anca_ae.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "ceb18c32",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "import h5py\n",
    "import time\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from pathlib import Path\n",
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "79e2690c",
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "def show_scatter(data: np.ndarray, color: np.ndarray):\n",
    "    fig = plt.figure()\n",
    "    ax = fig.add_subplot(111, projection='3d')\n",
    "    ax.set_facecolor('white')\n",
    "    ff = ax.scatter(data[:, 0], data[:, 1], data[:, 2], c=color)\n",
    "    plt.colorbar(ff)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "421b4551",
   "metadata": {},
   "outputs": [],
   "source": [
    "results_dir = Path(\"anca_ae_results\")\n",
    "results_dir.mkdir(exist_ok=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "a7cb3497",
   "metadata": {},
   "outputs": [],
   "source": [
    "data_files = list(Path(\"data\").glob(\"**/*.npy\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "abe5a216",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1107"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(data_files)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "271b85f1",
   "metadata": {},
   "outputs": [],
   "source": [
    "positions = [np.load(p) for p in data_files]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "07f2385c",
   "metadata": {},
   "outputs": [],
   "source": [
    "positions = np.concatenate(positions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "75ff9765",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(866708, 3, 665)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "positions.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "56f12013",
   "metadata": {},
   "outputs": [],
   "source": [
    "from mdlearn.data.preprocess.align import iterative_means_align"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "105c58b9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of coords array in iterative_means: (866708, 3, 665)\n",
      "Iteration #1 with an error of 5.4459624439403855\n",
      "Iteration #2 with an error of 0.0010506558835777014\n",
      "Iteration #3 with an error of 2.984254692027876e-06\n"
     ]
    }
   ],
   "source": [
    "itr, avg_coords, e_rmsd, coords = iterative_means_align(\n",
    "    positions, eps=0.001, max_iter=10, num_workers=75, verbose=True\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "f481ab57",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.save(results_dir / \"coords.npy\", coords)\n",
    "np.save(results_dir / \"e_rmsd.npy\", e_rmsd)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "70602fce",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[<AxesSubplot:title={'center':'0'}>]], dtype=object)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pd.DataFrame(e_rmsd[-1]).hist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "ef03d420",
   "metadata": {},
   "outputs": [],
   "source": [
    "subset_coords = coords[::10, :, :]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "ef09e159",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(86671, 3, 665)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "subset_coords.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "aec60a7c",
   "metadata": {},
   "outputs": [],
   "source": [
    "from mdlearn.data.preprocess.decorrelation.spatial import SD2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "127312b3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2nd order Spatial Decorrelation -> Looking for 1995 sources\n",
      "2nd order Spatial Decorrelation -> Removing the mean value\n",
      "2nd order Spatial Decorrelation -> Whitening the data\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/lambda_stor/homes/abrace/src/mdlearn/mdlearn/data/preprocess/decorrelation/spatial.py:99: RuntimeWarning: invalid value encountered in sqrt\n",
      "  scales = np.sqrt(S)\n"
     ]
    }
   ],
   "source": [
    "Y, S, B, U = SD2(subset_coords.reshape(-1, 665*3), m=665*3, verbose=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "1b518d69",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f2f9d967bd0>]"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "plt.semilogy(S, 'ro-')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "9ed8c94e",
   "metadata": {},
   "outputs": [],
   "source": [
    "from mdlearn.data.preprocess.decorrelation.spatial import SD4"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "63b03bfc",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4th order Spatial Decorrelation -> Estimating cumulant matrices\n",
      "SD4 -> Contrast optimization by joint diagonalization\n",
      "SD4 -> Sweep #0\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #1\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #2\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #3\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #4\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #5\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #6\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #7\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #8\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #9\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #10\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #11\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #12\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #13\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #14\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #15\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #16\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #17\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #18\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #19\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #20\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #21\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #22\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #23\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #24\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #25\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #26\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #27\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #28\n",
      "completed in 779 rotations\n",
      "SD4 -> Sweep #29\n",
