dataframe['time_until_disrupt'].values)\n",
    "        else:\n",
    "            y = dataframe['time_until_disrupt'].values\n",
    "        \n",
    "        return x,y \n",
    "    \n",
    "    Nt = int(0.7 * len(dataset))\n",
    "    train = dataset[:Nt]\n",
    "    test_val = dataset[Nt:]\n",
    "    \n",
    "    Nv = int(0.5 *len(test_val))\n",
    "    val = test_val[:Nv]\n",
    "    test = test_val[Nv:]\n",
    "    print(\"Total dataset size: \",len(dataset))\n",
    "    print(\"Number of Training data points: \",len(train))\n",
    "    print(\"Number of Validation data points: \",len(val))\n",
    "    print(\"Number of Testing data points: \",len(test))\n",
    "    \n",
    "    train_x,train_y = return_x_y(train,log_time)\n",
    "    val_x,val_y = return_x_y(val,log_time)\n",
    "    test_x,test_y = return_x_y(test,log_time)\n",
    "    \n",
    "    return train_x,train_y,val_x,val_y,test_x,test_y\n",
    "\n",
    "train_x,train_y,val_x,val_y,test_x,test_y = create_train_test_split(df_regression,log_time=log_time)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9439b94e",
   "metadata": {},
   "source": [
    "Lets scale our data. Here we will use MinMax scaling on the interval (-1,1). You can experiment with other scaling such as StandardScalers (z-score normalization) if you wish."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 253,
   "id": "e646aee0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Features\n",
      "Training: 1.0000000000000004 -1.0000000000000002\n",
      "Validation: 1.0 -1.0\n",
      "Testing: 1.0000000000000002 -1.0\n",
      " \n",
      "Targets\n",
      "Training: 1.7724999487400055 0.0004999637603759\n",
      "Validation: 1.7924999594688416 0.0001000165939331\n",
      "Testing: 1.7524999380111694 0.0009999275207519\n"
     ]
    }
   ],
   "source": [
    "feature_scaler = MinMaxScaler((-1,1))\n",
    "feature_scaler.fit(df_regression.drop(columns=['time_until_disrupt']).to_numpy()) # Lets use global statistics for simplicity here.\n",
    "\n",
    "x_train_scaled = feature_scaler.transform(train_x)\n",
    "x_val_scaled = feature_scaler.transform(val_x)\n",
    "x_test_scaled = feature_scaler.transform(test_x)\n",
    "\n",
    "print(\"Features\")\n",
    "print(\"Training:\",x_train_scaled.max(),x_train_scaled.min())\n",
    "print(\"Validation:\",x_val_scaled.max(),x_val_scaled.min())\n",
    "print(\"Testing:\",x_test_scaled.max(),x_test_scaled.min())\n",
    "\n",
    "if log_time: # Declared above\n",
    "    target_scaler = MinMaxScaler((0,1))\n",
    "    target_scaler.fit(np.log(df_regression['time_until_disrupt'].values.reshape(-1,1))) # Using global statistics for simplicity here\n",
    "    \n",
    "    y_train_scaled = target_scaler.transform(train_y.reshape(-1,1)).flatten()\n",
    "    y_val_scaled = target_scaler.transform(val_y.reshape(-1,1)).flatten()\n",
    "    y_test_scaled = target_scaler.transform(test_y.reshape(-1,1)).flatten()\n",
    "else:\n",
    "    y_train_scaled = train_y.copy()\n",
    "    y_val_scaled = val_y.copy()\n",
    "    y_test_scaled = test_y.copy()\n",
    "    \n",
    "print(\" \")\n",
    "print(\"Targets\")\n",
    "print(\"Training:\",y_train_scaled.max(),y_train_scaled.min())\n",
    "print(\"Validation:\",y_val_scaled.max(),y_val_scaled.min())\n",
    "print(\"Testing:\",y_test_scaled.max(),y_test_scaled.min())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e542f9ae",
   "metadata": {},
   "source": [
    "### Model Definition\n",
    "\n",
    "We will use a simple Deep Neural Network (DNN), commonly referred to as a Multi-Layer Perceptron (MLP).\n",
    "\n",
    "We are going to make this as explicit as possible for clarity. We could use nn.Sequential to package this more neatly if we so choose, and also be more dynamic in our model creation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 254,
   "id": "7c3085d8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MLP(\n",
      "  (init_layer): Linear(in_features=28, out_features=64, bias=True)\n",
      "  (B0): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  (L1): Linear(in_features=64, out_features=256, bias=True)\n",
      "  (B1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  (L2): Linear(in_features=256, out_features=512, bias=True)\n",
      "  (B2): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  (L3): Linear(in_features=512, out_features=256, bias=True)\n",
      "  (B3): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  (L4): Linear(in_features=256, out_features=128, bias=True)\n",
      "  (B4): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  (output): Linear(in_features=128, out_features=1, bias=True)\n",
      "  (activation): SELU()\n",
      "  (output_activation): ReLU()\n",
      ")\n"
     ]
    }
   ],
   "source": [
    "import torch\n",
    "from torch import nn\n",
    "from torch.nn import BatchNorm1d\n",
    "\n",
    "class MLP(nn.Module):\n",
    "    def __init__(self,input_shape,output_shape):\n",
    "        super(MLP, self).__init__()\n",
    "\n",
    "        self.init_layer = nn.Linear(input_shape,64)\n",
    "        self.B0 = nn.BatchNorm1d(64)\n",
    "        self.L1 = nn.Linear(64,256)\n",
    "        self.B1 = nn.BatchNorm1d(256)\n",
    "        self.L2 = nn.Linear(256,512)\n",
