basic_r2.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2025-01-04 06:59:14.494546: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.\n",
      "2025-01-04 06:59:14.528254: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.\n",
      "To enable the following instructions: SSE4.1 SSE4.2 AVX AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "import sklearn.metrics\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.metrics import r2_score\n",
    "import keras\n",
    "import pickle\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import StandardScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_data(gyro='EC2', mode=-1):\n",
    "    # Define your input and output columns\n",
    "    input_columns = [\n",
    "        'pol_ang', 'tor_ang', 'Bt_center', 'Ip', 'R0', 'Z0', 'aminor', 'betan', \n",
    "        'elong', 'li', 'volume', \n",
    "        'ne_val_0', 'ne_val_10', 'ne_val_20'\n",
    "    ]\n",
    "\n",
    "    output_columns = ['rho_pol', 'R', 'Z', 'CD_eta']\n",
    "\n",
    "    # Load the data\n",
    "    df_initial = pd.read_csv(\"TORBEAM_all_df.csv\")  # replace with your CSV file path\n",
    "\n",
    "    # Narrow down to a specific gyro and mode\n",
    "    temp = df_initial[df_initial['gyro'] == gyro]\n",
    "    temp = temp[temp['mode'] == mode]\n",
    "    # Remove rho_pol of -1\n",
    "    temp = temp[temp['rho_pol'] != -1]\n",
    "    df = temp.dropna()\n",
    "\n",
    "    # Extract input features and output targets\n",
    "    X = df[input_columns].values\n",
    "    y = df[output_columns].values\n",
    "\n",
    "    # Optional: Perform a train-validation-test split\n",
    "    X_train, X_temp, y_train, y_temp = train_test_split(X, y, test_size=0.2, random_state=42)\n",
    "    X_val, X_test, y_val, y_test   = train_test_split(X_temp, y_temp, test_size=0.5, random_state=42)\n",
    "\n",
    "    # Optional: Normalize features (recommended for neural networks)\n",
    "    scaler_X = StandardScaler()\n",
    "    X_train = scaler_X.fit_transform(X_train)\n",
    "    X_val = scaler_X.transform(X_val)\n",
    "    X_test = scaler_X.transform(X_test)\n",
    "\n",
    "    # If desired, you can also scale the outputs, especially if their ranges differ significantly.\n",
    "    scaler_y = StandardScaler()\n",
    "    y_train = scaler_y.fit_transform(y_train)\n",
    "    y_val = scaler_y.transform(y_val)\n",
    "    y_test = scaler_y.transform(y_test)\n",
    "\n",
    "    return X_train, X_val, X_test, y_train, y_val, y_test, scaler_X, scaler_y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023/2023 [==============================] - 1s 577us/step\n",
      "2023/2023 [==============================] - 1s 578us/step\n",
      "2023/2023 [==============================] - 1s 625us/step\n",
      "2023/2023 [==============================] - 1s 592us/step\n",
      "2057/2057 [==============================] - 1s 577us/step\n",
      "2057/2057 [==============================] - 1s 579us/step\n",
      "2057/2057 [==============================] - 1s 597us/step\n",
      "2057/2057 [==============================] - 2s 712us/step\n",
      "1996/1996 [==============================] - 1s 579us/step\n",
      "1996/1996 [==============================] - 1s 578us/step\n",
      "1996/1996 [==============================] - 1s 591us/step\n",
      "1996/1996 [==============================] - 1s 612us/step\n",
      "2057/2057 [==============================] - 1s 591us/step\n",
      "2057/2057 [==============================] - 1s 583us/step\n",
      "2057/2057 [==============================] - 1s 621us/step\n",
      "2057/2057 [==============================] - 1s 586us/step\n",
      "1998/1998 [==============================] - 1s 582us/step\n",
      "1998/1998 [==============================] - 1s 584us/step\n",
      "1998/1998 [==============================] - 1s 606us/step\n",
      "1998/1998 [==============================] - 1s 602us/step\n",
      "2056/2056 [==============================] - 1s 588us/step\n",
      "2056/2056 [==============================] - 1s 583us/step\n",
      "2056/2056 [==============================] - 1s 592us/step\n",
      "2056/2056 [==============================] - 1s 604us/step\n"
     ]
    }
   ],
   "source": [
    "output_names = ['rho_pol', 'R', 'Z', 'CD_eta']\n",
    "ind = 0\n",
    "full_df = None\n",
    "\n",
    "for gyro in ['EC2', 'EC4', 'EC5']:\n",
    "    for mode in [-1, 1]:\n",