      "completed in 779 rotations\n",
      "SD4 -> Sweep #30\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #31\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #32\n",
      "completed in 778 rotations\n",
      "SD4 -> Sweep #33\n",
      "completed in 779 rotations\n",
      "SD4 -> Sweep #34\n",
      "completed in 779 rotations\n",
      "SD4 -> Sweep #35\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #36\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #37\n",
      "completed in 779 rotations\n",
      "SD4 -> Sweep #38\n",
      "completed in 779 rotations\n",
      "SD4 -> Sweep #39\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #40\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #41\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #42\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #43\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #44\n",
      "completed in 779 rotations\n",
      "SD4 -> Sweep #45\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #46\n",
      "completed in 777 rotations\n",
      "SD4 -> Sweep #47\n",
      "completed in 779 rotations\n",
      "SD4 -> Sweep #48\n",
      "completed in 779 rotations\n",
      "SD4 -> Sweep #49\n",
      "completed in 779 rotations\n",
      "SD4 -> Sweep #50\n",
      "completed in 777 rotations\n",
      "SD4 -> Sweep #51\n",
      "completed in 779 rotations\n",
      "SD4 -> Sweep #52\n",
      "completed in 779 rotations\n",
      "SD4 -> Sweep #53\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #54\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #55\n",
      "completed in 779 rotations\n",
      "SD4 -> Sweep #56\n",
      "completed in 778 rotations\n",
      "SD4 -> Sweep #57\n",
      "completed in 778 rotations\n",
      "SD4 -> Sweep #58\n",
      "completed in 779 rotations\n",
      "SD4 -> Sweep #59\n",
      "completed in 777 rotations\n",
      "SD4 -> Sweep #60\n",
      "completed in 780 rotations\n",
      "SD4 -> Sweep #61\n",
      "completed in 778 rotations\n",
      "SD4 -> Sweep #62\n",
      "completed in 776 rotations\n",
      "SD4 -> Sweep #63\n",
      "completed in 779 rotations\n",
      "SD4 -> Sweep #64\n",
      "completed in 778 rotations\n",
      "SD4 -> Sweep #65\n",
      "completed in 779 rotations\n",
      "SD4 -> Sweep #66\n",
      "completed in 778 rotations\n",
      "SD4 -> Sweep #67\n",
      "completed in 778 rotations\n",
      "SD4 -> Sweep #68\n",
      "completed in 778 rotations\n",
      "SD4 -> Sweep #69\n",
      "completed in 778 rotations\n",
      "SD4 -> Sweep #70\n",
      "completed in 778 rotations\n",
      "SD4 -> Sweep #71\n",
      "completed in 779 rotations\n",
      "SD4 -> Sweep #72\n",
      "completed in 777 rotations\n",
      "SD4 -> Sweep #73\n",
      "completed in 778 rotations\n",
      "SD4 -> Sweep #74\n",
      "completed in 776 rotations\n",
      "SD4 -> Sweep #75\n",
      "completed in 776 rotations\n",
      "SD4 -> Sweep #76\n",
      "completed in 776 rotations\n",
      "SD4 -> Sweep #77\n",
      "completed in 775 rotations\n",
      "SD4 -> Sweep #78\n",
      "completed in 776 rotations\n",
      "SD4 -> Sweep #79\n",
      "completed in 776 rotations\n",
      "SD4 -> Sweep #80\n",
      "completed in 776 rotations\n",
      "SD4 -> Sweep #81\n",
      "completed in 776 rotations\n",
      "SD4 -> Sweep #82\n",
      "completed in 776 rotations\n",
      "SD4 -> Sweep #83\n",
      "completed in 776 rotations\n",
      "SD4 -> Sweep #84\n",
      "completed in 778 rotations\n",
      "SD4 -> Sweep #85\n",
      "completed in 777 rotations\n",
      "SD4 -> Sweep #86\n",
      "completed in 776 rotations\n",
      "SD4 -> Sweep #87\n",
      "completed in 777 rotations\n",
      "SD4 -> Sweep #88\n",
      "completed in 775 rotations\n",
      "SD4 -> Sweep #89\n",
      "completed in 776 rotations\n",
      "SD4 -> Sweep #90\n",
      "completed in 775 rotations\n",
      "SD4 -> Sweep #91\n",
      "completed in 775 rotations\n",
      "SD4 -> Sweep #92\n",
      "completed in 775 rotations\n",
      "SD4 -> Sweep #93\n",
      "completed in 774 rotations\n",
      "SD4 -> Sweep #94\n",
      "completed in 774 rotations\n",
      "SD4 -> Sweep #95\n",
      "completed in 774 rotations\n",
      "SD4 -> Sweep #96\n",
      "completed in 772 rotations\n",
      "SD4 -> Sweep #97\n",
      "completed in 771 rotations\n",
      "SD4 -> Sweep #98\n",
      "completed in 772 rotations\n",
      "SD4 -> Sweep #99\n",
      "completed in 772 rotations\n",
      "SD4 -> Sweep #100\n",
      "completed in 770 rotations\n",
      "SD4 -> Sweep #101\n",
      "completed in 772 rotations\n",
      "SD4 -> Sweep #102\n",
      "completed in 770 rotations\n",
      "SD4 -> Sweep #103\n",
      "completed in 770 rotations\n",
      "SD4 -> Sweep #104\n",
      "completed in 769 rotations\n",
      "SD4 -> Sweep #105\n",
      "completed in 769 rotations\n",
      "SD4 -> Sweep #106\n",
      "completed in 770 rotations\n",
      "SD4 -> Sweep #107\n",
      "completed in 767 rotations\n",
      "SD4 -> Sweep #108\n",
      "completed in 766 rotations\n",
      "SD4 -> Sweep #109\n",
      "completed in 769 rotations\n",
      "SD4 -> Sweep #110\n",
      "completed in 765 rotations\n",
      "SD4 -> Sweep #111\n",
      "completed in 766 rotations\n",
      "SD4 -> Sweep #112\n",
      "completed in 765 rotations\n",
      "SD4 -> Sweep #113\n",
      "completed in 765 rotations\n",
      "SD4 -> Sweep #114\n",
      "completed in 762 rotations\n",
      "SD4 -> Sweep #115\n",
      "completed in 766 rotations\n",
      "SD4 -> Sweep #116\n",
      "completed in 761 rotations\n",
      "SD4 -> Sweep #117\n",
      "completed in 762 rotations\n",
      "SD4 -> Sweep #118\n",
      "completed in 760 rotations\n",
      "SD4 -> Sweep #119\n",
      "completed in 763 rotations\n",
      "SD4 -> Sweep #120\n",
      "completed in 759 rotations\n",
      "SD4 -> Sweep #121\n",
      "completed in 757 rotations\n",
      "SD4 -> Sweep #122\n",
      "completed in 754 rotations\n",
      "SD4 -> Sweep #123\n",
      "completed in 760 rotations\n",
      "SD4 -> Sweep #124\n",
      "completed in 752 rotations\n",
      "SD4 -> Sweep #125\n",
      "completed in 752 rotations\n",
      "SD4 -> Sweep #126\n",
      "completed in 749 rotations\n",
      "SD4 -> Sweep #127\n",