    "        self.B2 = nn.BatchNorm1d(512)\n",
    "        self.L3 = nn.Linear(512,256)\n",
    "        self.B3 = nn.BatchNorm1d(256)\n",
    "        self.L4 = nn.Linear(256,128)\n",
    "        self.B4 = nn.BatchNorm1d(128)\n",
    "        self.output = nn.Linear(128,output_shape)\n",
    "        self.activation = nn.SELU()\n",
    "        self.output_activation = nn.ReLU()\n",
    "    def forward(self,x):\n",
    "\n",
    "        x = self.activation(self.B0(self.init_layer(x))) # SELU(BNorm(Linear(x)))\n",
    "        x = self.activation(self.B1(self.L1(x)))\n",
    "        x = self.activation(self.B2(self.L2(x)))\n",
    "        x = self.activation(self.B3(self.L3(x)))\n",
    "        x = self.activation(self.B4(self.L4(x)))\n",
    "        x = self.output_activation(self.output(x)).squeeze(1) # Time must be > 0, ReLU is good choice here.\n",
    "\n",
    "        return x\n",
    "     \n",
    "regressor = MLP(input_shape=x_train_scaled.shape[1],output_shape=1) # Singular output\n",
    "print(regressor)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bd242698",
   "metadata": {},
   "source": [
    "Lets see if we get output."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 255,
   "id": "b3d42365",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([1.1473, 0.0000, 0.0000], grad_fn=<SqueezeBackward1>)"
      ]
     },
     "execution_count": 255,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regressor(torch.tensor(x_train_scaled[:3]).float())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "334897aa",
   "metadata": {},
   "source": [
    "We are going to use built in pytorch functions for creating our data loading pipeline.\n",
    "Since we have simple tabular data, we can use TensorDataset(x,y).\n",
    "\n",
    "Note that we are going to need to convert these numpy arrays to torch.tensor() objects first."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 256,
   "id": "bc7338c6",
   "metadata": {},
   "outputs": [],
   "source": [
    "from torch.utils.data import TensorDataset, DataLoader\n",
    "\n",
    "train_dataset = TensorDataset(torch.tensor(x_train_scaled),torch.tensor(y_train_scaled))\n",
    "val_dataset = TensorDataset(torch.tensor(x_val_scaled),torch.tensor(y_val_scaled))\n",
    "test_dataset = TensorDataset(torch.tensor(x_test_scaled),torch.tensor(y_test_scaled))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fdc54a87",
   "metadata": {},
   "source": [
    "We can also build custom datasets, which you will see in next weeks lectures on normalizing flows. We will explain in more detail how those work then, but for now you should have knowledge of the basic method of getting data from these objects."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 257,
   "id": "b28c821a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(tensor([ 0.1039, -0.6623,  0.1553, -0.8241, -0.2839,  0.0161, -0.5669, -0.9937,\n",
       "         -0.7154,  0.4758,  0.1006, -0.9606, -0.5235, -0.7937, -0.3066,  0.1707,\n",
       "          0.5250, -0.9998, -0.8750, -0.7934, -0.3216, -0.0954,  0.0335, -0.0196,\n",
       "         -0.6106, -0.5459, -0.9736,  0.4466], dtype=torch.float64),\n",
       " tensor(0.0160, dtype=torch.float64))"
      ]
     },
     "execution_count": 257,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_dataset.__getitem__(0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d3d558e6",
   "metadata": {},
   "source": [
    "All pytorch datasets must have a __getitem__() function. This will be imporant when we want to pass these objects to a DataLoader(). Again, we will go into more detail on all of this next week. But lets get familiar with the basic usage.\n",
    "\n",
    "We know that when we train deep learning models we will use batches of inputs to compute gradients. An efficient way to batch your data is to use the DataLoader() functions from pytorch. This will pass a series of random indices to your TensorDataset's __getitem__() function and form batches."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 224,
   "id": "df62dfc3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Batch 0 -> x shape:  torch.Size([10, 28])  y shape:  torch.Size([10])\n"
     ]
    }
   ],
   "source": [
    "train_loader = DataLoader(train_dataset,batch_size=10)\n",
    "\n",
    "for i,data in enumerate(train_loader):\n",
    "    x = data[0]\n",
    "    y = data[1]\n",
    "    print(\"Batch {0}\".format(i),\"-> x shape: \",x.shape,\" y shape: \",y.shape)\n",
    "    break"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "296d6749",
   "metadata": {},
   "source": [
    "Now lets create a function to return dataloaders for the three datasets we have created."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 258,
   "id": "7f350461",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create dataloaders to iterate.\n",
    "# We take as input the three TensorDatasets along with the batch sizes we want to use.\n",
    "def CreateLoaders(train_dataset,val_dataset,test_dataset,train_batch,val_batch,test_batch):\n",
    "    train_loader = DataLoader(train_dataset,\n",
    "                            batch_size=train_batch,\n",
    "                            shuffle=True)\n",
    "    val_loader =  DataLoader(val_dataset,\n",
    "                            batch_size=val_batch,\n",