    "        X_train, X_val, X_test, y_train, y_val, y_test, _, _ = get_data(gyro=gyro, mode=mode)\n",
    "\n",
    "        for num_units in [60,80,100,120]:\n",
    "            if mode == -1:\n",
    "                mode_name = 'xmode'\n",
    "                plt_name = 'X-Mode'\n",
    "            else:\n",
    "                mode_name = 'omode'\n",
    "                plt_name = 'O-Mode'\n",
    "            model = keras.saving.load_model(f'models/{gyro}_{mode_name}_{num_units}.keras')\n",
    "            y_pred = model.predict(X_test)\n",
    "            with open(f'models/{gyro}_{mode_name}_{num_units}.pkl', 'rb') as f:\n",
    "                training_info = pickle.load(f)\n",
    "\n",
    "            accuracy = training_info['accuracy']\n",
    "            history = training_info['history']\n",
    "            scaler_X = training_info['scaler_X']\n",
    "            scaler_y = training_info['scaler_y']\n",
    "\n",
    "            y_test_plot = scaler_y.inverse_transform(y_test)\n",
    "            y_pred_plot = scaler_y.inverse_transform(y_pred)\n",
    "\n",
    "\n",
    "            d = {}\n",
    "            d['gyro'] = gyro\n",
    "            d['mode'] = mode\n",
    "            d['num_units'] = num_units\n",
    "            for i, output_name in enumerate(output_names):\n",
    "                r2_score = sklearn.metrics.r2_score(y_test_plot[:, i], y_pred_plot[:, i])\n",
    "                d[output_name] = r2_score\n",
    "\n",
    "            df = pd.DataFrame(d, index=[ind])\n",
    "            if full_df is None:\n",
    "                full_df = df\n",
    "            else:\n",
    "                full_df = pd.concat([full_df, df])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "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>gyro</th>\n",
       "      <th>mode</th>\n",
       "      <th>num_units</th>\n",
       "      <th>rho_pol</th>\n",
       "      <th>R</th>\n",
       "      <th>Z</th>\n",
       "      <th>CD_eta</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>EC2</td>\n",
       "      <td>-1</td>\n",
       "      <td>60</td>\n",
       "      <td>0.995904</td>\n",
       "      <td>0.993927</td>\n",
       "      <td>0.998411</td>\n",
       "      <td>0.994615</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>EC2</td>\n",
       "      <td>-1</td>\n",
       "      <td>80</td>\n",
       "      <td>0.996461</td>\n",
       "      <td>0.994624</td>\n",
       "      <td>0.998486</td>\n",
       "      <td>0.996158</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>EC2</td>\n",
       "      <td>-1</td>\n",
       "      <td>100</td>\n",
       "      <td>0.997170</td>\n",
       "      <td>0.994897</td>\n",
       "      <td>0.998918</td>\n",
       "      <td>0.997012</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>EC2</td>\n",
       "      <td>-1</td>\n",
       "      <td>120</td>\n",
       "      <td>0.997328</td>\n",
       "      <td>0.994874</td>\n",
       "      <td>0.998863</td>\n",
       "      <td>0.997165</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  gyro  mode  num_units   rho_pol         R         Z    CD_eta\n",
       "0  EC2    -1         60  0.995904  0.993927  0.998411  0.994615\n",
       "0  EC2    -1         80  0.996461  0.994624  0.998486  0.996158\n",
       "0  EC2    -1        100  0.997170  0.994897  0.998918  0.997012\n",
       "0  EC2    -1        120  0.997328  0.994874  0.998863  0.997165"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "full_df[np.logical_and(full_df['gyro'] == 'EC2', full_df['mode'] == -1)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Look at normalizationns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('models/EC2_xmode_60.pkl', 'rb') as f:\n",
    "    training_info = pickle.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['accuracy', 'history', 'scaler_X', 'scaler_y'])"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "training_info.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "EC2 0.31741317934357893\n",
      "EC4 -0.34219577294018877\n",
      "EC5 -0.33584843979968826\n",
      "EC2 0.1037795888951573\n",
      "EC4 0.11791940037597994\n",
      "EC5 0.11580169210313876\n"
     ]
    }
   ],
   "source": [
    "ind = 0\n",
    "for i in range(ind,ind+1):\n",
    "    for gyro in ['EC2', 'EC4', 'EC5']:\n",
    "        with open(f'models/{gyro}_xmode_60.pkl', 'rb') as f:\n",
    "            training_info = pickle.load(f)\n",
