fmda/rnn_class.ipynb

   1 {
   2  "cells": [
   3   {
   4    "cell_type": "markdown",
   5    "id": "ae5031bf-2e51-4caa-b83c-6cdb68926331",
   6    "metadata": {},
   7    "source": [
   8     "# v2 Demonstrations/testing of the class structure"
   9    ]
  10   },
  11   {
  12    "cell_type": "code",
  13    "execution_count": null,
  14    "id": "8b71194e-c5f8-488c-94d4-64b480805d44",
  15    "metadata": {},
  16    "outputs": [],
  17    "source": [
  18     "# Environment\n",
  19     "import numpy as np\n",
  20     "import pandas as pd\n",
  21     "import tensorflow as tf\n",
  22     "import matplotlib.pyplot as plt\n",
  23     "import sys\n",
  24     "# Local modules\n",
  25     "sys.path.append('..')\n",
  26     "import reproducibility\n",
  27     "from utils import print_dict_summary\n",
  28     "from data_funcs import load_and_fix_data, rmse\n",
  29     "from abc import ABC, abstractmethod\n",
  30     "from tensorflow.keras.callbacks import Callback\n",
  31     "from sklearn.metrics import mean_squared_error\n",
  32     "from utils import hash2"
  33    ]
  34   },
  35   {
  36    "cell_type": "code",
  37    "execution_count": null,
  38    "id": "040a3833-210e-4625-b101-6f33c484b127",
  39    "metadata": {},
  40    "outputs": [],
  41    "source": [
  42     "reproducibility_file='version_control/reproducibility_dict0.pickle'\n",
  43     "\n",
  44     "repro={}\n",
  45     "repro.update(load_and_fix_data(reproducibility_file))\n",
  46     "print_dict_summary(repro)\n",
  47     "\n",
  48     "case = 'case11'\n",
  49     "case_data=repro[case]\n",
  50     "case_data[\"h2\"]=300"
  51    ]
  52   },
  53   {
  54    "cell_type": "markdown",
  55    "id": "cf248365-2b11-4f86-8fa4-5d18cbf27ee8",
  56    "metadata": {},
  57    "source": [
  58     "## Stateful Batch Training"
  59    ]
  60   },
  61   {
  62    "cell_type": "code",
  63    "execution_count": null,
  64    "id": "0b3675ad-f7d9-46e1-ad60-7ad5432699ec",
  65    "metadata": {},
  66    "outputs": [],
  67    "source": [
  68     "from moisture_rnn import RNN, create_rnn_data2\n",
  69     "import logging\n",
  70     "from utils import logging_setup\n",
  71     "from moisture_rnn_pkl import pkl2train\n",
  72     "logging_setup()"
  73    ]
  74   },
  75   {
  76    "cell_type": "code",
  77    "execution_count": null,
  78    "id": "8f8df5cb-64f5-4a61-9aba-d4fc44751610",
  79    "metadata": {},
  80    "outputs": [],
  81    "source": [
  82     "reproducibility_file='data/reproducibility_dict2.pickle'\n",
  83     "train = pkl2train([reproducibility_file])"
  84    ]
  85   },
  86   {
  87    "cell_type": "code",
  88    "execution_count": null,
  89    "id": "024ea0f7-3574-4cc2-81d2-ad2af74b9ec9",
  90    "metadata": {},
  91    "outputs": [],
  92    "source": [
  93     "import yaml\n",
  94     "\n",
  95     "with open(\"params.yaml\") as file:\n",
  96     "    params = yaml.safe_load(file)[\"rnn_repro\"]\n",
  97     "params"
  98    ]
  99   },
 100   {
 101    "cell_type": "code",
 102    "execution_count": null,
 103    "id": "21ed1f73-885c-43c3-8869-0877b06f8ad5",
 104    "metadata": {},
 105    "outputs": [],
 106    "source": [
 107     "\n",
 108     "rnn_dat = create_rnn_data2(train[\"reproducibility\"], params)"
 109    ]
 110   },
 111   {
 112    "cell_type": "code",
 113    "execution_count": null,
 114    "id": "41a62222-9771-4b7a-b508-b82f4b38b46c",