    "                            shuffle=False)\n",
    "    test_loader =  DataLoader(test_dataset,\n",
    "                            batch_size=val_batch,\n",
    "                            shuffle=False)\n",
    "    return train_loader,val_loader,test_loader\n",
    "\n",
    "train_loader,val_loader,test_loader = CreateLoaders(train_dataset,val_dataset,test_dataset,10,10,10)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c8d0cc86",
   "metadata": {},
   "source": [
    "### Training Function\n",
    "\n",
    "We almost have everything we need to start training our model.\n",
    "\n",
    "1. Datasets in a pytorch format.\n",
    "2. DataLoaders for batch creation.\n",
    "3. A training function - lets build this.\n",
    "\n",
    "Lets write the training function in such a way that we can use it for any network.\n",
    "\n",
    "There are two main components:\n",
    "\n",
    "Training loop\n",
    "\n",
    "Validation loop\n",
    "\n",
    "We also need a clever way to store our parameters. Lets use a dictionary to do this.\n",
    "\n",
    "Note that we are going to assume GPU usage here, so use your notebook on sciclone!\n",
    "\n",
    "You can see the information regarding your GPU with the command below:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 286,
   "id": "bc809a77",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wed Jun  5 10:27:28 2024       \n",
      "+-----------------------------------------------------------------------------------------+\n",
      "| NVIDIA-SMI 551.68                 Driver Version: 551.68         CUDA Version: 12.4     |\n",
      "|-----------------------------------------+------------------------+----------------------+\n",
      "| GPU  Name                     TCC/WDDM  | Bus-Id          Disp.A | Volatile Uncorr. ECC |\n",
      "| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |\n",
      "|                                         |                        |               MIG M. |\n",
      "|=========================================+========================+======================|\n",
      "|   0  NVIDIA GeForce RTX 4090      WDDM  |   00000000:01:00.0  On |                  Off |\n",
      "|  0%   34C    P8             19W /  450W |    2156MiB /  24564MiB |      6%      Default |\n",
      "|                                         |                        |                  N/A |\n",
      "+-----------------------------------------+------------------------+----------------------+\n",
      "                                                                                         \n",
      "+-----------------------------------------------------------------------------------------+\n",
      "| Processes:                                                                              |\n",
      "|  GPU   GI   CI        PID   Type   Process name                              GPU Memory |\n",
      "|        ID   ID                                                               Usage      |\n",
      "|=========================================================================================|\n",
      "|    0   N/A  N/A      5640    C+G   ...oogle\\Chrome\\Application\\chrome.exe      N/A      |\n",
      "|    0   N/A  N/A      5756    C+G   ...5n1h2txyewy\\ShellExperienceHost.exe      N/A      |\n",
      "|    0   N/A  N/A      6936    C+G   ...ems\\FX\\SubAgent\\AlienFXSubAgent.exe      N/A      |\n",
      "|    0   N/A  N/A      8076    C+G   C:\\Windows\\explorer.exe                     N/A      |\n",
      "|    0   N/A  N/A      8160      C   ...James\\.conda\\envs\\ptorch\\python.exe      N/A      |\n",
      "|    0   N/A  N/A      9568    C+G   ...nt.CBS_cw5n1h2txyewy\\SearchHost.exe      N/A      |\n",
      "|    0   N/A  N/A      9592    C+G   ...2txyewy\\StartMenuExperienceHost.exe      N/A      |\n",
      "|    0   N/A  N/A     12132    C+G   ...t.LockApp_cw5n1h2txyewy\\LockApp.exe      N/A      |\n",
      "|    0   N/A  N/A     17240    C+G   ...Programs\\Microsoft VS Code\\Code.exe      N/A      |\n",
      "|    0   N/A  N/A     23292    C+G   ...__8wekyb3d8bbwe\\WindowsTerminal.exe      N/A      |\n",
      "|    0   N/A  N/A     31464    C+G   ...siveControlPanel\\SystemSettings.exe      N/A      |\n",
      "|    0   N/A  N/A     31864    C+G   ...ekyb3d8bbwe\\PhoneExperienceHost.exe      N/A      |\n",
      "|    0   N/A  N/A     33840    C+G   ...on\\125.0.2535.85\\msedgewebview2.exe      N/A      |\n",
      "|    0   N/A  N/A     37792    C+G   ...wekyb3d8bbwe\\XboxGameBarWidgets.exe      N/A      |\n",
      "|    0   N/A  N/A     42028    C+G   ...crosoft\\Edge\\Application\\msedge.exe      N/A      |\n",
      "+-----------------------------------------------------------------------------------------+\n"
     ]
    }
   ],
   "source": [
    "!nvidia-smi"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7b43a5c3",
   "metadata": {},
   "source": [
    "If you don't see an output with the above command please let me know and we will sort it."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 305,
   "id": "6ea18421",
   "metadata": {},
   "outputs": [],
   "source": [
    "config = {\"seed\":8,\n",
    "          \"name\": \"MyRegressionModel\",\n",
    "          \"run_val\":1,\n",