    "\n",
    "        print(f'{gyro} {training_info[\"scaler_X\"].mean_[i]}')\n",
    "\n",
    "    for gyro in ['EC2', 'EC4', 'EC5']:\n",
    "        with open(f'models/{gyro}_xmode_60.pkl', 'rb') as f:\n",
    "            training_info = pickle.load(f)\n",
    "        print(f'{gyro} {training_info[\"scaler_X\"].scale_[i]}')\n",
    "    # (training_info['scaler_X'].mean_[2:])\n",
    "    # (training_info['scaler_X'].scale_[2:])\n",
    "    #print(training_info['scaler_y'].mean_)\n",
    "    # print(training_info['scaler_y'].scale_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = keras.saving.load_model(f'models/EC2_xmode_60.keras')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "module 'keras' has no attribute 'tensor'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[16], line 17\u001b[0m\n\u001b[1;32m      1\u001b[0m trial_input \u001b[38;5;241m=\u001b[39m np\u001b[38;5;241m.\u001b[39marray([\u001b[38;5;241m0.304999\u001b[39m, \n\u001b[1;32m      2\u001b[0m                         \u001b[38;5;241m0.75498\u001b[39m, \n\u001b[1;32m      3\u001b[0m                         \u001b[38;5;241m0.781511\u001b[39m, \n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m     14\u001b[0m                         \u001b[38;5;241m0.52868\u001b[39m\n\u001b[1;32m     15\u001b[0m                         ])\n\u001b[1;32m     16\u001b[0m trial_input \u001b[38;5;241m=\u001b[39m trial_input\u001b[38;5;241m.\u001b[39mreshape(\u001b[38;5;241m1\u001b[39m, \u001b[38;5;241m-\u001b[39m\u001b[38;5;241m1\u001b[39m)\n\u001b[0;32m---> 17\u001b[0m trial_input \u001b[38;5;241m=\u001b[39m \u001b[43mkeras\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mtensor\u001b[49m(trial_input)\n\u001b[1;32m     18\u001b[0m output \u001b[38;5;241m=\u001b[39m model\u001b[38;5;241m.\u001b[39mpredict(trial_input)\n\u001b[1;32m     19\u001b[0m \u001b[38;5;28;01mwith\u001b[39;00m \u001b[38;5;28mopen\u001b[39m(\u001b[38;5;124mf\u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mmodels/\u001b[39m\u001b[38;5;132;01m{\u001b[39;00mgyro\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m_xmode_60.pkl\u001b[39m\u001b[38;5;124m'\u001b[39m, \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mrb\u001b[39m\u001b[38;5;124m'\u001b[39m) \u001b[38;5;28;01mas\u001b[39;00m f:\n",
      "\u001b[0;31mAttributeError\u001b[0m: module 'keras' has no attribute 'tensor'"
     ]
    }
   ],
   "source": [
    "trial_input = np.array([0.304999, \n",
    "                        0.75498, \n",
    "                        0.781511, \n",
    "                        0.099258,\n",
    "                        -3.818018, \n",
    "                        0.520651, \n",
    "                        -1.254566,\n",
    "                        -0.447334, \n",
    "                        -4.406280, \n",
    "                        -1.818545, \n",
    "                        -3.939891, \n",
    "                        0.035734, \n",
    "                        0.430360, \n",
    "                        0.52868\n",
    "                        ])\n",
    "trial_input = trial_input.reshape(1, -1)\n",
    "\n",
    "output = model.predict(trial_input)\n",
    "with open(f'models/{gyro}_xmode_60.pkl', 'rb') as f:\n",
    "    training_info = pickle.load(f)\n",
    "\n",
    "print(training_info['scaler_y'].inverse_transform(output))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 3.39217331e-01  1.70878825e+00 -6.74200280e-02 -2.70605993e-04]\n",
      "[0.11401295 0.04625591 0.12267106 0.00925621]\n"
     ]
    }
   ],
   "source": [
    "with open(f'models/EC5_omode_60.pkl', 'rb') as f:\n",
    "    training_info = pickle.load(f)\n",
    "\n",
    "print(training_info['scaler_y'].mean_)\n",
    "print(training_info['scaler_y'].scale_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Check PCS calculation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### EC2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[-0.53583927 -0.00451886  0.05483941 -0.04150851 -1.97538548  0.49219788\n",
      "  -0.37763258 -0.41612103 -4.1282275  -1.67131156 -3.61911192  0.01422088\n",
      "   0.47756198  0.58284796]]\n",
      "1/1 [==============================] - 0s 42ms/step\n",
      "[-0.58367443  0.5007925 ]\n",
      "[1.7221699  0.17023885]\n"
     ]
    }
   ],
   "source": [
    "model = keras.saving.load_model(f'models/EC2_xmode_60.keras')\n",
    "with open(f'models/EC2_xmode_60.pkl', 'rb') as f:\n",
    "    training_info = pickle.load(f)\n",
