fmda/rnn_workshop.ipynb

   1 {
   2  "cells": [
   3   {
   4    "cell_type": "markdown",
   5    "id": "244c2fb0-4339-476c-a2db-a641e124e25a",
   6    "metadata": {},
   7    "source": [
   8     "# v2.2 exploration trying to make it work better"
   9    ]
  10   },
  11   {
  12    "cell_type": "code",
  13    "execution_count": null,
  14    "id": "e6cc7920-e380-4b81-bac0-cd6840450e9a",
  15    "metadata": {},
  16    "outputs": [],
  17    "source": [
  18     "# Environment\n",
  19     "import os\n",
  20     "import os.path as osp\n",
  21     "import numpy as np\n",
  22     "import pandas as pd\n",
  23     "import tensorflow as tf\n",
  24     "import matplotlib.pyplot as plt\n",
  25     "import sys\n",
  26     "# Local modules\n",
  27     "sys.path.append('..')\n",
  28     "import reproducibility\n",
  29     "import pandas as pd\n",
  30     "from utils import print_dict_summary\n",
  31     "from data_funcs import rmse, build_train_dict, combine_nested, subset_by_features\n",
  32     "# from moisture_rnn import RNNParams, RNNData, RNN, RNN_LSTM, rnn_data_wrap\n",
  33     "from moisture_rnn_pkl import pkl2train\n",
  34     "from tensorflow.keras.callbacks import Callback\n",
  35     "from utils import hash2\n",
  36     "import copy\n",
  37     "import logging\n",
  38     "import pickle\n",
  39     "from utils import logging_setup, read_yml, read_pkl, hash_ndarray, hash_weights, str2time\n",
  40     "import yaml\n",
  41     "import copy\n",
  42     "import time"
  43    ]
  44   },
  45   {
  46    "cell_type": "code",
  47    "execution_count": null,
  48    "id": "f58e8839-bf0e-4995-b966-c09e4df001ce",
  49    "metadata": {},
  50    "outputs": [],
  51    "source": [
  52     "logging_setup()"
  53    ]
  54   },
  55   {
  56    "cell_type": "markdown",
  57    "id": "fae67b50-f916-45a7-bcc7-61995ba39449",
  58    "metadata": {},
  59    "source": [
  60     "## Test Data"
  61    ]
  62   },
  63   {
  64    "cell_type": "code",
  65    "execution_count": null,
  66    "id": "3efed1fa-9cda-4934-8a6c-edcf179c8755",
  67    "metadata": {},
  68    "outputs": [],
  69    "source": [
  70     "file_paths = ['data/fmda_rocky_202403-05_f05.pkl']"
  71    ]
  72   },
  73   {
  74    "cell_type": "code",
  75    "execution_count": null,
  76    "id": "28fd3746-1861-4afa-ab7e-ac449fbed322",
  77    "metadata": {},
  78    "outputs": [],
  79    "source": [
  80     "# Params used for data filtering\n",
  81     "params_data = read_yml(\"params_data.yaml\") \n",
  82     "params_data"
  83    ]
  84   },
  85   {
  86    "cell_type": "code",
  87    "execution_count": null,
  88    "id": "c45cb8ef-41fc-4bf7-b506-dad5fd24abb3",
  89    "metadata": {},
  90    "outputs": [],
  91    "source": [
  92     "dat = read_pkl(file_paths[0])"
  93    ]
  94   },
  95   {
  96    "cell_type": "code",
  97    "execution_count": null,
  98    "id": "3c960d69-4f8a-4abb-a5d9-ed6cf98f899b",
  99    "metadata": {},
 100    "outputs": [],
 101    "source": [
 102     "import importlib\n",
 103     "import data_funcs\n",
 104     "importlib.reload(data_funcs)\n",
 105     "from data_funcs import build_train_dict"
 106    ]
 107   },
 108   {
 109    "cell_type": "code",
 110    "execution_count": null,
 111    "id": "369cd913-85cb-4855-a80c-817d84637852",
 112    "metadata": {},
 113    "outputs": [],
 114    "source": [
 115     "params_data.update({'hours': None})"
 116    ]
 117   },
 118   {
 119    "cell_type": "code",
 120    "execution_count": null,
 121    "id": "8cdc2ce8-45b4-4caa-81d9-646271ff2e97",
 122    "metadata": {
 123     "scrolled": true
 124    },
 125    "outputs": [],