    "          \"model\": {\n",
    "               \"input_shape\":28,\n",
    "               \"output_shape\": 1,\n",
    "           },\n",
    "          \"optimizer\": {\n",
    "                \"lr\": 7e-4,\n",
    "            },\n",
    "          \"num_epochs\": 100,\n",
    "          \"dataloader\": {\n",
    "            \"train\": {\n",
    "                \"batch_size\": 64\n",
    "            },\n",
    "            \"val\": {\n",
    "                \"batch_size\": 64\n",
    "            },\n",
    "            \"test\": {\n",
    "                \"batch_size\": 64\n",
    "            },\n",
    "          },\n",
    "            \"output\": {\n",
    "                \"dir\":\"./\"\n",
    "            }\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4220964f",
   "metadata": {},
   "source": [
    "Below is our training function. It will operate as follows:\n",
    "\n",
    "1. We set seeds for commonly used packages for reproducability. \n",
    "2. Create a directory with name corresponding to the name field in the above dictionary.\n",
    "3. We create our dataloaders - we only need train/val, but test will be created anyways and not used.\n",
    "4. We create an instance of our MLP (funciton defined above)\n",
    "5. We create an instance of an optimizer - Adam.\n",
    "6. Define a learning rate scheduler - CosineAnnealing. We will see what this looks like graphically.\n",
    "7. Define our loss function - SmoothL1Loss - https://pytorch.org/docs/stable/generated/torch.nn.SmoothL1Loss.html\n",
    "8. For each epoch iterate over all batches in the training loader and train.\n",
    "9. At the end of each epoch, iterate over the validation loader - DO NOT apply gradients."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 306,
   "id": "63001949",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import json\n",
    "import random\n",
    "import pkbar\n",
    "import torch.optim as optim\n",
    "from torch.optim import lr_scheduler\n",
    "import torch.nn as nn\n",
    "import datetime\n",
    "import shutil\n",
    "import time\n",
    "\n",
    "def trainer(config,train_dataset,val_dataset,test_dataset):\n",
    "    # Setup random seed\n",
    "    torch.manual_seed(config['seed'])\n",
    "    np.random.seed(config['seed'])\n",
    "    random.seed(config['seed'])\n",
    "    torch.cuda.manual_seed(config['seed'])\n",
    "\n",
    "    # Create experiment name\n",
    "    exp_name = config['name'] \n",
    "    print(exp_name)\n",
    "\n",
    "    # Create directory structure\n",
    "    output_folder = config['output']['dir']\n",
    "    output_path = os.path.join(output_folder,exp_name)\n",
    "    if os.path.exists(output_path):\n",
    "        timestamp = time.time()\n",
    "        dt_object = datetime.datetime.fromtimestamp(timestamp)\n",
    "        formatted_time = dt_object.strftime('%H_%M_%S')\n",
    "        output_path = output_path +\"_\"+ str(formatted_time)\n",
    "        \n",
    "    os.mkdir(output_path)\n",
    "    \n",
    "    with open(os.path.join(output_path,'config.json'),'w') as outfile:\n",
    "        json.dump(config, outfile)\n",
    "\n",
    "\n",
    "       # Load the dataset\n",
    "    print('Creating Loaders.')\n",
    "    train_batch_size = config['dataloader']['train']['batch_size']\n",
    "    val_batch_size = config['dataloader']['val']['batch_size']\n",
    "    test_batch_size = config['dataloader']['val']['batch_size']\n",
    "\n",
    "    train_loader,val_loader,test_loader = CreateLoaders(train_dataset,val_dataset,test_dataset,train_batch_size,val_batch_size,test_batch_size)\n",
    "    \n",
    "    history = {'train_loss':[],'val_loss':[],'lr':[]}\n",
    "\n",
    "\n",
    "    print(\"Training Size: {0}\".format(len(train_loader.dataset)))\n",
    "    print(\"Validation Size: {0}\".format(len(val_loader.dataset)))\n",
    "\n",
    "    # Create the model\n",
    "    input_shape = config['model']['input_shape']\n",
    "    output_shape = config['model']['output_shape']\n",
    "    net = MLP(input_shape=input_shape,output_shape=output_shape)\n",
    "    t_params = sum(p.numel() for p in net.parameters())\n",
    "    print(\"Network Parameters: \",t_params)\n",
    "    device = torch.device('cuda')\n",
    "    net.to('cuda')\n",
    "\n",
    "    # Optimizer\n",
    "    num_epochs=int(config['num_epochs'])\n",
    "    lr = float(config['optimizer']['lr'])\n",
    "\n",
    "    optimizer = optim.Adam(list(filter(lambda p: p.requires_grad, net.parameters())), lr=lr)\n",
    "    num_steps = len(train_loader) * num_epochs\n",
    "    scheduler = torch.optim.lr_scheduler.CosineAnnealingLR(optimizer=optimizer, T_max=num_steps, last_epoch=-1,\n",
    "                                                           eta_min=0)\n",
    "\n",
    "    startEpoch = 0\n",
    "    global_step = 0\n",
    "\n",
    "    print('===========  Optimizer  ==================:')\n",
    "    print('      LR:', lr)\n",
    "    print('      num_epochs:', num_epochs)\n",
    "    print('')\n",
    "    \n",
    "    \n",
    "    # Loss Function \n",
    "    loss_function = nn.SmoothL1Loss()\n",
    "    for epoch in range(startEpoch,num_epochs):\n",
    "\n",
    "        kbar = pkbar.Kbar(target=len(train_loader), epoch=epoch, num_epochs=num_epochs, width=20, always_stateful=False)\n",