      "completed in 748 rotations\n",
      "SD4 -> Sweep #128\n",
      "completed in 749 rotations\n",
      "SD4 -> Sweep #129\n",
      "completed in 746 rotations\n",
      "SD4 -> Sweep #130\n",
      "completed in 735 rotations\n",
      "SD4 -> Sweep #131\n",
      "completed in 747 rotations\n",
      "SD4 -> Sweep #132\n",
      "completed in 728 rotations\n",
      "SD4 -> Sweep #133\n",
      "completed in 734 rotations\n",
      "SD4 -> Sweep #134\n",
      "completed in 725 rotations\n",
      "SD4 -> Sweep #135\n",
      "completed in 738 rotations\n",
      "SD4 -> Sweep #136\n",
      "completed in 721 rotations\n",
      "SD4 -> Sweep #137\n",
      "completed in 724 rotations\n",
      "SD4 -> Sweep #138\n",
      "completed in 724 rotations\n",
      "SD4 -> Sweep #139\n",
      "completed in 720 rotations\n",
      "SD4 -> Sweep #140\n",
      "completed in 714 rotations\n",
      "SD4 -> Sweep #141\n",
      "completed in 711 rotations\n",
      "SD4 -> Sweep #142\n",
      "completed in 706 rotations\n",
      "SD4 -> Sweep #143\n",
      "completed in 702 rotations\n",
      "SD4 -> Sweep #144\n",
      "completed in 706 rotations\n",
      "SD4 -> Sweep #145\n",
      "completed in 693 rotations\n",
      "SD4 -> Sweep #146\n",
      "completed in 694 rotations\n",
      "SD4 -> Sweep #147\n",
      "completed in 688 rotations\n",
      "SD4 -> Sweep #148\n",
      "completed in 686 rotations\n",
      "SD4 -> Sweep #149\n",
      "completed in 670 rotations\n",
      "SD4 -> Sweep #150\n",
      "completed in 686 rotations\n",
      "SD4 -> Sweep #151\n",
      "completed in 657 rotations\n",
      "SD4 -> Sweep #152\n",
      "completed in 666 rotations\n",
      "SD4 -> Sweep #153\n",
      "completed in 655 rotations\n",
      "SD4 -> Sweep #154\n",
      "completed in 658 rotations\n",
      "SD4 -> Sweep #155\n",
      "completed in 630 rotations\n",
      "SD4 -> Sweep #156\n",
      "completed in 654 rotations\n",
      "SD4 -> Sweep #157\n",
      "completed in 620 rotations\n",
      "SD4 -> Sweep #158\n",
      "completed in 621 rotations\n",
      "SD4 -> Sweep #159\n",
      "completed in 613 rotations\n",
      "SD4 -> Sweep #160\n",
      "completed in 609 rotations\n",
      "SD4 -> Sweep #161\n",
      "completed in 601 rotations\n",
      "SD4 -> Sweep #162\n",
      "completed in 598 rotations\n",
      "SD4 -> Sweep #163\n",
      "completed in 573 rotations\n",
      "SD4 -> Sweep #164\n",
      "completed in 574 rotations\n",
      "SD4 -> Sweep #165\n",
      "completed in 558 rotations\n",
      "SD4 -> Sweep #166\n",
      "completed in 563 rotations\n",
      "SD4 -> Sweep #167\n",
      "completed in 548 rotations\n",
      "SD4 -> Sweep #168\n",
      "completed in 543 rotations\n",
      "SD4 -> Sweep #169\n",
      "completed in 513 rotations\n",
      "SD4 -> Sweep #170\n",
      "completed in 531 rotations\n",
      "SD4 -> Sweep #171\n",
      "completed in 501 rotations\n",
      "SD4 -> Sweep #172\n",
      "completed in 493 rotations\n",
      "SD4 -> Sweep #173\n",
      "completed in 500 rotations\n",
      "SD4 -> Sweep #174\n",
      "completed in 475 rotations\n",
      "SD4 -> Sweep #175\n",
      "completed in 479 rotations\n",
      "SD4 -> Sweep #176\n",
      "completed in 449 rotations\n",
      "SD4 -> Sweep #177\n",
      "completed in 463 rotations\n",
      "SD4 -> Sweep #178\n",
      "completed in 443 rotations\n",
      "SD4 -> Sweep #179\n",
      "completed in 431 rotations\n",
      "SD4 -> Sweep #180\n",
      "completed in 404 rotations\n",
      "SD4 -> Sweep #181\n",
      "completed in 427 rotations\n",
      "SD4 -> Sweep #182\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "completed in 393 rotations\n",
      "SD4 -> Sweep #183\n",
      "completed in 400 rotations\n",
      "SD4 -> Sweep #184\n",
      "completed in 378 rotations\n",
      "SD4 -> Sweep #185\n",
      "completed in 376 rotations\n",
      "SD4 -> Sweep #186\n",
      "completed in 359 rotations\n",
      "SD4 -> Sweep #187\n",
      "completed in 345 rotations\n",
      "SD4 -> Sweep #188\n",
      "completed in 336 rotations\n",
      "SD4 -> Sweep #189\n",
      "completed in 345 rotations\n",
      "SD4 -> Sweep #190\n",
      "completed in 318 rotations\n",
      "SD4 -> Sweep #191\n",
      "completed in 338 rotations\n",
      "SD4 -> Sweep #192\n",
      "completed in 282 rotations\n",
      "SD4 -> Sweep #193\n",
      "completed in 305 rotations\n",
      "SD4 -> Sweep #194\n",
      "completed in 284 rotations\n",
      "SD4 -> Sweep #195\n",
      "completed in 283 rotations\n",
      "SD4 -> Sweep #196\n",
      "completed in 254 rotations\n",
      "SD4 -> Sweep #197\n",
      "completed in 273 rotations\n",
      "SD4 -> Sweep #198\n",
      "completed in 248 rotations\n",
      "SD4 -> Sweep #199\n",
      "completed in 239 rotations\n",
      "SD4 -> Sweep #200\n",
      "completed in 245 rotations\n",
      "SD4 -> Sweep #201\n",
      "completed in 246 rotations\n",
      "SD4 -> Sweep #202\n",
      "completed in 211 rotations\n",
      "SD4 -> Sweep #203\n",
      "completed in 189 rotations\n",
      "SD4 -> Sweep #204\n",
      "completed in 206 rotations\n",
      "SD4 -> Sweep #205\n",
      "completed in 201 rotations\n",
      "SD4 -> Sweep #206\n",
      "completed in 184 rotations\n",
      "SD4 -> Sweep #207\n",
      "completed in 200 rotations\n",
      "SD4 -> Sweep #208\n",
      "completed in 173 rotations\n",
      "SD4 -> Sweep #209\n",
      "completed in 168 rotations\n",
      "SD4 -> Sweep #210\n",
      "completed in 159 rotations\n",
      "SD4 -> Sweep #211\n",
      "completed in 161 rotations\n",
      "SD4 -> Sweep #212\n",
      "completed in 150 rotations\n",
      "SD4 -> Sweep #213\n",
      "completed in 152 rotations\n",
      "SD4 -> Sweep #214\n",
      "completed in 147 rotations\n",
      "SD4 -> Sweep #215\n",
      "completed in 134 rotations\n",
      "SD4 -> Sweep #216\n",
      "completed in 121 rotations\n",
      "SD4 -> Sweep #217\n",
      "completed in 121 rotations\n",
      "SD4 -> Sweep #218\n",
      "completed in 125 rotations\n",
      "SD4 -> Sweep #219\n",
      "completed in 116 rotations\n",
      "SD4 -> Sweep #220\n",
      "completed in 112 rotations\n",
      "SD4 -> Sweep #221\n",
      "completed in 91 rotations\n",
      "SD4 -> Sweep #222\n",