 126    "source": [
 127     "train3 = build_train_dict(file_paths, params_data, spatial=False, forecast_step=3, drop_na=True)\n"
 128    ]
 129   },
 130   {
 131    "cell_type": "code",
 132    "execution_count": null,
 133    "id": "3c4548ae-caa4-4bc4-9122-9f24e7e59ef7",
 134    "metadata": {},
 135    "outputs": [],
 136    "source": []
 137   },
 138   {
 139    "cell_type": "code",
 140    "execution_count": null,
 141    "id": "3dbb6f24-4435-47b3-90c6-6176582b0d4c",
 142    "metadata": {},
 143    "outputs": [],
 144    "source": []
 145   },
 146   {
 147    "cell_type": "markdown",
 148    "id": "6322f0bc-107d-40a5-96dc-804495085a99",
 149    "metadata": {
 150     "jp-MarkdownHeadingCollapsed": true
 151    },
 152    "source": [
 153     "## Test Other ML"
 154    ]
 155   },
 156   {
 157    "cell_type": "code",
 158    "execution_count": null,
 159    "id": "12992b9a-407f-4131-ac61-e1dc338386bf",
 160    "metadata": {},
 161    "outputs": [],
 162    "source": [
 163     "params = read_yml(\"params.yaml\", subkey='xgb')\n",
 164     "params"
 165    ]
 166   },
 167   {
 168    "cell_type": "code",
 169    "execution_count": null,
 170    "id": "f214fdf8-bb76-4912-8f8c-5d0c8c1230c2",
 171    "metadata": {},
 172    "outputs": [],
 173    "source": [
 174     "dat = read_pkl(\"data/train.pkl\")"
 175    ]
 176   },
 177   {
 178    "cell_type": "code",
 179    "execution_count": null,
 180    "id": "888b7805-15f6-4c09-a05b-7aed7d253f6e",
 181    "metadata": {},
 182    "outputs": [],
 183    "source": [
 184     "cases = [*dat.keys()]"
 185    ]
 186   },
 187   {
 188    "cell_type": "code",
 189    "execution_count": null,
 190    "id": "375055d8-c070-4639-9561-e47d3f21f1f8",
 191    "metadata": {},
 192    "outputs": [],
 193    "source": [
 194     "rnn_dat = RNNData(dat[cases[10]], params['scaler'], params['features_list'])\n",
 195     "rnn_dat.train_test_split(\n",
 196     "    time_fracs = [.8, .1, .1]\n",
 197     ")\n",
 198     "rnn_dat.scale_data()"
 199    ]
 200   },
 201   {
 202    "cell_type": "code",
 203    "execution_count": null,
 204    "id": "e79f8dc8-5cf8-4190-b4ff-e640f61bd78b",
 205    "metadata": {},
 206    "outputs": [],
 207    "source": [
 208     "from moisture_models import XGB, RF, LM"
 209    ]
 210   },
 211   {
 212    "cell_type": "code",
 213    "execution_count": null,
 214    "id": "b3aeb47f-261e-4e29-9eeb-67215e5628f6",
 215    "metadata": {},
 216    "outputs": [],
 217    "source": [
 218     "mod = XGB(params)"
 219    ]
 220   },
 221   {
 222    "cell_type": "code",
 223    "execution_count": null,
 224    "id": "cae9a20d-1caf-45aa-a9c4-aef21b65d9c8",
 225    "metadata": {},
 226    "outputs": [],
 227    "source": [
 228     "mod.params"
 229    ]
 230   },
 231   {
 232    "cell_type": "code",
 233    "execution_count": null,
 234    "id": "68a07b25-c586-4fc4-a3d5-c857354e7a2c",
 235    "metadata": {},
 236    "outputs": [],
 237    "source": [
 238     "mod.fit(rnn_dat.X_train, rnn_dat.y_train)"
 239    ]
 240   },
 241   {
 242    "cell_type": "code",
 243    "execution_count": null,
 244    "id": "c8f88819-0a7a-4420-abb9-56a47015a4de",
 245    "metadata": {},
 246    "outputs": [],
 247    "source": [
 248     "preds = mod.predict(rnn_dat.X_test)"
 249    ]
 250   },
 251   {
 252    "cell_type": "code",
 253    "execution_count": null,
 254    "id": "cb7cdf14-74d6-45e4-bc1b-7d4d47dd41ac",
 255    "metadata": {},
 256    "outputs": [],
 257    "source": [
 258     "rmse(preds, rnn_dat.y_test)"
 259    ]
 260   },
 261   {
 262    "cell_type": "code",
 263    "execution_count": null,
 264    "id": "74d478c7-8c01-448e-9a00-dd0e1ee8e325",