    "\n",
    "        net.train()\n",
    "        running_loss = 0.0\n",
    "\n",
    "        for i, data in enumerate(train_loader):\n",
    "            input  = data[0].to('cuda').float()\n",
    "            y = data[1].to('cuda').float()\n",
    "\n",
    "            optimizer.zero_grad()\n",
    "\n",
    "            with torch.set_grad_enabled(True):\n",
    "                y_hat = net(input)\n",
    "                \n",
    "            loss = loss_function(y,y_hat)\n",
    "\n",
    "            loss.backward()\n",
    "            #torch.nn.utils.clip_grad_norm_(net.parameters(), max_norm=0.5,error_if_nonfinite=True)\n",
    "            optimizer.step()\n",
    "            scheduler.step()\n",
    "\n",
    "            running_loss += loss.item() * input.shape[0]\n",
    "            kbar.update(i, values=[(\"loss\", loss.item())])\n",
    "            global_step += 1\n",
    "\n",
    "\n",
    "        history['train_loss'].append(running_loss / len(train_loader.dataset))\n",
    "        history['lr'].append(scheduler.get_last_lr()[0])\n",
    "\n",
    "\n",
    "        ######################\n",
    "        ## validation phase ##\n",
    "        ######################\n",
    "        if bool(config['run_val']):\n",
    "            net.eval()\n",
    "            val_loss = 0.0\n",
    "            with torch.no_grad():\n",
    "                for i, data in enumerate(val_loader):\n",
    "                    input  = data[0].to('cuda').float()\n",
    "                    y = data[1].to('cuda').float()\n",
    "                    y_hat = net(input)\n",
    "                    loss = loss_function(y,y_hat)\n",
    "\n",
    "                    val_loss += loss\n",
    "\n",
    "            val_loss = val_loss.cpu().numpy() / len(val_loader)\n",
    "\n",
    "            history['val_loss'].append(val_loss)\n",
    "\n",
    "            kbar.add(1, values=[(\"val_loss\", val_loss.item())])\n",
    "\n",
    "            name_output_file = config['name']+'_epoch{:02d}_val_loss_{:.6f}.pth'.format(epoch, val_loss)\n",
    "\n",
    "        else:\n",
    "            kbar.add(1,values=[('val_loss',0.)])\n",
    "            name_output_file = config['name']+'_epoch{:02d}_train_loss_{:.6f}.pth'.format(epoch, running_loss / len(train_loader.dataset))\n",
    "\n",
    "        filename = os.path.join(output_path, name_output_file)\n",
    "\n",
    "        checkpoint={}\n",
    "        checkpoint['net_state_dict'] = net.state_dict()\n",
    "        checkpoint['optimizer'] = optimizer.state_dict()\n",
    "        checkpoint['scheduler'] = scheduler.state_dict()\n",
    "        checkpoint['epoch'] = epoch\n",
    "        checkpoint['history'] = history\n",
    "        checkpoint['global_step'] = global_step\n",
    "\n",
    "        torch.save(checkpoint,filename)\n",
    "\n",
    "        print('')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 307,
   "id": "b5774b53",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MyRegressionModel\n",
      "Creating Loaders.\n",
      "Training Size: 44450\n",
      "Validation Size: 9525\n",
      "Network Parameters:  316865\n",
      "===========  Optimizer  ==================:\n",
      "      LR: 0.0007\n",
      "      num_epochs: 100\n",
      "\n",
      "Epoch: 1/100\n",
      "695/695 [====================] - 5s 7ms/step - loss: 0.0512 - val_loss: 0.0361\n",
      "\n",
      "Epoch: 2/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0374 - val_loss: 0.0366\n",
      "\n",
      "Epoch: 3/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0353 - val_loss: 0.0324\n",
      "\n",
      "Epoch: 4/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0336 - val_loss: 0.0309\n",
      "\n",
      "Epoch: 5/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0318 - val_loss: 0.0278\n",
      "\n",
      "Epoch: 6/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0311 - val_loss: 0.0306\n",
      "\n",
      "Epoch: 7/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0296 - val_loss: 0.0296\n",
      "\n",
      "Epoch: 8/100\n",
      "695/695 [====================] - 7s 11ms/step - loss: 0.0295 - val_loss: 0.0273\n",
      "\n",
      "Epoch: 9/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0284 - val_loss: 0.0298\n",
      "\n",
      "Epoch: 10/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0276 - val_loss: 0.0254\n",
      "\n",
      "Epoch: 11/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0269 - val_loss: 0.0241\n",
      "\n",
      "Epoch: 12/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0263 - val_loss: 0.0258\n",
      "\n",
      "Epoch: 13/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0255 - val_loss: 0.0235\n",
      "\n",
      "Epoch: 14/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0248 - val_loss: 0.0254\n",
      "\n",
      "Epoch: 15/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0243 - val_loss: 0.0220\n",
      "\n",
      "Epoch: 16/100\n",
      "695/695 [====================] - 7s 11ms/step - loss: 0.0240 - val_loss: 0.0247\n",
      "\n",
      "Epoch: 17/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0234 - val_loss: 0.0240\n",
      "\n",
      "Epoch: 18/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0230 - val_loss: 0.0222\n",