 115    "metadata": {},
 116    "outputs": [],
 117    "source": [
 118     "# Update Params for Reproducibility\n",
 119     "\n",
 120     "params.update({\n",
 121     "    'epochs':200,\n",
 122     "    'dropout': [0, 0], # NOTE: length must match total number of layers, default is 1 hidden recurrent layer and 1 dense output layer\n",
 123     "    'recurrent_dropout': 0, # Length must match number of recurrent layers    \n",
 124     "    'rnn_units': 20\n",
 125     "})"
 126    ]
 127   },
 128   {
 129    "cell_type": "code",
 130    "execution_count": null,
 131    "id": "fc961d9e-2e66-472f-8fe4-ec0e888d1b04",
 132    "metadata": {},
 133    "outputs": [],
 134    "source": [
 135     "reproducibility.set_seed()\n",
 136     "rnn = RNN(params)"
 137    ]
 138   },
 139   {
 140    "cell_type": "code",
 141    "execution_count": null,
 142    "id": "080b491f-48ab-4592-8d73-2fbd5ccb82b4",
 143    "metadata": {},
 144    "outputs": [],
 145    "source": [
 146     "rnn.fit(rnn_dat[\"X_train\"], rnn_dat[\"y_train\"])\n",
 147     "# rnn.fit(X_train, y_train)"
 148    ]
 149   },
 150   {
 151    "cell_type": "code",
 152    "execution_count": null,
 153    "id": "882a1902-cdee-46bb-a8de-762465df03cd",
 154    "metadata": {},
 155    "outputs": [],
 156    "source": [
 157     "# preds = rnn.predict(np.reshape(X,(1, X.shape[0], features)))\n",
 158     "print(hash2(rnn_dat['X']))\n",
 159     "print(hash2(rnn_dat['y']))\n",
 160     "preds = rnn.predict(rnn_dat['X'])\n",
 161     "rmse(preds, rnn_dat['y'])"
 162    ]
 163   },
 164   {
 165    "cell_type": "code",
 166    "execution_count": null,
 167    "id": "b9cae53e-2130-4cba-9fc7-901747b6349e",
 168    "metadata": {},
 169    "outputs": [],
 170    "source": [
 171     "reproducibility.set_seed()\n",
 172     "rnn = RNN(params)\n",
 173     "rnn.run_model(rnn_dat)"
 174    ]
 175   },
 176   {
 177    "cell_type": "markdown",
 178    "id": "9d61b952-dbf8-4935-b795-974d9d5d6bbf",
 179    "metadata": {},
 180    "source": [
 181     "## Physics Initialized"
 182    ]
 183   },
 184   {
 185    "cell_type": "code",
 186    "execution_count": null,
 187    "id": "80466af6-a786-483d-8c61-7214a8dd0b67",
 188    "metadata": {},
 189    "outputs": [],
 190    "source": []
 191   },
 192   {
 193    "cell_type": "markdown",
 194    "id": "fd8f1b91-80db-4086-a44d-4bcc695466e4",
 195    "metadata": {},
 196    "source": [
 197     "---\n",
 198     "\n",
 199     "## New Developments"
 200    ]
 201   },
 202   {
 203    "cell_type": "markdown",
 204    "id": "fb3c3a94-e3ff-4efe-98f6-f29896cc1066",
 205    "metadata": {},
 206    "source": [
 207     "### Other Hyperparams"
 208    ]
 209   },
 210   {
 211    "cell_type": "code",
 212    "execution_count": null,
 213    "id": "0db3fe44-e056-49c5-88ac-00b1c43f3beb",
 214    "metadata": {},
 215    "outputs": [],
 216    "source": [
 217     "params.update({\n",
 218     "    'activation': ['sigmoid', 'relu'], # Length must match total number of layers\n",
 219     "    'dropout': [0.2, 0.2], # NOTE: length must match total number of layers, default is 1 hidden recurrent layer and 1 dense output layer\n",
 220     "    'recurrent_dropout': 0.2, # Length must match number of recurrent layers\n",
 221     "    'learning_rate': 0.003,\n",
 222     "    'rnn_units': 9,\n",
 223     "    'epochs': 100\n",
 224     "})"
 225    ]