 265    "metadata": {},
 266    "outputs": [],
 267    "source": [
 268     "plt.plot(rnn_dat.y_test)\n",
 269     "plt.plot(preds)"
 270    ]
 271   },
 272   {
 273    "cell_type": "code",
 274    "execution_count": null,
 275    "id": "c5441014-c39a-4414-a779-95b81e1ed6a8",
 276    "metadata": {},
 277    "outputs": [],
 278    "source": [
 279     "params = read_yml(\"params.yaml\", subkey='rf')\n",
 280     "rnn_dat = RNNData(dat[cases[10]], features_list = ['Ed', 'Ew', 'solar', 'wind', 'rain'])\n",
 281     "rnn_dat.train_test_split(\n",
 282     "    time_fracs = [.8, .1, .1]\n",
 283     ")"
 284    ]
 285   },
 286   {
 287    "cell_type": "code",
 288    "execution_count": null,
 289    "id": "cafe711a-20cb-4bd3-a4bc-4995a843a021",
 290    "metadata": {},
 291    "outputs": [],
 292    "source": [
 293     "import importlib\n",
 294     "import moisture_models\n",
 295     "importlib.reload(moisture_models)"
 296    ]
 297   },
 298   {
 299    "cell_type": "code",
 300    "execution_count": null,
 301    "id": "ee45f7d6-f57f-4ff6-995a-527565565f94",
 302    "metadata": {},
 303    "outputs": [],
 304    "source": [
 305     "params"
 306    ]
 307   },
 308   {
 309    "cell_type": "code",
 310    "execution_count": null,
 311    "id": "fafe76e5-0212-4bd1-a058-535935a08780",
 312    "metadata": {},
 313    "outputs": [],
 314    "source": [
 315     "mod2 = RF(params)\n",
 316     "mod2.fit(rnn_dat.X_train, rnn_dat.y_train.flatten())\n",
 317     "preds2 = mod2.predict(rnn_dat.X_test)\n",
 318     "print(rmse(preds2, rnn_dat.y_test.flatten()))\n",
 319     "plt.plot(rnn_dat.y_test)\n",
 320     "plt.plot(preds2)"
 321    ]
 322   },
 323   {
 324    "cell_type": "code",
 325    "execution_count": null,
 326    "id": "c0ab4244-996c-49af-bf4a-8b0c47b0b6db",
 327    "metadata": {},
 328    "outputs": [],
 329    "source": [
 330     "from moisture_models import RF\n",
 331     "mod2 = RF(params)"
 332    ]
 333   },
 334   {
 335    "cell_type": "code",
 336    "execution_count": null,
 337    "id": "aa6c33fd-db35-4c77-9eee-fdb39a934959",
 338    "metadata": {},
 339    "outputs": [],
 340    "source": []
 341   },
 342   {
 343    "cell_type": "code",
 344    "execution_count": null,
 345    "id": "c5598bfe-2d87-4d23-869e-aff127782462",
 346    "metadata": {},
 347    "outputs": [],
 348    "source": [
 349     "params = read_yml(\"params.yaml\", subkey='lm')\n",
 350     "rnn_dat = RNNData(dat[cases[10]], features_list = ['Ed', 'Ew', 'solar', 'wind', 'rain'])\n",
 351     "rnn_dat.train_test_split(\n",
 352     "    time_fracs = [.8, .1, .1]\n",
 353     ")\n",
 354     "mod = LM(params)"
 355    ]
 356   },
 357   {
 358    "cell_type": "code",
 359    "execution_count": null,
 360    "id": "d828c15c-4078-4967-abff-c1fd15d4696d",
 361    "metadata": {},
 362    "outputs": [],
 363    "source": [
 364     "mod.fit(rnn_dat.X_train, rnn_dat.y_train)\n",
 365     "preds = mod.predict(rnn_dat.X_test)\n",
 366     "print(rmse(preds2, rnn_dat.y_test.flatten()))"
 367    ]
 368   },
 369   {
 370    "cell_type": "code",
 371    "execution_count": null,
 372    "id": "8496a32a-8269-4d6b-953e-7f33fe626789",
 373    "metadata": {},
 374    "outputs": [],
 375    "source": []
 376   },
 377   {
 378    "cell_type": "code",
 379    "execution_count": null,
 380    "id": "75ce8bf3-6efb-4dc7-b895-def92f6ce6b4",
 381    "metadata": {},
 382    "outputs": [],
 383    "source": []
 384   },
 385   {
 386    "cell_type": "markdown",
 387    "id": "d6e089d9-e466-45bb-80f2-15c563ae21ad",
 388    "metadata": {},
 389    "source": [
 390     "## Class RNN "
 391    ]
 392   },
 393   {
 394    "cell_type": "code",