      "\n",
      "Epoch: 19/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0226 - val_loss: 0.0206\n",
      "\n",
      "Epoch: 20/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0219 - val_loss: 0.0214\n",
      "\n",
      "Epoch: 21/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0216 - val_loss: 0.0211\n",
      "\n",
      "Epoch: 22/100\n",
      "695/695 [====================] - 7s 11ms/step - loss: 0.0214 - val_loss: 0.0203\n",
      "\n",
      "Epoch: 23/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0208 - val_loss: 0.0213\n",
      "\n",
      "Epoch: 24/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0204 - val_loss: 0.0194\n",
      "\n",
      "Epoch: 25/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0200 - val_loss: 0.0212\n",
      "\n",
      "Epoch: 26/100\n",
      "695/695 [====================] - 7s 11ms/step - loss: 0.0196 - val_loss: 0.0180\n",
      "\n",
      "Epoch: 27/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0193 - val_loss: 0.0208\n",
      "\n",
      "Epoch: 28/100\n",
      "695/695 [====================] - 7s 11ms/step - loss: 0.0188 - val_loss: 0.0198\n",
      "\n",
      "Epoch: 29/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0190 - val_loss: 0.0201\n",
      "\n",
      "Epoch: 30/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0185 - val_loss: 0.0169\n",
      "\n",
      "Epoch: 31/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0182 - val_loss: 0.0168\n",
      "\n",
      "Epoch: 32/100\n",
      "695/695 [====================] - 7s 11ms/step - loss: 0.0179 - val_loss: 0.0175\n",
      "\n",
      "Epoch: 33/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0179 - val_loss: 0.0169\n",
      "\n",
      "Epoch: 34/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0175 - val_loss: 0.0162\n",
      "\n",
      "Epoch: 35/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0173 - val_loss: 0.0149\n",
      "\n",
      "Epoch: 36/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0169 - val_loss: 0.0162\n",
      "\n",
      "Epoch: 37/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0168 - val_loss: 0.0168\n",
      "\n",
      "Epoch: 38/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0166 - val_loss: 0.0163\n",
      "\n",
      "Epoch: 39/100\n",
      "695/695 [====================] - 7s 11ms/step - loss: 0.0165 - val_loss: 0.0148\n",
      "\n",
      "Epoch: 40/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0161 - val_loss: 0.0154\n",
      "\n",
      "Epoch: 41/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0159 - val_loss: 0.0148\n",
      "\n",
      "Epoch: 42/100\n",
      "695/695 [====================] - 7s 11ms/step - loss: 0.0157 - val_loss: 0.0142\n",
      "\n",
      "Epoch: 43/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0156 - val_loss: 0.0147\n",
      "\n",
      "Epoch: 44/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0151 - val_loss: 0.0144\n",
      "\n",
      "Epoch: 45/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0151 - val_loss: 0.0140\n",
      "\n",
      "Epoch: 46/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0149 - val_loss: 0.0147\n",
      "\n",
      "Epoch: 47/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0148 - val_loss: 0.0141\n",
      "\n",
      "Epoch: 48/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0145 - val_loss: 0.0142\n",
      "\n",
      "Epoch: 49/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0143 - val_loss: 0.0130\n",
      "\n",
      "Epoch: 50/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0142 - val_loss: 0.0139\n",
      "\n",
      "Epoch: 51/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0141 - val_loss: 0.0134\n",
      "\n",
      "Epoch: 52/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0139 - val_loss: 0.0141\n",
      "\n",
      "Epoch: 53/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0139 - val_loss: 0.0133\n",
      "\n",
      "Epoch: 54/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0136 - val_loss: 0.0133\n",
      "\n",
      "Epoch: 55/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0133 - val_loss: 0.0120\n",
      "\n",
      "Epoch: 56/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0131 - val_loss: 0.0122\n",
      "\n",
      "Epoch: 57/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0132 - val_loss: 0.0128\n",
      "\n",
      "Epoch: 58/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0130 - val_loss: 0.0124\n",
      "\n",
      "Epoch: 59/100\n",
      "695/695 [====================] - 7s 11ms/step - loss: 0.0128 - val_loss: 0.0121\n",
      "\n",
      "Epoch: 60/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0125 - val_loss: 0.0122\n",
      "\n",
      "Epoch: 61/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0126 - val_loss: 0.0119\n",
      "\n",
      "Epoch: 62/100\n",
      "695/695 [====================] - 7s 11ms/step - loss: 0.0125 - val_loss: 0.0117\n",
      "\n",
      "Epoch: 63/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0122 - val_loss: 0.0119\n",
      "\n",
      "Epoch: 64/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0120 - val_loss: 0.0119\n",
      "\n",
      "Epoch: 65/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0123 - val_loss: 0.0112\n",
      "\n",