    "\n",
    "raw_input = np.array([0.261804, \n",
    "                        0.0, \n",
    "                        -1.791502, \n",
    "                        -515634.718,\n",
    "                        1.77, \n",
    "                        0.0, \n",
    "                        0.459118*100,\n",
    "                        0.950630, \n",
    "                        1.450798, \n",
    "                        0.849321, \n",
    "                        10.160793*1e6, \n",
    "                        3.598558, \n",
    "                        3.115731, \n",
    "                        0.209706\n",
    "                        ])\n",
    "\n",
    "raw_input = raw_input.reshape(1, -1)\n",
    "\n",
    "trial_input = training_info['scaler_X'].transform(raw_input)\n",
    "print(trial_input)\n",
    "\n",
    "output = model.predict(trial_input)\n",
    "print(output[0,1:3])\n",
    "print(training_info['scaler_y'].inverse_transform(output)[0,1:3])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "12777734.913684051\n",
      "723089.522511209\n"
     ]
    }
   ],
   "source": [
    "print(training_info['scaler_X'].mean_[10])\n",
    "print(training_info['scaler_X'].scale_[10])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 0s 44ms/step\n",
      "[[ 1.7654405  -0.51966643  1.0948038   0.7874182 ]]\n"
     ]
    }
   ],
   "source": [
    "inputs = [-5.89540962e-01,-5.31455719e-01,+4.50318004e-01,-1.09321609e+00,-1.48777925e+00,\n",
    "-1.53913852e+00,-1.17093095e+00,+6.44183951e-01,+1.04089078e+00,+1.42375586e+00,\n",
    "-9.33865478e-01,+1.88388224e+00,+7.64007621e-01,-7.92732182e-03]\n",
    "\n",
    "raw_input = np.array(inputs).reshape(1, -1)\n",
    "\n",
    "model = keras.saving.load_model(f'models/EC2_xmode_60.keras')\n",
    "\n",
    "output = model.predict(raw_input)\n",
    "print(output)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.553693235, -0.296740413, 0.507193923, 0.0895579234]"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[+5.53693235e-01,-2.96740413e-01,+5.07193923e-01,+8.95579234e-02]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.020018024"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.layers[0].get_weights()[0][0,0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 2.00180244e-02,  2.25528162e-02,  6.89424992e-01,\n",
       "        -9.32391081e-03,  1.36226714e-02, -2.27822224e-03,\n",
       "         3.52142123e-03,  2.21876390e-02,  7.84411654e-02,\n",
       "         1.01493308e-02,  7.84593225e-02, -1.67687722e-02,\n",
       "        -1.17378682e-02, -6.32737298e-03,  6.84254229e-01,\n",
       "         2.81457826e-02, -3.24052584e-04, -2.81280398e-01,\n",
       "         1.67875551e-02,  5.01174806e-03,  2.26888992e-02,\n",
       "        -2.69525591e-03,  3.37144703e-01, -7.26191550e-02,\n",
       "        -3.07646573e-01,  5.07269084e-01,  3.19633546e-04,\n",
       "        -1.83256436e-02, -2.28847891e-01,  4.82691191e-02,\n",
       "         2.91315943e-01, -6.57258136e-03,  3.55967171e-02,\n",
       "         2.17133388e-02,  1.57921836e-01,  3.00484687e-01,\n",
       "        -2.18333807e-04, -1.91089232e-02, -4.07347409e-03,\n",
       "         3.73413190e-02, -5.87272227e-01,  7.31675863e-01,\n",
       "        -7.20458567e-01, -2.65119411e-03,  1.87123148e-03,\n",
       "        -9.72618997e-01,  5.14735188e-03, -1.38757611e-02,\n",
       "        -1.61703415e-02,  1.10167670e+00,  7.72577524e-02,\n",
       "        -1.04267383e+00, -7.57085904e-03,  4.61522758e-01,\n",
       "        -8.89508799e-03, -4.93031710e-01, -2.56484728e-02,\n",
       "         8.32663849e-03,  2.63658613e-02, -1.22366734e-01],\n",
       "       [-1.48721606e-01, -1.60662115e-01, -2.64470756e-01,\n",
       "        -5.48956208e-02, -8.80968571e-03,  3.63123050e-04,\n",
       "         6.05580062e-02,  2.14693099e-01,  1.02882779e+00,\n",
       "        -1.77148685e-01,  2.58957654e-01,  1.53966378e-02,\n",
       "         5.12554586e-01, -1.41119439e-04,  2.64938146e-01,\n",
       "         1.37609756e+00, -7.42075145e-02,  1.38141643e-02,\n",
       "        -1.72849447e-02, -1.28625017e-02, -2.80915290e-01,\n",
       "        -2.09966034e-01,  3.42056423e-01, -9.60419944e-04,\n",
       "         1.49688246e-02,  1.16884279e+00, -1.16634425e-02,\n",
       "        -5.55790186e-01, -2.30983114e+00, -9.80700776e-02,\n",
       "        -1.62122416e+00, -1.81893371e-02,  3.73056233e-01,\n",