      "completed in 119 rotations\n",
      "SD4 -> Sweep #223\n",
      "completed in 96 rotations\n",
      "SD4 -> Sweep #224\n",
      "completed in 111 rotations\n",
      "SD4 -> Sweep #225\n",
      "completed in 91 rotations\n",
      "SD4 -> Sweep #226\n",
      "completed in 80 rotations\n",
      "SD4 -> Sweep #227\n",
      "completed in 71 rotations\n",
      "SD4 -> Sweep #228\n",
      "completed in 69 rotations\n",
      "SD4 -> Sweep #229\n",
      "completed in 53 rotations\n",
      "SD4 -> Sweep #230\n",
      "completed in 65 rotations\n",
      "SD4 -> Sweep #231\n",
      "completed in 66 rotations\n",
      "SD4 -> Sweep #232\n",
      "completed in 56 rotations\n",
      "SD4 -> Sweep #233\n",
      "completed in 55 rotations\n",
      "SD4 -> Sweep #234\n",
      "completed in 50 rotations\n",
      "SD4 -> Sweep #235\n",
      "completed in 36 rotations\n",
      "SD4 -> Sweep #236\n",
      "completed in 34 rotations\n",
      "SD4 -> Sweep #237\n",
      "completed in 32 rotations\n",
      "SD4 -> Sweep #238\n",
      "completed in 42 rotations\n",
      "SD4 -> Sweep #239\n",
      "completed in 27 rotations\n",
      "SD4 -> Sweep #240\n",
      "completed in 35 rotations\n",
      "SD4 -> Sweep #241\n",
      "completed in 34 rotations\n",
      "SD4 -> Sweep #242\n",
      "completed in 28 rotations\n",
      "SD4 -> Sweep #243\n",
      "completed in 17 rotations\n",
      "SD4 -> Sweep #244\n",
      "completed in 14 rotations\n",
      "SD4 -> Sweep #245\n",
      "completed in 9 rotations\n",
      "SD4 -> Sweep #246\n",
      "completed in 22 rotations\n",
      "SD4 -> Sweep #247\n",
      "completed in 21 rotations\n",
      "SD4 -> Sweep #248\n",
      "completed in 14 rotations\n",
      "SD4 -> Sweep #249\n",
      "completed in 10 rotations\n",
      "SD4 -> Sweep #250\n",
      "completed in 11 rotations\n",
      "SD4 -> Sweep #251\n",
      "completed in 7 rotations\n",
      "SD4 -> Sweep #252\n",
      "completed in 1 rotations\n",
      "SD4 -> Sweep #253\n",
      "completed in 2 rotations\n",
      "SD4 -> Sweep #254\n",
      "completed in 3 rotations\n",
      "SD4 -> Sweep #255\n",
      "completed in 2 rotations\n",
      "SD4 -> Sweep #256\n",
      "completed in 0 rotations\n",
      "SD4 -> Total of 143423 Givens rotations\n",
      "SD4 -> Sorting the components\n",
      "SD4 -> Fixing the signs\n"
     ]
    }
   ],
   "source": [
    "\n",
    "# Use the first 60 principal components\n",
    "W = SD4(Y[0:40,:], m=40, U=U[0:40,:], verbose=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "798c6736",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.save(results_dir / \"W.npy\", W)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "5f6c7232",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(40, 1995)"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "W.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "4b5ebac3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1995, 866708)"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "coordsAll = np.reshape(coords, (len(coords), coords.shape[1] * coords.shape[2])).T\n",
    "avgCoordsAll = np.mean(coordsAll, 1)\n",
    "tmpAll = np.reshape(np.tile(avgCoordsAll, coords.shape[0]), (coords.shape[0], coords.shape[1] * coords.shape[2])).T \n",
    "caDevsMDall = coordsAll - tmpAll\n",
    "caDevsMDall.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "1b994f05",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(40, 866708)"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ZPrj4 = W.dot(caDevsMDall)\n",
    "ZPrj4.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "9dfbaa9b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_scatter(ZPrj4.T[:, :3], e_rmsd[-1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "e21f668a",
   "metadata": {},
   "outputs": [],
   "source": [
    "from mdlearn.nn.models.ae.linear import LinearAETrainer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "3aa96d37",
   "metadata": {},
   "outputs": [],
   "source": [
    "trainer = LinearAETrainer(\n",
    "    input_dim=40,\n",
    "    latent_dim=16,\n",
    "    hidden_neurons=[32, 24],\n",
    "    epochs=500,\n",
    "    verbose=True,\n",
    "    checkpoint_log_every=10,\n",
    "    plot_log_every=100,\n",
    "    plot_n_samples=8000,\n",
    "    plot_method=\"raw\",\n",
    "    device=\"cpu\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "a4314416",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'seed': 42,\n",
       " 'scalar_dset_names': [],\n",
       " 'in_gpu_memory': False,\n",
       " 'num_data_workers': 0,\n",
       " 'persistent_workers': False,\n",
       " 'prefetch_factor': 2,\n",
       " 'split_pct': 0.8,\n",
       " 'split_method': 'random',\n",
       " 'batch_size': 128,\n",
       " 'shuffle': True,\n",
       " 'device': device(type='cpu'),\n",
       " 'epochs': 500,\n",
       " 'verbose': True,\n",
       " 'clip_grad_max_norm': 10.0,\n",
       " 'checkpoint_log_every': 10,\n",
       " 'plot_log_every': 100,\n",
       " 'plot_n_samples': 8000,\n",
       " 'plot_method': 'raw',\n",
       " 'train_subsample_pct': 1.0,\n",
       " 'valid_subsample_pct': 1.0,\n",
       " 'use_wandb': False,\n",
       " 'optimizer_name': 'RMSprop',\n",
       " 'optimizer_hparams': {'lr': 0.001, 'weight_decay': 1e-05},\n",
       " 'scheduler_name': None,\n",
       " 'scheduler_hparams': {},\n",
       " 'model': LinearAE(\n",
       "   (encoder): DenseNet(\n",
       "     (model): Sequential(\n",
       "       (0): Linear(in_features=40, out_features=32, bias=True)\n",
       "       (1): ReLU()\n",
       "       (2): Linear(in_features=32, out_features=24, bias=True)\n",
       "       (3): ReLU()\n",
       "       (4): Linear(in_features=24, out_features=16, bias=True)\n",
       "     )\n",
       "   )\n",
       "   (decoder): DenseNet(\n",
       "     (model): Sequential(\n",
       "       (0): Linear(in_features=16, out_features=24, bias=True)\n",
       "       (1): ReLU()\n",
       "       (2): Linear(in_features=24, out_features=32, bias=True)\n",
       "       (3): ReLU()\n",
       "       (4): Linear(in_features=32, out_features=40, bias=True)\n",
       "     )\n",
       "   )\n",
       " ),\n",
       " 'optimizer': RMSprop (\n",
       " Parameter Group 0\n",
       "     alpha: 0.99\n",