      "Epoch: 66/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0120 - val_loss: 0.0113\n",
      "\n",
      "Epoch: 67/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0118 - val_loss: 0.0121\n",
      "\n",
      "Epoch: 68/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0116 - val_loss: 0.0108\n",
      "\n",
      "Epoch: 69/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0115 - val_loss: 0.0115\n",
      "\n",
      "Epoch: 70/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0115 - val_loss: 0.0117\n",
      "\n",
      "Epoch: 71/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0114 - val_loss: 0.0111\n",
      "\n",
      "Epoch: 72/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0112 - val_loss: 0.0104\n",
      "\n",
      "Epoch: 73/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0112 - val_loss: 0.0106\n",
      "\n",
      "Epoch: 74/100\n",
      "695/695 [====================] - 7s 11ms/step - loss: 0.0110 - val_loss: 0.0108\n",
      "\n",
      "Epoch: 75/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0112 - val_loss: 0.0105\n",
      "\n",
      "Epoch: 76/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0109 - val_loss: 0.0104\n",
      "\n",
      "Epoch: 77/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0108 - val_loss: 0.0109\n",
      "\n",
      "Epoch: 78/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0108 - val_loss: 0.0104\n",
      "\n",
      "Epoch: 79/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0107 - val_loss: 0.0104\n",
      "\n",
      "Epoch: 80/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0106 - val_loss: 0.0104\n",
      "\n",
      "Epoch: 81/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0104 - val_loss: 0.0102\n",
      "\n",
      "Epoch: 82/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0105 - val_loss: 0.0100\n",
      "\n",
      "Epoch: 83/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0104 - val_loss: 0.0098\n",
      "\n",
      "Epoch: 84/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0102 - val_loss: 0.0100\n",
      "\n",
      "Epoch: 85/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "695/695 [====================] - 8s 11ms/step - loss: 0.0103 - val_loss: 0.0099\n",
      "\n",
      "Epoch: 86/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0101 - val_loss: 0.0099\n",
      "\n",
      "Epoch: 87/100\n",
      "695/695 [====================] - 7s 11ms/step - loss: 0.0102 - val_loss: 0.0099\n",
      "\n",
      "Epoch: 88/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0101 - val_loss: 0.0100\n",
      "\n",
      "Epoch: 89/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0101 - val_loss: 0.0099\n",
      "\n",
      "Epoch: 90/100\n",
      "695/695 [====================] - 7s 11ms/step - loss: 0.0100 - val_loss: 0.0099\n",
      "\n",
      "Epoch: 91/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0099 - val_loss: 0.0103\n",
      "\n",
      "Epoch: 92/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0101 - val_loss: 0.0098\n",
      "\n",
      "Epoch: 93/100\n",
      "695/695 [====================] - 7s 11ms/step - loss: 0.0099 - val_loss: 0.0108\n",
      "\n",
      "Epoch: 94/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0100 - val_loss: 0.0104\n",
      "\n",
      "Epoch: 95/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0098 - val_loss: 0.0109\n",
      "\n",
      "Epoch: 96/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0099 - val_loss: 0.0098\n",
      "\n",
      "Epoch: 97/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0099 - val_loss: 0.0097\n",
      "\n",
      "Epoch: 98/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0098 - val_loss: 0.0095\n",
      "\n",
      "Epoch: 99/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0098 - val_loss: 0.0095\n",
      "\n",
      "Epoch: 100/100\n",
      "695/695 [====================] - 8s 11ms/step - loss: 0.0098 - val_loss: 0.0096\n",
      "\n"
     ]
    }
   ],
   "source": [
    "trainer(config,train_dataset,val_dataset,test_dataset)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e569df8f",
   "metadata": {},
   "source": [
    "Lets take a look at what the learning rate scheduler has done for us:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 308,
   "id": "ec310ba9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dict_keys(['net_state_dict', 'optimizer', 'scheduler', 'epoch', 'history', 'global_step'])\n"
     ]
    }
   ],
   "source": [
    "dicte = torch.load(os.path.join(config['name'],os.listdir(\"MyRegressionModel\")[-1]))\n",
    "print(dicte.keys())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "39be14ff",
   "metadata": {},
   "source": [
    "Within each .pth file, we have stored the weights at a specific epoch, along with a variety of other useful information. First lets plot the training and validation losses, along with the learning rate."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 309,
   "id": "cade360c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_loss = dicte['history']['train_loss']\n",
    "val_loss = dicte['history']['val_loss']\n",
    "learning_rate = dicte['history']['lr']\n",
    "\n",
    "plt.plot(train_loss,'r-',label=\"Training Loss\")\n",
    "plt.plot(val_loss,'b-',label='Validation Loss')\n",