       "        -3.96853918e-03,  2.54469728e+00,  8.36189080e-04,\n",
       "         1.28039971e-01, -9.71666202e-02, -6.25731945e-01,\n",
       "        -6.50964528e-02,  2.17167631e-01,  2.77020745e-02,\n",
       "        -4.22534868e-02, -1.63828939e-01,  6.99255662e-03,\n",
       "         4.79414780e-03, -1.52385415e-04, -1.60722286e-01,\n",
       "        -5.72890043e-02, -2.17519980e-03,  7.16770411e-01,\n",
       "        -5.76418638e-02,  1.59725361e-02, -6.34021387e-02,\n",
       "         2.79148430e-01,  1.80364400e-02,  2.43893522e-03,\n",
       "         2.66385600e-02,  1.11442078e-02, -1.56360161e+00],\n",
       "       [-1.22557282e-01, -2.91669071e-01,  3.62379998e-01,\n",
       "         1.75387692e-02,  1.12846203e-01, -1.76636633e-02,\n",
       "        -6.40636563e-01,  5.79259247e-02,  9.63199884e-04,\n",
       "        -7.12427855e-01, -3.68974239e-01, -7.42788494e-01,\n",
       "         7.33888671e-02, -2.83750623e-01, -1.63637236e-01,\n",
       "        -1.82127953e-01,  9.85037923e-01,  8.14877510e-01,\n",
       "         7.83676445e-01,  1.44075736e-01, -3.76424491e-02,\n",
       "        -6.51637614e-02,  5.75735122e-02,  4.68241215e-01,\n",
       "         4.45197195e-01, -7.60528520e-02,  3.98121364e-02,\n",
       "         6.96476921e-02, -2.85201252e-01, -6.27489835e-02,\n",
       "        -1.88674718e-01,  3.73045892e-01,  9.29205585e-03,\n",
       "        -3.22570711e-01, -4.48848248e-01, -2.81393319e-01,\n",
       "         2.34178036e-01,  3.56063247e-02,  6.38059497e-01,\n",
       "         8.37154806e-01, -2.78894842e-01,  1.89548150e-01,\n",
       "         1.63941860e-01, -1.94391400e-01,  4.22450304e-02,\n",
       "        -3.57602149e-01, -6.20213389e-01,  1.52947283e+00,\n",
       "        -7.83001840e-01,  2.79017001e-01, -1.15315092e+00,\n",
       "        -1.37425419e-02,  9.27565277e-01,  5.36127687e-02,\n",
       "        -8.84320736e-02,  7.38299266e-02,  2.38624468e-01,\n",
       "         4.74617705e-02,  4.68512893e-01, -5.65937869e-02],\n",
       "       [-7.19936311e-01,  8.63750875e-02,  1.83514535e-01,\n",
       "         1.18905270e+00, -6.44723892e-01,  1.32341146e+00,\n",
       "        -1.25652254e+00, -3.44830155e-01,  7.47318054e-03,\n",
       "        -7.12238431e-01, -3.94526005e-01, -3.20893973e-01,\n",
       "        -7.18507290e-01, -3.03932160e-01,  8.46899673e-03,\n",
       "         1.96728885e-01,  2.39596423e-02,  3.21959674e-01,\n",
       "         3.99024159e-01,  1.70337651e-02,  4.90341872e-01,\n",
       "         2.44083583e-01, -2.09696237e-02, -1.01986980e+00,\n",
       "        -2.88549274e-01, -9.52916220e-02, -6.55423522e-01,\n",
       "        -1.45724416e-01,  8.71190205e-02, -3.37142609e-02,\n",
       "         3.69316549e-03, -2.27327704e-01, -7.88822174e-01,\n",
       "        -1.30613789e-01,  2.91510671e-02, -4.29584011e-02,\n",
       "        -7.22814426e-02,  1.06971018e-01,  8.64170492e-02,\n",
       "         4.18424755e-01, -1.43108159e-01, -5.81614524e-02,\n",
       "         1.08194485e-01,  6.50341958e-02,  1.25100732e-01,\n",
       "        -3.67972031e-02,  6.42766356e-01, -1.98588982e-01,\n",
       "         5.22518933e-01,  1.09838853e-02,  5.56986108e-02,\n",
       "         1.01237856e-02,  2.43576378e-01,  1.22827634e-01,\n",
       "         5.61289966e-01, -2.80648144e-03,  1.91453159e-01,\n",
       "        -6.88935071e-03,  2.67960168e-02,  9.02019888e-02],\n",
       "       [-1.33034080e-01, -7.47045875e-01, -1.80390775e-01,\n",
       "         5.78757189e-02, -1.00041699e+00, -3.98339391e-01,\n",
       "         4.67406273e-01, -9.65100706e-01,  4.43761870e-02,\n",
       "        -1.76094741e-01,  5.81111126e-02, -1.62563935e-01,\n",
       "        -4.30574864e-01,  1.49815232e-01, -3.79343554e-02,\n",
       "         1.47665545e-01,  3.59503686e-01, -1.37747794e-01,\n",
       "         1.81004137e-01,  3.09352010e-01,  6.07382178e-01,\n",
       "         7.48595178e-01, -3.01099177e-02, -6.52340353e-01,\n",
       "        -1.66530147e-01, -1.27424851e-01, -4.39583242e-01,\n",
       "         1.44131079e-01,  7.61887059e-02,  2.26185605e-01,\n",
       "        -1.02757672e-02, -6.29020393e-01, -9.33448076e-02,\n",
       "        -6.85775653e-04,  1.07251316e-01, -1.67425171e-01,\n",
       "         4.16049868e-01,  4.71770406e-01,  2.08178207e-01,\n",
       "        -3.83921236e-01,  1.47741452e-01,  1.48871079e-01,\n",
       "        -1.17822019e-02,  5.15192077e-02,  6.10993385e-01,\n",