       "     centered: False\n",
       "     eps: 1e-08\n",
       "     lr: 0.001\n",
       "     momentum: 0\n",
       "     weight_decay: 1e-05\n",
       " ),\n",
       " 'scheduler': None,\n",
       " 'loss_curve_': {'train': [], 'validation': []}}"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vars(trainer)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "4eb28893",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(866708, 40)"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = ZPrj4.T\n",
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "715c6add",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.save(results_dir / \"X.npy\", X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "id": "f187d5e0",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_subset = X[::10]\n",
    "scalars = {\"rmsd\": e_rmsd[-1][::10]}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "0c5cbb5b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(86671, 40)"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_subset.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "54ce71a9",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "os.environ[\"CUDA_VISBLE_DEVICES\"] = \"6\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "id": "b56254f7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====> Epoch: 1 Train:\tAvg loss: 0.8088\n",
      "====> Epoch: 1 Valid:\tAvg loss: 0.7362\n",
      "\n",
      "====> Epoch: 2 Train:\tAvg loss: 0.6882\n",
      "====> Epoch: 2 Valid:\tAvg loss: 0.6587\n",
      "\n",
      "====> Epoch: 3 Train:\tAvg loss: 0.6283\n",
      "====> Epoch: 3 Valid:\tAvg loss: 0.6056\n",
      "\n",
      "====> Epoch: 4 Train:\tAvg loss: 0.5811\n",
      "====> Epoch: 4 Valid:\tAvg loss: 0.5565\n",
      "\n",
      "====> Epoch: 5 Train:\tAvg loss: 0.5467\n",
      "====> Epoch: 5 Valid:\tAvg loss: 0.5329\n",
      "\n",
      "====> Epoch: 6 Train:\tAvg loss: 0.5279\n",
      "====> Epoch: 6 Valid:\tAvg loss: 0.5222\n",
      "\n",
      "====> Epoch: 7 Train:\tAvg loss: 0.5150\n",
      "====> Epoch: 7 Valid:\tAvg loss: 0.5100\n",
      "\n",
      "====> Epoch: 8 Train:\tAvg loss: 0.5054\n",
      "====> Epoch: 8 Valid:\tAvg loss: 0.4997\n",
      "\n",
      "====> Epoch: 9 Train:\tAvg loss: 0.4979\n",
      "====> Epoch: 9 Valid:\tAvg loss: 0.4916\n",
      "\n",
      "====> Epoch: 10 Train:\tAvg loss: 0.4916\n",
      "====> Epoch: 10 Valid:\tAvg loss: 0.4856\n",
      "\n",
      "====> Epoch: 11 Train:\tAvg loss: 0.4857\n",
      "====> Epoch: 11 Valid:\tAvg loss: 0.4804\n",
      "\n",
      "====> Epoch: 12 Train:\tAvg loss: 0.4774\n",
      "====> Epoch: 12 Valid:\tAvg loss: 0.4713\n",
      "\n",
      "====> Epoch: 13 Train:\tAvg loss: 0.4691\n",
      "====> Epoch: 13 Valid:\tAvg loss: 0.4626\n",
      "\n",
      "====> Epoch: 14 Train:\tAvg loss: 0.4637\n",
      "====> Epoch: 14 Valid:\tAvg loss: 0.4601\n",
      "\n",
      "====> Epoch: 15 Train:\tAvg loss: 0.4591\n",
      "====> Epoch: 15 Valid:\tAvg loss: 0.4569\n",
      "\n",
      "====> Epoch: 16 Train:\tAvg loss: 0.4542\n",
      "====> Epoch: 16 Valid:\tAvg loss: 0.4519\n",
      "\n",
      "====> Epoch: 17 Train:\tAvg loss: 0.4490\n",
      "====> Epoch: 17 Valid:\tAvg loss: 0.4459\n",
      "\n",
      "====> Epoch: 18 Train:\tAvg loss: 0.4448\n",
      "====> Epoch: 18 Valid:\tAvg loss: 0.4447\n",
      "\n",
      "====> Epoch: 19 Train:\tAvg loss: 0.4423\n",
      "====> Epoch: 19 Valid:\tAvg loss: 0.4390\n",
      "\n",
      "====> Epoch: 20 Train:\tAvg loss: 0.4402\n",
      "====> Epoch: 20 Valid:\tAvg loss: 0.4436\n",
      "\n",
      "====> Epoch: 21 Train:\tAvg loss: 0.4387\n",
      "====> Epoch: 21 Valid:\tAvg loss: 0.4403\n",
      "\n",
      "====> Epoch: 22 Train:\tAvg loss: 0.4368\n",
      "====> Epoch: 22 Valid:\tAvg loss: 0.4345\n",
      "\n",
      "====> Epoch: 23 Train:\tAvg loss: 0.4358\n",
      "====> Epoch: 23 Valid:\tAvg loss: 0.4330\n",
      "\n",
      "====> Epoch: 24 Train:\tAvg loss: 0.4346\n",
      "====> Epoch: 24 Valid:\tAvg loss: 0.4322\n",
      "\n",
      "====> Epoch: 25 Train:\tAvg loss: 0.4334\n",
      "====> Epoch: 25 Valid:\tAvg loss: 0.4492\n",
      "\n",
      "====> Epoch: 26 Train:\tAvg loss: 0.4325\n",
      "====> Epoch: 26 Valid:\tAvg loss: 0.4349\n",
      "\n",
      "====> Epoch: 27 Train:\tAvg loss: 0.4317\n",
      "====> Epoch: 27 Valid:\tAvg loss: 0.4319\n",
      "\n",
      "====> Epoch: 28 Train:\tAvg loss: 0.4309\n",
      "====> Epoch: 28 Valid:\tAvg loss: 0.4352\n",
      "\n",
      "====> Epoch: 29 Train:\tAvg loss: 0.4301\n",
      "====> Epoch: 29 Valid:\tAvg loss: 0.4327\n",
      "\n",
      "====> Epoch: 30 Train:\tAvg loss: 0.4291\n",
      "====> Epoch: 30 Valid:\tAvg loss: 0.4338\n",
      "\n",
      "====> Epoch: 31 Train:\tAvg loss: 0.4286\n",
      "====> Epoch: 31 Valid:\tAvg loss: 0.4290\n",
      "\n",
      "====> Epoch: 32 Train:\tAvg loss: 0.4278\n",
      "====> Epoch: 32 Valid:\tAvg loss: 0.4276\n",
      "\n",
      "====> Epoch: 33 Train:\tAvg loss: 0.4268\n",
      "====> Epoch: 33 Valid:\tAvg loss: 0.4290\n",
      "\n",
      "====> Epoch: 34 Train:\tAvg loss: 0.4260\n",
      "====> Epoch: 34 Valid:\tAvg loss: 0.4260\n",
      "\n",
      "====> Epoch: 35 Train:\tAvg loss: 0.4247\n",
      "====> Epoch: 35 Valid:\tAvg loss: 0.4308\n",
      "\n",
      "====> Epoch: 36 Train:\tAvg loss: 0.4232\n",
      "====> Epoch: 36 Valid:\tAvg loss: 0.4240\n",
      "\n",
      "====> Epoch: 37 Train:\tAvg loss: 0.4215\n",
      "====> Epoch: 37 Valid:\tAvg loss: 0.4185\n",
      "\n",
      "====> Epoch: 38 Train:\tAvg loss: 0.4202\n",
      "====> Epoch: 38 Valid:\tAvg loss: 0.4192\n",
      "\n",
      "====> Epoch: 39 Train:\tAvg loss: 0.4187\n",
      "====> Epoch: 39 Valid:\tAvg loss: 0.4186\n",
      "\n",
      "====> Epoch: 40 Train:\tAvg loss: 0.4176\n",
      "====> Epoch: 40 Valid:\tAvg loss: 0.4199\n",
      "\n",
      "====> Epoch: 41 Train:\tAvg loss: 0.4168\n",
      "====> Epoch: 41 Valid:\tAvg loss: 0.4170\n",
      "\n",
      "====> Epoch: 42 Train:\tAvg loss: 0.4158\n",
      "====> Epoch: 42 Valid:\tAvg loss: 0.4144\n",
      "\n",
      "====> Epoch: 43 Train:\tAvg loss: 0.4150\n",
      "====> Epoch: 43 Valid:\tAvg loss: 0.4134\n",
      "\n",
      "====> Epoch: 44 Train:\tAvg loss: 0.4143\n",
      "====> Epoch: 44 Valid:\tAvg loss: 0.4159\n",
      "\n",