 395    "execution_count": null,
 396    "id": "3d5792a1-53e3-4099-8630-1bd5e3f52dcc",
 397    "metadata": {},
 398    "outputs": [],
 399    "source": [
 400     "from tensorflow.keras import layers,models"
 401    ]
 402   },
 403   {
 404    "cell_type": "code",
 405    "execution_count": null,
 406    "id": "0962428e-1124-4e1f-8500-d02b26640204",
 407    "metadata": {},
 408    "outputs": [],
 409    "source": [
 410     "import importlib\n",
 411     "import moisture_rnn\n",
 412     "importlib.reload(moisture_rnn)\n",
 413     "from moisture_rnn import RNN, RNNParams"
 414    ]
 415   },
 416   {
 417    "cell_type": "code",
 418    "execution_count": null,
 419    "id": "a14f9c76-93eb-4b13-a11d-6ccb38285335",
 420    "metadata": {},
 421    "outputs": [],
 422    "source": [
 423     "params = RNNParams(read_yml(\"params.yaml\", subkey='rnn'))"
 424    ]
 425   },
 426   {
 427    "cell_type": "code",
 428    "execution_count": null,
 429    "id": "ed3dd798-6a40-4e90-b40b-accabe49fb35",
 430    "metadata": {},
 431    "outputs": [],
 432    "source": [
 433     "params.update({\n",
 434     "    'hidden_layers': ['lstm', 'conv1d', 'dense'],\n",
 435     "    'hidden_units': [32, 32, 16],\n",
 436     "    'hidden_activation': ['tanh', 'relu', 'relu'],\n",
 437     "    'return_sequences': True\n",
 438     "})"
 439    ]
 440   },
 441   {
 442    "cell_type": "code",
 443    "execution_count": null,
 444    "id": "e559d0d7-5847-4fd0-81e4-7d3ca92147dd",
 445    "metadata": {},
 446    "outputs": [],
 447    "source": [
 448     "import importlib\n",
 449     "import moisture_rnn\n",
 450     "importlib.reload(moisture_rnn)\n",
 451     "from moisture_rnn import RNN, rnn_data_wrap"
 452    ]
 453   },
 454   {
 455    "cell_type": "code",
 456    "execution_count": null,
 457    "id": "7c1627f9-f011-4159-98a2-1b5973929e71",
 458    "metadata": {},
 459    "outputs": [],
 460    "source": [
 461     "reproducibility.set_seed()\n",
 462     "mod = RNN(params)"
 463    ]
 464   },
 465   {
 466    "cell_type": "code",
 467    "execution_count": null,
 468    "id": "5dbc66c0-ccb5-46c2-a073-1fa7a5be750a",
 469    "metadata": {},
 470    "outputs": [],
 471    "source": [
 472     "mod.model_train.summary()"
 473    ]
 474   },
 475   {
 476    "cell_type": "code",
 477    "execution_count": null,
 478    "id": "882c5872-a017-4d9c-90be-88e692dd33e8",
 479    "metadata": {},
 480    "outputs": [],
 481    "source": [
 482     "mod.model_predict.summary()"
 483    ]
 484   },
 485   {
 486    "cell_type": "code",
 487    "execution_count": null,
 488    "id": "30498201-3798-484d-922f-974909b195af",
 489    "metadata": {},
 490    "outputs": [],
 491    "source": [
 492     "rnn_dat = rnn_data_wrap(combine_nested(train3), params)"
 493    ]
 494   },
 495   {
 496    "cell_type": "code",
 497    "execution_count": null,
 498    "id": "e213ffd7-d26c-41ce-8e2b-b17368fdd7a8",
 499    "metadata": {},
 500    "outputs": [],
 501    "source": [
 502     "params.update({\n",
 503     "    'loc_batch_reset': rnn_dat.n_seqs # Used to reset hidden state when location changes for a given batch\n",
 504     "})"
 505    ]
 506   },
 507   {
 508    "cell_type": "code",
 509    "execution_count": null,
 510    "id": "74e599b6-7f4d-4175-a5f1-de892e72ebd4",
 511    "metadata": {},
 512    "outputs": [],
 513    "source": [
 514     "m, errs = mod.run_model(rnn_dat)"
 515    ]
 516   },
 517   {
 518    "cell_type": "code",
 519    "execution_count": null,
 520    "id": "f894d203-d277-48f3-bb57-a610f162361f",
 521    "metadata": {},
 522    "outputs": [],
 523    "source": [
 524     "errs.mean()"
 525    ]
 526   },
 527   {