       "        -2.77043693e-02,  5.22004485e-01, -1.02347359e-02,\n",
       "        -3.64870094e-02, -2.62955856e-02, -5.11254966e-02,\n",
       "         3.11438628e-02,  1.03243597e-01, -4.31935042e-01,\n",
       "        -3.04279268e-01, -1.15218073e-01,  1.55012310e-01,\n",
       "        -3.33956368e-02, -3.95447791e-01,  1.63462423e-02],\n",
       "       [ 6.41860217e-02, -7.48621002e-02,  1.69471592e-01,\n",
       "        -5.08022010e-01,  1.39986485e-01, -3.45178694e-01,\n",
       "        -2.60839611e-01,  5.67923367e-01,  2.13390533e-02,\n",
       "         1.18130080e-01,  2.12983206e-01,  8.91414210e-02,\n",
       "         4.88836877e-02,  1.06601506e-01,  1.56927511e-01,\n",
       "        -5.10340892e-02, -7.63823837e-02, -2.92983055e-01,\n",
       "         3.62471640e-02, -5.36608025e-02, -1.53078079e-01,\n",
       "        -4.84004706e-01,  9.40688774e-02, -1.00617267e-01,\n",
       "         4.71887551e-02,  6.06659591e-01,  8.34207594e-01,\n",
       "        -1.58453897e-01, -1.18600965e-01,  6.53942108e-01,\n",
       "         6.24879263e-02,  3.48589689e-01, -3.38030607e-01,\n",
       "        -8.01968272e-04,  1.03284074e-02, -1.60346255e-02,\n",
       "         3.75551760e-01,  3.70602794e-02, -1.01353577e-03,\n",
       "         3.38178098e-01, -1.34214997e-01,  1.88864708e-01,\n",
       "        -4.08130288e-01,  1.91072851e-01,  1.62295535e-01,\n",
       "        -2.77786344e-01,  1.60445124e-01,  1.02423400e-01,\n",
       "        -3.13806057e-01,  1.16221838e-01, -1.92087255e-02,\n",
       "         7.11787865e-02,  2.04789951e-01, -3.04755867e-01,\n",
       "         1.25314504e-01, -1.60455406e-01, -8.43596384e-02,\n",
       "        -3.81166875e-01, -1.69912189e-01,  8.11633281e-03],\n",
       "       [-5.61004937e-01,  2.84859121e-01,  4.33891743e-01,\n",
       "         2.32508719e-01, -1.74415838e-02, -3.18929285e-01,\n",
       "        -1.10607900e-01, -1.95908993e-01,  8.99358988e-02,\n",
       "        -1.53998777e-01,  1.20378740e-01, -1.10022336e-01,\n",
       "        -2.88405836e-01,  3.89650017e-01,  1.04176067e-01,\n",
       "         1.47052124e-01,  4.68191177e-01,  7.01278225e-02,\n",
       "         2.52639890e-01, -9.48091689e-03,  1.76417857e-01,\n",
       "         6.21010624e-02, -6.23729825e-02, -4.68112051e-01,\n",
       "        -3.88447225e-01, -7.63303787e-02, -5.37990451e-01,\n",
       "         2.72893876e-01, -1.23569407e-01, -2.55231019e-02,\n",
       "         2.03110687e-02,  4.95226607e-02, -1.24040239e-01,\n",
       "         1.27199858e-01,  1.57364711e-01, -1.25094727e-01,\n",
       "        -3.32628101e-01, -3.20588320e-01,  1.26584321e-01,\n",
       "        -3.50674123e-01, -2.78557509e-01, -1.76811125e-02,\n",
       "        -1.53358877e-01,  2.25910440e-01,  1.55198470e-01,\n",
       "         8.70498712e-04, -3.50228585e-02,  3.21160108e-02,\n",
       "         7.12269664e-01,  1.14377528e-01,  3.48360017e-02,\n",
       "         1.95235819e-01,  1.22819981e-02,  1.39197949e-02,\n",
       "         2.28993312e-01, -3.60316575e-01,  1.08549751e-01,\n",
       "         2.80880351e-02,  3.18328254e-02,  1.87346235e-01],\n",
       "       [ 5.86161494e-01,  6.44084811e-02, -2.54934192e-01,\n",
       "         7.85826206e-01,  7.96317220e-01,  3.80352661e-02,\n",
       "        -6.26452804e-01, -6.58205524e-02,  2.02548116e-01,\n",
       "        -4.11960661e-01,  2.21693784e-01, -2.28052080e-01,\n",
       "         1.07728727e-01,  7.94460773e-01, -4.22901630e-01,\n",
       "        -3.96655276e-02,  5.66380799e-01, -7.36069679e-02,\n",
       "        -3.50789577e-02,  7.93379620e-02, -1.87713325e-01,\n",
       "         2.88644731e-02,  1.75216883e-01, -9.41394269e-01,\n",
       "         1.09856129e-01,  2.33530745e-01, -2.43383691e-01,\n",
       "         5.39486587e-01, -3.57579626e-02,  2.36131623e-01,\n",
       "         5.96880503e-02,  4.25579280e-01,  2.37876639e-01,\n",
       "        -6.33210003e-01, -2.20377129e-02,  3.88914376e-01,\n",
       "        -1.19102009e-01, -2.75886148e-01,  1.35958204e-02,\n",
       "        -3.60682869e-04,  2.08062083e-01, -2.12973580e-02,\n",
       "        -4.86489944e-02,  4.62277353e-01, -8.71659636e-01,\n",
       "        -3.71185727e-02, -1.12487578e+00,  2.37942889e-01,\n",
       "         3.82958263e-01, -6.57167956e-02,  1.06977120e-01,\n",
       "        -1.44199878e-01, -1.59291232e+00, -1.64283469e-01,\n",