      "====> Epoch: 45 Train:\tAvg loss: 0.4134\n",
      "====> Epoch: 45 Valid:\tAvg loss: 0.4175\n",
      "\n",
      "====> Epoch: 46 Train:\tAvg loss: 0.4126\n",
      "====> Epoch: 46 Valid:\tAvg loss: 0.4182\n",
      "\n",
      "====> Epoch: 47 Train:\tAvg loss: 0.4118\n",
      "====> Epoch: 47 Valid:\tAvg loss: 0.4091\n",
      "\n",
      "====> Epoch: 48 Train:\tAvg loss: 0.4110\n",
      "====> Epoch: 48 Valid:\tAvg loss: 0.4145\n",
      "\n",
      "====> Epoch: 49 Train:\tAvg loss: 0.4104\n",
      "====> Epoch: 49 Valid:\tAvg loss: 0.4123\n",
      "\n",
      "====> Epoch: 50 Train:\tAvg loss: 0.4098\n",
      "====> Epoch: 50 Valid:\tAvg loss: 0.4153\n",
      "\n",
      "====> Epoch: 51 Train:\tAvg loss: 0.4089\n",
      "====> Epoch: 51 Valid:\tAvg loss: 0.4059\n",
      "\n",
      "====> Epoch: 52 Train:\tAvg loss: 0.4081\n",
      "====> Epoch: 52 Valid:\tAvg loss: 0.4091\n",
      "\n",
      "====> Epoch: 53 Train:\tAvg loss: 0.4075\n",
      "====> Epoch: 53 Valid:\tAvg loss: 0.4091\n",
      "\n",
      "====> Epoch: 54 Train:\tAvg loss: 0.4069\n",
      "====> Epoch: 54 Valid:\tAvg loss: 0.4119\n",
      "\n",
      "====> Epoch: 55 Train:\tAvg loss: 0.4064\n",
      "====> Epoch: 55 Valid:\tAvg loss: 0.4060\n",
      "\n",
      "====> Epoch: 56 Train:\tAvg loss: 0.4058\n",
      "====> Epoch: 56 Valid:\tAvg loss: 0.4066\n",
      "\n",
      "====> Epoch: 57 Train:\tAvg loss: 0.4053\n",
      "====> Epoch: 57 Valid:\tAvg loss: 0.4025\n",
      "\n",
      "====> Epoch: 58 Train:\tAvg loss: 0.4049\n",
      "====> Epoch: 58 Valid:\tAvg loss: 0.4057\n",
      "\n",
      "====> Epoch: 59 Train:\tAvg loss: 0.4044\n",
      "====> Epoch: 59 Valid:\tAvg loss: 0.4049\n",
      "\n",
      "====> Epoch: 60 Train:\tAvg loss: 0.4042\n",
      "====> Epoch: 60 Valid:\tAvg loss: 0.4042\n",
      "\n",
      "====> Epoch: 61 Train:\tAvg loss: 0.4038\n",
      "====> Epoch: 61 Valid:\tAvg loss: 0.4132\n",
      "\n",
      "====> Epoch: 62 Train:\tAvg loss: 0.4035\n",
      "====> Epoch: 62 Valid:\tAvg loss: 0.4020\n",
      "\n",
      "====> Epoch: 63 Train:\tAvg loss: 0.4031\n",
      "====> Epoch: 63 Valid:\tAvg loss: 0.3988\n",
      "\n",
      "====> Epoch: 64 Train:\tAvg loss: 0.4028\n",
      "====> Epoch: 64 Valid:\tAvg loss: 0.4098\n",
      "\n",
      "====> Epoch: 65 Train:\tAvg loss: 0.4022\n",
      "====> Epoch: 65 Valid:\tAvg loss: 0.4081\n",
      "\n",
      "====> Epoch: 66 Train:\tAvg loss: 0.4019\n",
      "====> Epoch: 66 Valid:\tAvg loss: 0.4016\n",
      "\n",
      "====> Epoch: 67 Train:\tAvg loss: 0.4017\n",
      "====> Epoch: 67 Valid:\tAvg loss: 0.4022\n",
      "\n",
      "====> Epoch: 68 Train:\tAvg loss: 0.4014\n",
      "====> Epoch: 68 Valid:\tAvg loss: 0.4041\n",
      "\n",
      "====> Epoch: 69 Train:\tAvg loss: 0.4010\n",
      "====> Epoch: 69 Valid:\tAvg loss: 0.4116\n",
      "\n",
      "====> Epoch: 70 Train:\tAvg loss: 0.4008\n",
      "====> Epoch: 70 Valid:\tAvg loss: 0.4050\n",
      "\n",
      "====> Epoch: 71 Train:\tAvg loss: 0.4003\n",
      "====> Epoch: 71 Valid:\tAvg loss: 0.4103\n",
      "\n",
      "====> Epoch: 72 Train:\tAvg loss: 0.4002\n",
      "====> Epoch: 72 Valid:\tAvg loss: 0.3990\n",
      "\n",
      "====> Epoch: 73 Train:\tAvg loss: 0.3999\n",
      "====> Epoch: 73 Valid:\tAvg loss: 0.4004\n",
      "\n",
      "====> Epoch: 74 Train:\tAvg loss: 0.3998\n",
      "====> Epoch: 74 Valid:\tAvg loss: 0.3977\n",
      "\n",
      "====> Epoch: 75 Train:\tAvg loss: 0.3994\n",
      "====> Epoch: 75 Valid:\tAvg loss: 0.4059\n",
      "\n",
      "====> Epoch: 76 Train:\tAvg loss: 0.3994\n",
      "====> Epoch: 76 Valid:\tAvg loss: 0.4009\n",
      "\n",
      "====> Epoch: 77 Train:\tAvg loss: 0.3990\n",
      "====> Epoch: 77 Valid:\tAvg loss: 0.4038\n",
      "\n",
      "====> Epoch: 78 Train:\tAvg loss: 0.3991\n",
      "====> Epoch: 78 Valid:\tAvg loss: 0.4037\n",
      "\n",
      "====> Epoch: 79 Train:\tAvg loss: 0.3989\n",
      "====> Epoch: 79 Valid:\tAvg loss: 0.3993\n",
      "\n",
      "====> Epoch: 80 Train:\tAvg loss: 0.3984\n",
      "====> Epoch: 80 Valid:\tAvg loss: 0.3997\n",
      "\n",
      "====> Epoch: 81 Train:\tAvg loss: 0.3984\n",
      "====> Epoch: 81 Valid:\tAvg loss: 0.4051\n",
      "\n",
      "====> Epoch: 82 Train:\tAvg loss: 0.3982\n",
      "====> Epoch: 82 Valid:\tAvg loss: 0.3966\n",
      "\n",
      "====> Epoch: 83 Train:\tAvg loss: 0.3979\n",
      "====> Epoch: 83 Valid:\tAvg loss: 0.3959\n",
      "\n",
      "====> Epoch: 84 Train:\tAvg loss: 0.3979\n",
      "====> Epoch: 84 Valid:\tAvg loss: 0.3972\n",
      "\n",
      "====> Epoch: 85 Train:\tAvg loss: 0.3977\n",
      "====> Epoch: 85 Valid:\tAvg loss: 0.3998\n",
      "\n",
      "====> Epoch: 86 Train:\tAvg loss: 0.3976\n",
      "====> Epoch: 86 Valid:\tAvg loss: 0.3963\n",
      "\n",
      "====> Epoch: 87 Train:\tAvg loss: 0.3973\n",
      "====> Epoch: 87 Valid:\tAvg loss: 0.4096\n",
      "\n",
      "====> Epoch: 88 Train:\tAvg loss: 0.3971\n",
      "====> Epoch: 88 Valid:\tAvg loss: 0.3968\n",
      "\n",
      "====> Epoch: 89 Train:\tAvg loss: 0.3971\n",
      "====> Epoch: 89 Valid:\tAvg loss: 0.3949\n",
      "\n",
      "====> Epoch: 90 Train:\tAvg loss: 0.3968\n",
      "====> Epoch: 90 Valid:\tAvg loss: 0.3994\n",
      "\n",
      "====> Epoch: 91 Train:\tAvg loss: 0.3968\n",
      "====> Epoch: 91 Valid:\tAvg loss: 0.3980\n",
      "\n",
      "====> Epoch: 92 Train:\tAvg loss: 0.3965\n",
      "====> Epoch: 92 Valid:\tAvg loss: 0.3993\n",
      "\n",
      "====> Epoch: 93 Train:\tAvg loss: 0.3965\n",
      "====> Epoch: 93 Valid:\tAvg loss: 0.3978\n",
      "\n",
      "====> Epoch: 94 Train:\tAvg loss: 0.3964\n",
      "====> Epoch: 94 Valid:\tAvg loss: 0.3978\n",
      "\n",
      "====> Epoch: 95 Train:\tAvg loss: 0.3963\n",
      "====> Epoch: 95 Valid:\tAvg loss: 0.3935\n",
      "\n",
      "====> Epoch: 96 Train:\tAvg loss: 0.3962\n",
      "====> Epoch: 96 Valid:\tAvg loss: 0.3941\n",
      "\n",
      "====> Epoch: 97 Train:\tAvg loss: 0.3961\n",
      "====> Epoch: 97 Valid:\tAvg loss: 0.3941\n",
      "\n",
      "====> Epoch: 98 Train:\tAvg loss: 0.3959\n",
      "====> Epoch: 98 Valid:\tAvg loss: 0.3983\n",
      "\n",
      "====> Epoch: 99 Train:\tAvg loss: 0.3957\n",
      "====> Epoch: 99 Valid:\tAvg loss: 0.3997\n",
      "\n",