    "plt.legend(fontsize=20)\n",
    "plt.xlabel('Epoch',fontsize=20)\n",
    "plt.ylabel('Loss',fontsize=20)\n",
    "plt.tick_params(axis='both', labelsize=18)\n",
    "plt.show()\n",
    "\n",
    "plt.plot(learning_rate,'k-',label=r\"Learning Rate - $\\eta$\")\n",
    "plt.legend(fontsize=20)\n",
    "plt.xlabel('Epoch',fontsize=20)\n",
    "plt.ylabel(r'$\\eta$',fontsize=20)\n",
    "plt.tick_params(axis='both', labelsize=18)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6e324042",
   "metadata": {},
   "source": [
    "Notice how we have decreased the learning rate at each step corresponding to a cosine distribution. Specifically, we follow the formula below:\n",
    "\n",
    "\n",
    "\n",
    "$\\eta_{t} = \\eta_{\\text{min}} + \\frac{1}{2} (\\eta_{\\text{max}} - \\eta_{\\text{min}}) \\left(1 + \\cos\\left(\\frac{T_{\\text{cur}}}{T_{\\text{max}}} \\pi\\right)\\right)$\n",
    "\n",
    "$\\eta_{t+1} = \\eta_{t} + \\frac{1}{2} (\\eta_{\\text{max}} - \\eta_{\\text{min}}) \\left(1 - \\cos\\left(\\frac{T_{\\text{cur}}}{T_{\\text{max}}} \\pi\\right)\\right)$\n",
    "\n",
    "where:\n",
    "\n",
    "- $\\eta_{t}$ is the learning rate at time \\(t\\),\n",
    "- $\\eta_{t+1}$ is the learning rate at time \\(t+1\\),\n",
    "- $\\eta_{\\text{min}}$ is the minimum learning rate,\n",
    "- $\\eta_{\\text{max}}$ is the maximum learning rate,\n",
    "- $T_{\\text{cur}}$ is the current time step,\n",
    "- $T_{\\text{max}}$ is the maximum time step.\n",
    "\n",
    "\n",
    "with $T_{\\text{cur}} \\neq (2k+1)T_{\\text{max}}$,\n",
    "\n",
    "and $ T_{\\text{cur}} = (2k+1)T_{\\text{max}}. $\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9004d863",
   "metadata": {},
   "source": [
    "### Testing the Performance \n",
    "\n",
    "We have trained our model, and at each epoch we have saved the weights of the model to .pth file. This will be located in a folder corresponding to the “name” field of the config dictionary.\n",
    "\n",
    "Lets load in the model, and look at some metrics."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 310,
   "id": "d2083b36",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Network Parameters:  316865\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<All keys matched successfully>"
      ]
     },
     "execution_count": 310,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "input_shape = config['model']['input_shape']\n",
    "output_shape = config['model']['output_shape']\n",
    "net = MLP(input_shape=input_shape,output_shape=output_shape)\n",
    "t_params = sum(p.numel() for p in net.parameters())\n",
    "print(\"Network Parameters: \",t_params)\n",
    "device = torch.device('cuda')\n",
    "net.to('cuda')\n",
    "\n",
    "dicte = torch.load(os.path.join(config['name'],os.listdir(\"MyRegressionModel\")[-1]))\n",
    "net.load_state_dict(dicte[\"net_state_dict\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 311,
   "id": "c28c9d4b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "127/149 [================>...] - ETA: 0s"
     ]
    }
   ],
   "source": [
    "test_loader = DataLoader(test_dataset,batch_size=config['dataloader']['test']['batch_size'])\n",
    "net.eval() # Eval mode \n",
    "\n",
    "predictions = []\n",
    "y_true = []\n",
    "kbar = pkbar.Kbar(target=len(test_loader), width=20, always_stateful=False)\n",
    "for i,data in enumerate(test_loader):\n",
    "    x = data[0].to('cuda').float()\n",
    "    \n",
    "    \n",
    "    with torch.set_grad_enabled(False): # Same as with torch.no_grad():\n",
    "        y_hat = net(x).detach().cpu().numpy()\n",
    " \n",
    "        if log_time: # Decalared above\n",
    "            y_hat = np.exp(target_scaler.inverse_transform(y_hat.reshape(-1,1))).flatten()\n",
    "            y_true.append(np.exp(target_scaler.inverse_transform(data[1].numpy().reshape(-1,1))).flatten())\n",
    "        else:\n",
    "            y_true.append(data[1].numpy())\n",
    "            \n",
    "    predictions.append(y_hat)\n",
    "    \n",
    "    kbar.update(i)\n",
    "    \n",
    "predictions = np.concatenate(predictions)\n",
    "y_true = np.concatenate(y_true)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b6b12852",
   "metadata": {},
   "source": [
    "Lets look at the relative residuals."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 312,
   "id": "201d968c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "relative_residuals = (predictions - y_true) / (y_true + 1e-7)\n",
    "\n",
    "plt.hist(relative_residuals,density=True,color='blue',histtype='step',label='Relative Residuals',range=[-1,1],bins=100)\n",
    "plt.title('Relative Residuals - Testing Data',fontsize=25,pad=10)\n",
    "plt.xlabel(r'$\\frac{\\hat{y} - y_{true.}}{y_{true.}}$',fontsize=25,labelpad=10)\n",
    "plt.ylabel('Density',fontsize=25,labelpad=10)\n",
    "plt.tick_params(axis='both', labelsize=18)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c7eb95a0",
   "metadata": {},
   "source": [
    "Uh oh. We have a large spike at -1. Can anyone hypothesize why this could be the case?\n",
    "\n",
    "Can you look into the predicted and true values to get some intuition?"
   ]
  }
 ],
 "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": "3.9.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}