       "         9.81759802e-02,  2.17121854e-01,  4.95656550e-01,\n",
       "        -1.29996979e+00,  1.37304425e-01, -3.19650918e-01],\n",
       "       [ 2.52216607e-02, -2.56138332e-02,  4.61545363e-02,\n",
       "         4.50731665e-01,  2.22487092e-01, -4.07136321e-01,\n",
       "         2.25290563e-02, -3.61794025e-01,  4.61502597e-02,\n",
       "        -2.20614344e-01,  9.72594768e-02,  9.80798900e-03,\n",
       "        -2.20922545e-01,  3.11635852e-01,  1.00439131e-01,\n",
       "         1.79815188e-01,  6.35466993e-01, -2.23035016e-03,\n",
       "         9.45169628e-02,  1.08608417e-01,  1.45295181e-03,\n",
       "         2.03960910e-01, -8.29623342e-02, -1.07103193e+00,\n",
       "        -2.83536434e-01, -1.30113110e-01, -4.44541007e-01,\n",
       "         1.66186228e-01, -8.37498084e-02,  6.47548884e-02,\n",
       "         8.59909039e-03, -2.66782701e-01, -2.46844608e-02,\n",
       "         1.80487365e-01,  1.07271418e-01, -6.17606603e-02,\n",
       "        -1.13304548e-01, -3.45237702e-01,  2.03390904e-02,\n",
       "        -3.04047287e-01,  1.17805433e-02,  1.33849129e-01,\n",
       "        -9.74733233e-02,  2.49967366e-01, -2.72063017e-01,\n",
       "        -1.47171151e-02, -3.47241922e-03,  5.28114587e-02,\n",
       "         6.14296436e-01,  7.35194460e-02,  2.20508035e-02,\n",
       "         9.93311554e-02,  2.57452857e-02,  1.47264063e-01,\n",
       "         2.91417781e-02, -5.63635468e-01, -1.75225228e-01,\n",
       "        -1.14744544e-01,  3.29159230e-01,  1.38847589e-01],\n",
       "       [-1.02611527e-01, -3.79416831e-02, -4.86700237e-02,\n",
       "         2.13953391e-01, -6.36937767e-02,  1.19693565e+00,\n",
       "        -5.74292600e-01,  2.72138774e-01, -8.81205685e-03,\n",
       "        -2.66252518e-01, -1.26213282e-01,  1.09322540e-01,\n",
       "        -1.77051350e-01, -8.13114122e-02,  1.29077584e-01,\n",
       "        -4.72224742e-01, -1.88097224e-01, -3.12995501e-02,\n",
       "        -6.59353137e-02,  6.86557710e-01,  4.44177032e-01,\n",
       "        -1.43762752e-01,  1.43369511e-01, -2.39449203e-01,\n",
       "         2.27370888e-01,  3.11041534e-01,  6.78025603e-01,\n",
       "         2.82370895e-01,  1.64074093e-01,  5.67797869e-02,\n",
       "         1.85810961e-02,  8.41330349e-01,  5.03068626e-01,\n",
       "        -4.84842509e-01, -8.10501203e-02, -1.13857679e-01,\n",
       "         1.13174811e-01, -3.84879529e-01, -2.00429149e-02,\n",
       "        -3.56986225e-01,  2.05067825e-02, -4.02272731e-01,\n",
       "         2.11444050e-01,  1.74960449e-01,  3.36445063e-01,\n",
       "         1.43297046e-01,  5.66734374e-01,  8.58146548e-02,\n",
       "        -3.19651544e-01,  2.90974118e-02,  4.04528007e-02,\n",
       "        -3.53590399e-02, -5.50822437e-01,  2.36577123e-01,\n",
       "        -2.21093535e-01,  5.49571335e-01,  1.56365454e-01,\n",
       "        -2.83186764e-01, -7.46889949e-01, -1.65457919e-01],\n",
       "       [ 7.27713048e-01,  3.02658767e-01, -3.36021394e-01,\n",
       "        -2.14271247e-01,  8.72043967e-01, -1.04097277e-01,\n",
       "         7.51670301e-01,  3.29242468e-01, -1.06086798e-01,\n",
       "        -1.00549065e-01, -1.46745965e-01, -4.05261099e-01,\n",
       "         3.61704081e-01, -8.39563459e-02, -2.09370181e-01,\n",
       "        -7.22357407e-02, -4.75385547e-01,  1.67300459e-02,\n",
       "         2.21087337e-01, -2.84025609e-01, -5.59758618e-02,\n",
       "        -1.65551469e-01,  4.11369763e-02, -7.17474937e-01,\n",
       "         2.23534435e-01,  1.23707011e-01, -6.35983288e-01,\n",
       "        -2.63255715e-01,  4.57943790e-02,  2.39675745e-01,\n",
       "         1.28469700e-02,  1.90843925e-01,  3.32607657e-01,\n",
       "        -6.18015192e-02, -1.27994433e-01,  1.55629039e-01,\n",
       "        -2.00963449e-02,  4.37465698e-01, -1.03136100e-01,\n",
       "         2.51333296e-01,  1.93501770e-01,  6.36109710e-02,\n",
       "         1.77593917e-01, -5.16639948e-01,  1.61084250e-01,\n",
       "        -4.25538383e-02, -4.94628668e-01, -6.96298927e-02,\n",
       "        -4.30302739e-01, -1.47208214e-01, -1.88872349e-02,\n",
       "        -2.26645216e-01, -2.13187728e-02, -2.54927397e-01,\n",
       "         1.79524615e-01,  4.08951402e-01,  1.29971787e-01,\n",
       "         2.61432648e-01,  5.81069477e-02, -1.61222979e-01],\n",
       "       [ 6.98716985e-03,  1.10422373e-01,  2.70412982e-01,\n",