 226   },
 227   {
 228    "cell_type": "code",
 229    "execution_count": null,
 230    "id": "78ccebaf-7ea4-465a-9af9-5bc3aad7cb25",
 231    "metadata": {},
 232    "outputs": [],
 233    "source": [
 234     "reproducibility.set_seed()\n",
 235     "rnn = RNN(params)"
 236    ]
 237   },
 238   {
 239    "cell_type": "code",
 240    "execution_count": null,
 241    "id": "70787b72-bbef-4774-919a-2fc725fbb6d7",
 242    "metadata": {},
 243    "outputs": [],
 244    "source": [
 245     "rnn.fit(rnn_dat[\"X_train\"], rnn_dat[\"y_train\"], verbose_fit=False)"
 246    ]
 247   },
 248   {
 249    "cell_type": "code",
 250    "execution_count": null,
 251    "id": "03afe3c9-480f-4e36-bd1f-f30a35927441",
 252    "metadata": {},
 253    "outputs": [],
 254    "source": [
 255     "preds = rnn.predict(rnn_dat[\"X\"])\n",
 256     "# np.sqrt(mean_squared_error(preds.flatten(), rnn_dat[\"y\"].flatten())\n",
 257     "rmse(preds, rnn_dat[\"y\"])"
 258    ]
 259   },
 260   {
 261    "cell_type": "markdown",
 262    "id": "6bc0e50e-5a34-42e7-9ae0-0ac688408e3a",
 263    "metadata": {},
 264    "source": [
 265     "### Validation Error"
 266    ]
 267   },
 268   {
 269    "cell_type": "code",
 270    "execution_count": null,
 271    "id": "6bb62317-707c-47a2-8a54-6c7276ae7d91",
 272    "metadata": {},
 273    "outputs": [],
 274    "source": [
 275     "params.update({\n",
 276     "    'train_frac': 0.5,\n",
 277     "    'val_frac': 0.1\n",
 278     "})\n",
 279     "rnn_dat = create_rnn_data2(train[\"reproducibility\"], params)"
 280    ]
 281   },
 282   {
 283    "cell_type": "code",
 284    "execution_count": null,
 285    "id": "e36bc1bd-31af-4369-9e8f-801d61a9aa25",
 286    "metadata": {},
 287    "outputs": [],
 288    "source": [
 289     "reproducibility.set_seed()\n",
 290     "rnn = RNN(params)\n",
 291     "rnn.fit(rnn_dat[\"X_train\"], rnn_dat[\"y_train\"], validation_data = (rnn_dat[\"X_val\"], rnn_dat[\"y_val\"]))"
 292    ]
 293   },
 294   {
 295    "cell_type": "code",
 296    "execution_count": null,
 297    "id": "e2fa46fd-4c08-4f92-96cf-e7ca3ff8b320",
 298    "metadata": {},
 299    "outputs": [],
 300    "source": [
 301     "import importlib\n",
 302     "import moisture_rnn\n",
 303     "importlib.reload(moisture_rnn)\n",
 304     "from moisture_rnn import RNN"
 305    ]
 306   },
 307   {
 308    "cell_type": "code",
 309    "execution_count": null,
 310    "id": "e534f3a6-faf6-4ce6-9f5f-b88af5928ec6",
 311    "metadata": {},
 312    "outputs": [],
 313    "source": [
 314     "reproducibility.set_seed()\n",
 315     "rnn = RNN(params)\n",
 316     "rnn.run_model(rnn_dat)"
 317    ]
 318   },
 319   {
 320    "cell_type": "code",
 321    "execution_count": null,
 322    "id": "2ee11823-a3e1-406a-85b7-3085aa685318",
 323    "metadata": {},
 324    "outputs": [],
 325    "source": []
 326   }
 327  ],
 328  "metadata": {
 329   "kernelspec": {
 330    "display_name": "Python 3 (ipykernel)",
 331    "language": "python",
 332    "name": "python3"
 333   },
 334   "language_info": {
 335    "codemirror_mode": {
 336     "name": "ipython",
 337     "version": 3
 338    },
 339    "file_extension": ".py",
 340    "mimetype": "text/x-python",
 341    "name": "python",
 342    "nbconvert_exporter": "python",
 343    "pygments_lexer": "ipython3",
 344    "version": "3.10.9"
 345   }
 346  },
 347  "nbformat": 4,
 348  "nbformat_minor": 5
 349 }