 528    "cell_type": "code",
 529    "execution_count": null,
 530    "id": "b875ea70-41f9-4550-982b-88380ad1b5a0",
 531    "metadata": {},
 532    "outputs": [],
 533    "source": [
 534     "params"
 535    ]
 536   },
 537   {
 538    "cell_type": "markdown",
 539    "id": "282cb651-b21f-401d-94c5-9e07530a9ba8",
 540    "metadata": {},
 541    "source": [
 542     "## RNN"
 543    ]
 544   },
 545   {
 546    "cell_type": "code",
 547    "execution_count": null,
 548    "id": "8c1894e3-5283-4e5e-83ae-9c386836a990",
 549    "metadata": {},
 550    "outputs": [],
 551    "source": [
 552     "import importlib \n",
 553     "import moisture_rnn\n",
 554     "importlib.reload(moisture_rnn)\n",
 555     "from moisture_rnn import RNN"
 556    ]
 557   },
 558   {
 559    "cell_type": "code",
 560    "execution_count": null,
 561    "id": "aa1b690f-edaa-4c97-893c-ec9a3a615ce1",
 562    "metadata": {},
 563    "outputs": [],
 564    "source": [
 565     "params = read_yml(\"params.yaml\", subkey=\"lstm\")\n",
 566     "params = RNNParams(params)\n",
 567     "params.update({\n",
 568     "    'dense_layers': 2,\n",
 569     "    'dense_units': 32\n",
 570     "})"
 571    ]
 572   },
 573   {
 574    "cell_type": "code",
 575    "execution_count": null,
 576    "id": "054ab015-4e41-4255-8b1a-843b61e3d21d",
 577    "metadata": {},
 578    "outputs": [],
 579    "source": [
 580     "params.update({'batch_schedule_type': 'step'})"
 581    ]
 582   },
 583   {
 584    "cell_type": "code",
 585    "execution_count": null,
 586    "id": "fa38f35a-d367-4df8-b2d3-7691ff4b0cf4",
 587    "metadata": {},
 588    "outputs": [],
 589    "source": [
 590     "rnn_dat = rnn_data_wrap(combine_nested(train3), params)\n",
 591     "reproducibility.set_seed(123)\n",
 592     "rnn = RNN(params)"
 593    ]
 594   },
 595   {
 596    "cell_type": "code",
 597    "execution_count": null,
 598    "id": "27d11b75-89e9-43a9-8801-7be7fb845b09",
 599    "metadata": {},
 600    "outputs": [],
 601    "source": [
 602     "rnn.model_train.summary()"
 603    ]
 604   },
 605   {
 606    "cell_type": "code",
 607    "execution_count": null,
 608    "id": "b9a0b3fb-aaab-4948-b6e6-824e9dcb92a7",
 609    "metadata": {},
 610    "outputs": [],
 611    "source": [
 612     "rnn.model_predict.summary()"
 613    ]
 614   },
 615   {
 616    "cell_type": "code",
 617    "execution_count": null,
 618    "id": "ade176b9-2844-43b6-b85e-5bb30414aa35",
 619    "metadata": {},
 620    "outputs": [],
 621    "source": [
 622     "rnn.params"
 623    ]
 624   },
 625   {
 626    "cell_type": "code",
 627    "execution_count": null,
 628    "id": "5945e6c1-6b3a-4b7d-ade2-b5788860ef18",
 629    "metadata": {},
 630    "outputs": [],
 631    "source": [
 632     "rnn.model_train.fit(rnn_dat.X_train, rnn_dat.y_train, validation_data=(rnn_dat.X_val, rnn_dat.y_val), \n",
 633     "                    verbose=True, epochs=20)"
 634    ]
 635   },
 636   {
 637    "cell_type": "code",
 638    "execution_count": null,
 639    "id": "2d123b2b-047e-4a04-b49e-6629cc22edc6",
 640    "metadata": {},
 641    "outputs": [],
 642    "source": [
 643     "rnn.model_predict.set_weights(rnn.model_train.get_weights())"
 644    ]
 645   },
 646   {
 647    "cell_type": "code",
 648    "execution_count": null,
 649    "id": "db57df64-d2ac-4b91-bbfc-71a5834ddf41",
 650    "metadata": {},
 651    "outputs": [],
 652    "source": [
 653     "rnn.model_predict.summary()"
 654    ]
 655   },
 656   {
 657    "cell_type": "code",
 658    "execution_count": null,
 659    "id": "0466887f-9833-4a6a-a0c7-a4d56f207d33",
 660    "metadata": {},
 661    "outputs": [],
 662    "source": [