       "        -1.54538810e-01, -4.22634989e-01,  3.57423246e-01,\n",
       "         3.61305177e-02, -1.07170358e-01, -1.07728504e-01,\n",
       "         4.21516687e-01, -9.56438422e-01,  5.62812090e-02,\n",
       "         2.56993353e-01,  2.63079219e-02,  5.41131161e-02,\n",
       "        -7.69833848e-02,  1.06129444e+00, -4.35697913e-01,\n",
       "        -1.49662152e-01,  2.11154789e-01, -4.85075042e-02,\n",
       "         1.01982594e-01, -1.74448133e-01,  1.41462111e+00,\n",
       "        -3.99528503e-01,  7.08967686e-01, -7.38933563e-01,\n",
       "         4.77859862e-02, -4.84590173e-01,  7.10411429e-01,\n",
       "        -1.05184607e-01,  8.60960722e-01, -8.72070134e-01,\n",
       "         1.06485701e+00,  5.70282564e-02, -6.99318051e-01,\n",
       "        -9.82852429e-02,  1.26693264e-01, -1.80672377e-01,\n",
       "         6.03481948e-01, -2.23162189e-01, -3.26706380e-01,\n",
       "         7.07433224e-01, -1.54336944e-01,  2.10099488e-01,\n",
       "        -2.36000679e-02,  4.08013798e-02, -1.15521006e-01,\n",
       "         4.38217252e-01,  6.88428402e-01, -2.14176297e-01,\n",
       "         1.17104858e-01, -4.04198021e-01, -5.83919466e-01,\n",
       "        -3.77319247e-01,  6.05944514e-01,  3.69109631e-01,\n",
       "         8.86713147e-01,  7.49345943e-02,  2.21824408e-01],\n",
       "       [-6.83670819e-01,  1.07011974e+00, -3.07285607e-01,\n",
       "        -7.52376139e-01, -2.59089857e-01,  2.17398927e-01,\n",
       "         1.78190842e-01,  9.65483904e-01, -7.86724091e-02,\n",
       "        -2.17772022e-01, -3.64487886e-01, -5.37642121e-01,\n",
       "         3.80311072e-01,  7.02428669e-02, -8.27691182e-02,\n",
       "         6.73328936e-01, -4.87373888e-01,  2.20130101e-01,\n",
       "         1.12705064e+00,  1.12355387e+00, -5.69035769e-01,\n",
       "        -6.23590887e-01,  3.40387642e-01,  3.67548257e-01,\n",
       "         1.12390757e+00,  3.15158755e-01,  2.14777770e-03,\n",
       "         4.20702398e-01, -7.13649318e-02, -5.45826197e-01,\n",
       "         1.35750547e-01,  5.90224266e-01,  5.38503170e-01,\n",
       "         6.92429423e-01, -4.31312174e-02,  4.70308289e-02,\n",
       "        -2.91115016e-01, -6.19149923e-01,  2.40072146e-01,\n",
       "         4.32288617e-01,  3.59341681e-01,  9.27050829e-01,\n",
       "         7.72972256e-02, -9.86204445e-01,  9.72241044e-01,\n",
       "         6.79965913e-02,  7.75403857e-01, -1.83085591e-01,\n",
       "        -6.74453437e-01, -1.13242045e-02, -1.17006056e-01,\n",
       "        -6.34768307e-02,  4.44251537e-01, -1.64393075e-02,\n",
       "         1.46727169e+00, -7.78604567e-01, -3.01983654e-02,\n",
       "         6.73787147e-02, -3.03066581e-01,  1.06477670e-01],\n",
       "       [ 4.25691791e-02, -9.99243259e-01, -4.98764142e-02,\n",
       "         5.96109390e-01,  2.87956208e-01,  4.26172704e-01,\n",
       "        -4.20641959e-01, -3.41155767e-01,  1.06628217e-01,\n",
       "        -1.12628318e-01,  4.12496388e-01,  9.84650627e-02,\n",
       "        -4.76479113e-01, -6.93141937e-01,  4.67675477e-02,\n",
       "        -1.52488962e-01,  6.90902710e-01, -5.16519770e-02,\n",
       "        -2.44142607e-01, -5.07391512e-01, -1.00621030e-01,\n",
       "         1.13821320e-01, -2.44796351e-01, -3.69937330e-01,\n",
       "        -6.23005331e-01, -5.82799792e-01,  1.97845325e-01,\n",
       "        -3.53474855e-01,  1.29683763e-01,  3.81331831e-01,\n",
       "        -1.32884204e-01, -8.36711466e-01, -4.34343606e-01,\n",
       "        -2.68046021e-01, -1.71086602e-02,  1.10033900e-01,\n",
       "         3.45418245e-01,  6.58050895e-01,  2.03223035e-01,\n",
       "        -7.12672770e-01, -4.87725772e-02, -4.65840459e-01,\n",
       "        -3.27375054e-01,  8.85483980e-01, -7.82916784e-01,\n",
       "        -1.45548940e-01,  2.59757727e-01,  2.21036728e-02,\n",
       "         4.38789874e-01,  1.55696988e-01,  6.45831525e-02,\n",
       "        -6.08395934e-02, -1.99496821e-01,  5.44028759e-01,\n",
       "        -1.10713017e+00,  5.27263105e-01, -3.79147172e-01,\n",
       "        -5.30832931e-02,  1.30256608e-01,  2.51410455e-02]], dtype=float32)"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.layers[0].get_weights()[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### EC4"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "k2c-env",
   "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": "3.11.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}