      "====> Epoch: 100 Train:\tAvg loss: 0.3957\n",
      "====> Epoch: 100 Valid:\tAvg loss: 0.3941\n",
      "\n",
      "====> Epoch: 101 Train:\tAvg loss: 0.3955\n",
      "====> Epoch: 101 Valid:\tAvg loss: 0.3965\n",
      "\n",
      "====> Epoch: 102 Train:\tAvg loss: 0.3956\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====> Epoch: 102 Valid:\tAvg loss: 0.3983\n",
      "\n",
      "====> Epoch: 103 Train:\tAvg loss: 0.3953\n",
      "====> Epoch: 103 Valid:\tAvg loss: 0.3954\n",
      "\n",
      "====> Epoch: 104 Train:\tAvg loss: 0.3951\n",
      "====> Epoch: 104 Valid:\tAvg loss: 0.3952\n",
      "\n",
      "====> Epoch: 105 Train:\tAvg loss: 0.3951\n",
      "====> Epoch: 105 Valid:\tAvg loss: 0.3963\n",
      "\n",
      "====> Epoch: 106 Train:\tAvg loss: 0.3951\n",
      "====> Epoch: 106 Valid:\tAvg loss: 0.3980\n",
      "\n",
      "====> Epoch: 107 Train:\tAvg loss: 0.3949\n",
      "====> Epoch: 107 Valid:\tAvg loss: 0.3929\n",
      "\n",
      "====> Epoch: 108 Train:\tAvg loss: 0.3949\n",
      "====> Epoch: 108 Valid:\tAvg loss: 0.3954\n",
      "\n",
      "====> Epoch: 109 Train:\tAvg loss: 0.3947\n",
      "====> Epoch: 109 Valid:\tAvg loss: 0.4000\n",
      "\n",
      "====> Epoch: 110 Train:\tAvg loss: 0.3948\n",
      "====> Epoch: 110 Valid:\tAvg loss: 0.3939\n",
      "\n",
      "====> Epoch: 111 Train:\tAvg loss: 0.3945\n",
      "====> Epoch: 111 Valid:\tAvg loss: 0.3990\n",
      "\n",
      "====> Epoch: 112 Train:\tAvg loss: 0.3945\n",
      "====> Epoch: 112 Valid:\tAvg loss: 0.3978\n",
      "\n",
      "====> Epoch: 113 Train:\tAvg loss: 0.3944\n",
      "====> Epoch: 113 Valid:\tAvg loss: 0.3948\n",
      "\n",
      "====> Epoch: 114 Train:\tAvg loss: 0.3943\n",
      "====> Epoch: 114 Valid:\tAvg loss: 0.4048\n",
      "\n",
      "====> Epoch: 115 Train:\tAvg loss: 0.3941\n",
      "====> Epoch: 115 Valid:\tAvg loss: 0.3950\n",
      "\n",
      "====> Epoch: 116 Train:\tAvg loss: 0.3941\n",
      "====> Epoch: 116 Valid:\tAvg loss: 0.3952\n",
      "\n",
      "====> Epoch: 117 Train:\tAvg loss: 0.3939\n",
      "====> Epoch: 117 Valid:\tAvg loss: 0.3938\n",
      "\n",
      "====> Epoch: 118 Train:\tAvg loss: 0.3940\n",
      "====> Epoch: 118 Valid:\tAvg loss: 0.3947\n",
      "\n",
      "====> Epoch: 119 Train:\tAvg loss: 0.3937\n",
      "====> Epoch: 119 Valid:\tAvg loss: 0.4006\n",
      "\n",
      "====> Epoch: 120 Train:\tAvg loss: 0.3938\n",
      "====> Epoch: 120 Valid:\tAvg loss: 0.3941\n",
      "\n",
      "====> Epoch: 121 Train:\tAvg loss: 0.3935\n",
      "====> Epoch: 121 Valid:\tAvg loss: 0.4003\n",
      "\n",
      "====> Epoch: 122 Train:\tAvg loss: 0.3935\n",
      "====> Epoch: 122 Valid:\tAvg loss: 0.3966\n",
      "\n",
      "====> Epoch: 123 Train:\tAvg loss: 0.3936\n",
      "====> Epoch: 123 Valid:\tAvg loss: 0.3949\n",
      "\n",
      "====> Epoch: 124 Train:\tAvg loss: 0.3933\n",
      "====> Epoch: 124 Valid:\tAvg loss: 0.3955\n",
      "\n",
      "====> Epoch: 125 Train:\tAvg loss: 0.3933\n",
      "====> Epoch: 125 Valid:\tAvg loss: 0.3998\n",
      "\n",
      "====> Epoch: 126 Train:\tAvg loss: 0.3932\n",
      "====> Epoch: 126 Valid:\tAvg loss: 0.3939\n",
      "\n",
      "====> Epoch: 127 Train:\tAvg loss: 0.3929\n",
      "====> Epoch: 127 Valid:\tAvg loss: 0.3990\n",
      "\n",
      "====> Epoch: 128 Train:\tAvg loss: 0.3930\n",
      "====> Epoch: 128 Valid:\tAvg loss: 0.3921\n",
      "\n",
      "====> Epoch: 129 Train:\tAvg loss: 0.3928\n",
      "====> Epoch: 129 Valid:\tAvg loss: 0.3931\n",
      "\n",
      "====> Epoch: 130 Train:\tAvg loss: 0.3928\n",
      "====> Epoch: 130 Valid:\tAvg loss: 0.3979\n",
      "\n",
      "====> Epoch: 131 Train:\tAvg loss: 0.3926\n",
      "====> Epoch: 131 Valid:\tAvg loss: 0.3973\n",
      "\n",
      "====> Epoch: 132 Train:\tAvg loss: 0.3924\n",
      "====> Epoch: 132 Valid:\tAvg loss: 0.3907\n",
      "\n",
      "====> Epoch: 133 Train:\tAvg loss: 0.3925\n",
      "====> Epoch: 133 Valid:\tAvg loss: 0.3924\n",
      "\n",
      "====> Epoch: 134 Train:\tAvg loss: 0.3921\n",
      "====> Epoch: 134 Valid:\tAvg loss: 0.3961\n",
      "\n",
      "====> Epoch: 135 Train:\tAvg loss: 0.3921\n",
      "====> Epoch: 135 Valid:\tAvg loss: 0.3962\n",
      "\n",
      "====> Epoch: 136 Train:\tAvg loss: 0.3919\n",
      "====> Epoch: 136 Valid:\tAvg loss: 0.3952\n",
      "\n",
      "====> Epoch: 137 Train:\tAvg loss: 0.3919\n",
      "====> Epoch: 137 Valid:\tAvg loss: 0.3976\n",
      "\n",
      "====> Epoch: 138 Train:\tAvg loss: 0.3917\n",
      "====> Epoch: 138 Valid:\tAvg loss: 0.3953\n",
      "\n",
      "====> Epoch: 139 Train:\tAvg loss: 0.3917\n",
      "====> Epoch: 139 Valid:\tAvg loss: 0.3951\n",
      "\n",
      "====> Epoch: 140 Train:\tAvg loss: 0.3916\n",
      "====> Epoch: 140 Valid:\tAvg loss: 0.3977\n",
      "\n",
      "====> Epoch: 141 Train:\tAvg loss: 0.3915\n",
      "====> Epoch: 141 Valid:\tAvg loss: 0.3903\n",
      "\n",
      "====> Epoch: 142 Train:\tAvg loss: 0.3915\n",
      "====> Epoch: 142 Valid:\tAvg loss: 0.3929\n",
      "\n",
      "====> Epoch: 143 Train:\tAvg loss: 0.3913\n",
      "====> Epoch: 143 Valid:\tAvg loss: 0.3914\n",
      "\n",
      "====> Epoch: 144 Train:\tAvg loss: 0.3910\n",
      "====> Epoch: 144 Valid:\tAvg loss: 0.3925\n",
      "\n",
      "====> Epoch: 145 Train:\tAvg loss: 0.3913\n",
      "====> Epoch: 145 Valid:\tAvg loss: 0.3937\n",
      "\n",
      "====> Epoch: 146 Train:\tAvg loss: 0.3910\n",
      "====> Epoch: 146 Valid:\tAvg loss: 0.3947\n",
      "\n",
      "====> Epoch: 147 Train:\tAvg loss: 0.3911\n",
      "====> Epoch: 147 Valid:\tAvg loss: 0.3885\n",
      "\n",
      "====> Epoch: 148 Train:\tAvg loss: 0.3910\n",
      "====> Epoch: 148 Valid:\tAvg loss: 0.3921\n",
      "\n",
      "====> Epoch: 149 Train:\tAvg loss: 0.3909\n",
      "====> Epoch: 149 Valid:\tAvg loss: 0.3920\n",
      "\n",
      "====> Epoch: 150 Train:\tAvg loss: 0.3909\n",
      "====> Epoch: 150 Valid:\tAvg loss: 0.3903\n",
      "\n",
      "====> Epoch: 151 Train:\tAvg loss: 0.3908\n",
      "====> Epoch: 151 Valid:\tAvg loss: 0.3928\n",
      "\n",
      "====> Epoch: 152 Train:\tAvg loss: 0.3906\n",
      "====> Epoch: 152 Valid:\tAvg loss: 0.3921\n",
      "\n",
      "====> Epoch: 153 Train:\tAvg loss: 0.3908\n",
      "====> Epoch: 153 Valid:\tAvg loss: 0.3898\n",
      "\n",