 663     "rnn_dat.X_test.shape"
 664    ]
 665   },
 666   {
 667    "cell_type": "code",
 668    "execution_count": null,
 669    "id": "1d3e630c-db69-4603-962e-95c576b45ac9",
 670    "metadata": {},
 671    "outputs": [],
 672    "source": [
 673     "preds = rnn.model_predict.predict(rnn_dat.X_test)"
 674    ]
 675   },
 676   {
 677    "cell_type": "code",
 678    "execution_count": null,
 679    "id": "8b8228a9-5b6d-4de1-8968-d40277edacd2",
 680    "metadata": {},
 681    "outputs": [],
 682    "source": [
 683     "preds.shape"
 684    ]
 685   },
 686   {
 687    "cell_type": "code",
 688    "execution_count": null,
 689    "id": "8b001dd8-ffd7-4fd1-bf11-413515ddc488",
 690    "metadata": {},
 691    "outputs": [],
 692    "source": [
 693     "rnn_dat.X_test.shape"
 694    ]
 695   },
 696   {
 697    "cell_type": "code",
 698    "execution_count": null,
 699    "id": "f96c6dbf-6ca8-451e-abc4-b68b8116871b",
 700    "metadata": {},
 701    "outputs": [],
 702    "source": [
 703     "squared_diff = np.square(preds - rnn_dat.y_test)\n",
 704     "mse = np.mean(squared_diff, axis=(1, 2))\n",
 705     "errs = np.sqrt(mse)\n",
 706     "errs.mean()"
 707    ]
 708   },
 709   {
 710    "cell_type": "markdown",
 711    "id": "5ef092ff-8af1-491a-b0bf-cc3e674330e0",
 712    "metadata": {},
 713    "source": [
 714     "## Phys Initialized"
 715    ]
 716   },
 717   {
 718    "cell_type": "code",
 719    "execution_count": null,
 720    "id": "5488628e-4552-4909-83e9-413fd6878bdd",
 721    "metadata": {},
 722    "outputs": [],
 723    "source": [
 724     "params.update({\n",
 725     "    'epochs':100,\n",
 726     "    'dense_layers': 0,\n",
 727     "    'activation': ['relu', 'relu'],\n",
 728     "    'phys_initialize': False,\n",
 729     "    'dropout': [0,0],\n",
 730     "    'space_fracs': [.8, .1, .1],\n",
 731     "    'scaler': None\n",
 732     "})"
 733    ]
 734   },
 735   {
 736    "cell_type": "code",
 737    "execution_count": null,
 738    "id": "ab7db7d6-949e-457d-90b9-22d9c5aa4739",
 739    "metadata": {},
 740    "outputs": [],
 741    "source": [
 742     "import importlib\n",
 743     "import moisture_rnn\n",
 744     "importlib.reload(moisture_rnn)\n",
 745     "from moisture_rnn import rnn_data_wrap"
 746    ]
 747   },
 748   {
 749    "cell_type": "code",
 750    "execution_count": null,
 751    "id": "d26cf1b2-2fad-409d-888f-4921b0ae4ba8",
 752    "metadata": {},
 753    "outputs": [],
 754    "source": [
 755     "params['scaler'] is None"
 756    ]
 757   },
 758   {
 759    "cell_type": "code",
 760    "execution_count": null,
 761    "id": "1c4627bc-0f90-44e6-9103-2efe5c5f439d",
 762    "metadata": {},
 763    "outputs": [],
 764    "source": [
 765     "rnn_dat = rnn_data_wrap(combine_nested(train3), params)"
 766    ]
 767   },
 768   {
 769    "cell_type": "code",
 770    "execution_count": null,
 771    "id": "56bdf26c-07e7-4e4a-a567-af7dd0f564d9",
 772    "metadata": {},
 773    "outputs": [],
 774    "source": [
 775     "reproducibility.set_seed()\n",
 776     "rnn = RNN(params)\n",
 777     "m, errs = rnn.run_model(rnn_dat)"
 778    ]
 779   },
 780   {
 781    "cell_type": "code",
 782    "execution_count": null,
 783    "id": "01227b79-98f3-4931-bdfc-ff08afa8be5f",
 784    "metadata": {},
 785    "outputs": [],
 786    "source": [
 787     "rnn.model_train.summary()"
 788    ]
 789   },
 790   {
 791    "cell_type": "code",
 792    "execution_count": null,
 793    "id": "918a8bf0-638b-4b4b-82fe-c6a1965a72dd",
 794    "metadata": {},
 795    "outputs": [],
 796    "source": [
 797     "errs.mean()"
 798    ]
 799   },
 800   {
 801    "cell_type": "code",
 802    "execution_count": null,
 803    "id": "37fdbb3a-3e83-4541-93b2-982b6d4cbe93",
 804    "metadata": {
 805     "scrolled": true
 806    },
 807    "outputs": [],
 808    "source": [
 809     "rnn_dat.X_train[:,:,0].mean()"
 810    ]
 811   },
 812   {
 813    "cell_type": "code",
 814    "execution_count": null,
 815    "id": "7ca41db1-72aa-44b6-b9dd-058735336ab3",
 816    "metadata": {},
 817    "outputs": [],
 818    "source": []
 819   },
 820   {
 821    "cell_type": "code",
 822    "execution_count": null,
 823    "id": "a592a4c9-cb3b-4174-8eaa-02afd00a1897",
 824    "metadata": {},
 825    "outputs": [],
 826    "source": [
 827     "rnn_dat['features_list']"
 828    ]
 829   },
 830   {
 831    "cell_type": "code",
 832    "execution_count": null,
 833    "id": "3832fb05-417c-4648-8e2e-7748c06b3768",
 834    "metadata": {},
 835    "outputs": [],
 836    "source": []
 837   },
 838   {
 839    "cell_type": "markdown",
 840    "id": "d2360aef-e9c4-4a71-922d-336e53b82537",
 841    "metadata": {},
 842    "source": [
 843     "## LSTM"
 844    ]
 845   },
 846   {
 847    "cell_type": "code",
 848    "execution_count": null,
 849    "id": "71d4e441-9bf1-4d57-bb37-091553e23212",
 850    "metadata": {},
 851    "outputs": [],
 852    "source": [
 853     "import importlib \n",
 854     "import moisture_rnn\n",
 855     "importlib.reload(moisture_rnn)\n",
 856     "from moisture_rnn import RNN_LSTM"
 857    ]
 858   },
 859   {
 860    "cell_type": "code",
 861    "execution_count": null,
 862    "id": "0f6ba896-e3be-4a9f-8a42-3df64aff7d63",
 863    "metadata": {},
 864    "outputs": [],
 865    "source": [
 866     "params = read_yml(\"params.yaml\", subkey=\"lstm\")\n",
 867     "params = RNNParams(params)"
 868    ]
 869   },
 870   {
 871    "cell_type": "code",
 872    "execution_count": null,
 873    "id": "a4cf567e-d623-4e14-b578-eed88b80d04e",
 874    "metadata": {},
 875    "outputs": [],
 876    "source": [
 877     "rnn_dat = rnn_data_wrap(combine_nested(train3), params)"
 878    ]
 879   },
 880   {
 881    "cell_type": "code",
 882    "execution_count": null,
 883    "id": "57bb5708-7be9-4474-abb4-3b7ff4bf79df",
 884    "metadata": {},
 885    "outputs": [],
 886    "source": [
 887     "params.update({\n",
 888     "    'loc_batch_reset': rnn_dat.n_seqs # Used to reset hidden state when location changes for a given batch\n",
 889     "})"
 890    ]
 891   },
 892   {
 893    "cell_type": "code",
 894    "execution_count": null,
 895    "id": "0157a6bc-3a99-4b87-a42c-ab770d19ae37",
 896    "metadata": {},
 897    "outputs": [],
 898    "source": [
 899     "from moisture_rnn import ResetStatesCallback, EarlyStoppingCallback\n",
 900     "params.update({'epochs': 50, 'learning_rate': 0.001, 'verbose_fit': True, 'rnn_layers': 2, 'rnn_units': 20, 'dense_layers': 1, 'dense_units': 10,\n",
 901     "              'activation': ['tanh', 'tanh'], 'features_list': rnn_dat.features_list,\n",
 902     "              'batch_schedule_type':'step', 'bmin': 10, 'bmax':rnn_dat.hours})\n",
 903     "reproducibility.set_seed(123)\n",
 904     "lstm = RNN_LSTM(params)\n",
 905     "\n",
 906     "history = lstm.model_train.fit(rnn_dat.X_train, rnn_dat.y_train, \n",
 907     "                    batch_size = params['batch_size'], epochs=params['epochs'], \n",
 908     "                    callbacks = [ResetStatesCallback(params),\n",
 909     "                                EarlyStoppingCallback(patience = 15)],\n",
 910     "                   validation_data = (rnn_dat.X_val, rnn_dat.y_val))\n",
 911     "              "
 912    ]
 913   },
 914   {
 915    "cell_type": "code",
 916    "execution_count": null,
 917    "id": "9b3c8d8d-ea50-44ea-8c0c-414e07cd01ac",
 918    "metadata": {},
 919    "outputs": [],
 920    "source": []
 921   },
 922   {
 923    "cell_type": "code",
 924    "execution_count": null,
 925    "id": "03063e3c-e8f4-451d-b0cf-25bd965cd9d6",
 926    "metadata": {},
 927    "outputs": [],
 928    "source": [
 929     "params = RNNParams(read_yml(\"params.yaml\", subkey=\"lstm\"))\n",
 930     "params.update({'epochs': 50, 'learning_rate': 0.001, 'verbose_fit': True, 'rnn_layers': 2, 'rnn_units': 20, 'dense_layers': 1, 'dense_units': 10,\n",
 931     "              'activation': ['tanh', 'tanh'], 'features_list': rnn_dat.features_list,\n",
 932     "              'batch_schedule_type':'step', 'bmin': 10, 'bmax':rnn_dat.hours})\n",
 933     "rnn_dat = rnn_data_wrap(combine_nested(train3), params)\n",
 934     "params.update({\n",
 935     "    'loc_batch_reset': rnn_dat.n_seqs # Used to reset hidden state when location changes for a given batch\n",
 936     "})\n",
 937     "reproducibility.set_seed(123)\n",
 938     "lstm = RNN_LSTM(params)\n",
 939     "m, errs = lstm.run_model(rnn_dat)"
 940    ]
 941   },
 942   {
 943    "cell_type": "code",
 944    "execution_count": null,
 945    "id": "be46a2dc-bf5c-4893-a1ee-a1682566f7a2",
 946    "metadata": {},
 947    "outputs": [],
 948    "source": [
 949     "errs.mean()"
 950    ]
 951   },
 952   {
 953    "cell_type": "code",
 954    "execution_count": null,
 955    "id": "0f319f37-7d13-41fd-95fa-66dbdfeab588",
 956    "metadata": {},
 957    "outputs": [],
 958    "source": []
 959   },
 960   {
 961    "cell_type": "code",
 962    "execution_count": null,
 963    "id": "b1252b08-62b9-4d24-add2-0f87d15b0ff2",
 964    "metadata": {},
 965    "outputs": [],
 966    "source": [
 967     "params = RNNParams(read_yml(\"params.yaml\", subkey=\"rnn\"))\n",
 968     "rnn_dat = rnn_data_wrap(combine_nested(train3), params)"
 969    ]
 970   },
 971   {
 972    "cell_type": "code",
 973    "execution_count": null,
 974    "id": "9281540b-eb26-4923-883b-1b31d8347634",
 975    "metadata": {},
 976    "outputs": [],
 977    "source": [
 978     "reproducibility.set_seed(123)\n",
 979     "rnn = RNN(params)\n",
 980     "m, errs = rnn.run_model(rnn_dat)"
 981    ]
 982   },
 983   {
 984    "cell_type": "code",
 985    "execution_count": null,
 986    "id": "8a0269b4-d6b7-4f20-8386-69814d7acaa3",
 987    "metadata": {},
 988    "outputs": [],
 989    "source": [
 990     "errs.mean()"
 991    ]
 992   },
 993   {
 994    "cell_type": "code",
 995    "execution_count": null,
 996    "id": "10b44de3-a0e9-49e4-9e03-873d69580c07",
 997    "metadata": {},
 998    "outputs": [],
 999    "source": []
1000   },
1001   {
1002    "cell_type": "code",
1003    "execution_count": null,
1004    "id": "27f4fee4-7fce-49c5-a455-97a90b754c13",
1005    "metadata": {},
1006    "outputs": [],
1007    "source": []
1008   },
1009   {
1010    "cell_type": "code",
1011    "execution_count": null,
1012    "id": "739d4b26-641e-47b2-a90a-67cd32215d05",
1013    "metadata": {},
1014    "outputs": [],
1015    "source": []
1016   }
1017  ],
1018  "metadata": {
1019   "kernelspec": {
1020    "display_name": "Python 3 (ipykernel)",
1021    "language": "python",
1022    "name": "python3"
1023   },
1024   "language_info": {
1025    "codemirror_mode": {
1026     "name": "ipython",
1027     "version": 3
1028    },
1029    "file_extension": ".py",
1030    "mimetype": "text/x-python",
1031    "name": "python",
1032    "nbconvert_exporter": "python",
1033    "pygments_lexer": "ipython3",
1034    "version": "3.12.5"
1035   }
1036  },
1037  "nbformat": 4,
1038  "nbformat_minor": 5
1039 }