      "====> Epoch: 154 Train:\tAvg loss: 0.3907\n",
      "====> Epoch: 154 Valid:\tAvg loss: 0.3907\n",
      "\n",
      "====> Epoch: 155 Train:\tAvg loss: 0.3906\n",
      "====> Epoch: 155 Valid:\tAvg loss: 0.3939\n",
      "\n",
      "====> Epoch: 156 Train:\tAvg loss: 0.3906\n",
      "====> Epoch: 156 Valid:\tAvg loss: 0.3924\n",
      "\n",
      "====> Epoch: 157 Train:\tAvg loss: 0.3905\n",
      "====> Epoch: 157 Valid:\tAvg loss: 0.3946\n",
      "\n",
      "====> Epoch: 158 Train:\tAvg loss: 0.3904\n",
      "====> Epoch: 158 Valid:\tAvg loss: 0.3914\n",
      "\n",
      "====> Epoch: 159 Train:\tAvg loss: 0.3905\n",
      "====> Epoch: 159 Valid:\tAvg loss: 0.3943\n",
      "\n",
      "====> Epoch: 160 Train:\tAvg loss: 0.3904\n",
      "====> Epoch: 160 Valid:\tAvg loss: 0.3888\n",
      "\n",
      "====> Epoch: 161 Train:\tAvg loss: 0.3903\n",
      "====> Epoch: 161 Valid:\tAvg loss: 0.3930\n",
      "\n",
      "====> Epoch: 162 Train:\tAvg loss: 0.3904\n",
      "====> Epoch: 162 Valid:\tAvg loss: 0.3950\n",
      "\n",
      "====> Epoch: 163 Train:\tAvg loss: 0.3902\n",
      "====> Epoch: 163 Valid:\tAvg loss: 0.3977\n",
      "\n",
      "====> Epoch: 164 Train:\tAvg loss: 0.3903\n",
      "====> Epoch: 164 Valid:\tAvg loss: 0.3948\n",
      "\n",
      "====> Epoch: 165 Train:\tAvg loss: 0.3901\n",
      "====> Epoch: 165 Valid:\tAvg loss: 0.3994\n",
      "\n",
      "====> Epoch: 166 Train:\tAvg loss: 0.3901\n",
      "====> Epoch: 166 Valid:\tAvg loss: 0.3931\n",
      "\n",
      "====> Epoch: 167 Train:\tAvg loss: 0.3901\n",
      "====> Epoch: 167 Valid:\tAvg loss: 0.3971\n",
      "\n",
      "====> Epoch: 168 Train:\tAvg loss: 0.3900\n",
      "====> Epoch: 168 Valid:\tAvg loss: 0.3926\n",
      "\n",
      "====> Epoch: 169 Train:\tAvg loss: 0.3901\n",
      "====> Epoch: 169 Valid:\tAvg loss: 0.3924\n",
      "\n",
      "====> Epoch: 170 Train:\tAvg loss: 0.3900\n",
      "====> Epoch: 170 Valid:\tAvg loss: 0.3902\n",
      "\n",
      "====> Epoch: 171 Train:\tAvg loss: 0.3900\n",
      "====> Epoch: 171 Valid:\tAvg loss: 0.3913\n",
      "\n",
      "====> Epoch: 172 Train:\tAvg loss: 0.3900\n",
      "====> Epoch: 172 Valid:\tAvg loss: 0.3954\n",
      "\n",
      "====> Epoch: 173 Train:\tAvg loss: 0.3899\n",
      "====> Epoch: 173 Valid:\tAvg loss: 0.3901\n",
      "\n",
      "====> Epoch: 174 Train:\tAvg loss: 0.3898\n",
      "====> Epoch: 174 Valid:\tAvg loss: 0.3942\n",
      "\n",
      "====> Epoch: 175 Train:\tAvg loss: 0.3897\n",
      "====> Epoch: 175 Valid:\tAvg loss: 0.3920\n",
      "\n",
      "====> Epoch: 176 Train:\tAvg loss: 0.3898\n",
      "====> Epoch: 176 Valid:\tAvg loss: 0.3885\n",
      "\n",
      "====> Epoch: 177 Train:\tAvg loss: 0.3897\n",
      "====> Epoch: 177 Valid:\tAvg loss: 0.3901\n",
      "\n",
      "====> Epoch: 178 Train:\tAvg loss: 0.3898\n",
      "====> Epoch: 178 Valid:\tAvg loss: 0.3917\n",
      "\n",
      "====> Epoch: 179 Train:\tAvg loss: 0.3896\n",
      "====> Epoch: 179 Valid:\tAvg loss: 0.3916\n",
      "\n",
      "====> Epoch: 180 Train:\tAvg loss: 0.3895\n",
      "====> Epoch: 180 Valid:\tAvg loss: 0.3914\n",
      "\n",
      "====> Epoch: 181 Train:\tAvg loss: 0.3895\n",
      "====> Epoch: 181 Valid:\tAvg loss: 0.3944\n",
      "\n",
      "====> Epoch: 182 Train:\tAvg loss: 0.3893\n",
      "====> Epoch: 182 Valid:\tAvg loss: 0.3880\n",
      "\n",
      "====> Epoch: 183 Train:\tAvg loss: 0.3894\n",
      "====> Epoch: 183 Valid:\tAvg loss: 0.3955\n",
      "\n",
      "====> Epoch: 184 Train:\tAvg loss: 0.3894\n",
      "====> Epoch: 184 Valid:\tAvg loss: 0.3885\n",
      "\n",
      "====> Epoch: 185 Train:\tAvg loss: 0.3893\n",
      "====> Epoch: 185 Valid:\tAvg loss: 0.3914\n",
      "\n",
      "====> Epoch: 186 Train:\tAvg loss: 0.3894\n",
      "====> Epoch: 186 Valid:\tAvg loss: 0.3925\n",
      "\n",
      "====> Epoch: 187 Train:\tAvg loss: 0.3893\n",
      "====> Epoch: 187 Valid:\tAvg loss: 0.3974\n",
      "\n",
      "====> Epoch: 188 Train:\tAvg loss: 0.3894\n",
      "====> Epoch: 188 Valid:\tAvg loss: 0.3883\n",
      "\n",
      "====> Epoch: 189 Train:\tAvg loss: 0.3894\n",
      "====> Epoch: 189 Valid:\tAvg loss: 0.3883\n",
      "\n",
      "====> Epoch: 190 Train:\tAvg loss: 0.3892\n",
      "====> Epoch: 190 Valid:\tAvg loss: 0.3902\n",
      "\n",
      "====> Epoch: 191 Train:\tAvg loss: 0.3891\n",
      "====> Epoch: 191 Valid:\tAvg loss: 0.3901\n",
      "\n",
      "====> Epoch: 192 Train:\tAvg loss: 0.3891\n",
      "====> Epoch: 192 Valid:\tAvg loss: 0.3933\n",
      "\n",
      "====> Epoch: 193 Train:\tAvg loss: 0.3891\n",
      "====> Epoch: 193 Valid:\tAvg loss: 0.3986\n",
      "\n",
      "====> Epoch: 194 Train:\tAvg loss: 0.3891\n",
      "====> Epoch: 194 Valid:\tAvg loss: 0.3952\n",
      "\n",
      "====> Epoch: 195 Train:\tAvg loss: 0.3890\n",
      "====> Epoch: 195 Valid:\tAvg loss: 0.3958\n",
      "\n",
      "====> Epoch: 196 Train:\tAvg loss: 0.3891\n",
      "====> Epoch: 196 Valid:\tAvg loss: 0.3922\n",
      "\n",
      "====> Epoch: 197 Train:\tAvg loss: 0.3889\n",
      "====> Epoch: 197 Valid:\tAvg loss: 0.3981\n",
      "\n",
      "====> Epoch: 198 Train:\tAvg loss: 0.3891\n",
      "====> Epoch: 198 Valid:\tAvg loss: 0.3928\n",
      "\n",
      "====> Epoch: 199 Train:\tAvg loss: 0.3888\n",
      "====> Epoch: 199 Valid:\tAvg loss: 0.3932\n",
      "\n",
      "====> Epoch: 200 Train:\tAvg loss: 0.3888\n",
      "====> Epoch: 200 Valid:\tAvg loss: 0.3910\n",
      "\n",
      "====> Epoch: 201 Train:\tAvg loss: 0.3889\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====> Epoch: 201 Valid:\tAvg loss: 0.4011\n",
      "\n",
      "====> Epoch: 202 Train:\tAvg loss: 0.3888\n",
      "====> Epoch: 202 Valid:\tAvg loss: 0.3901\n",
      "\n",
      "====> Epoch: 203 Train:\tAvg loss: 0.3888\n",
      "====> Epoch: 203 Valid:\tAvg loss: 0.3950\n",
      "\n",
      "====> Epoch: 204 Train:\tAvg loss: 0.3887\n",
      "====> Epoch: 204 Valid:\tAvg loss: 0.3891\n",
      "\n",
      "====> Epoch: 205 Train:\tAvg loss: 0.3888\n",
      "====> Epoch: 205 Valid:\tAvg loss: 0.3910\n",
      "\n",
      "====> Epoch: 206 Train:\tAvg loss: 0.3887\n",
      "====> Epoch: 206 Valid:\tAvg loss: 0.3885\n",
      "\n",
      "====> Epoch: 207 Train:\tAvg loss: 0.3888\n",
      "====> Epoch: 207 Valid:\tAvg loss: 0.3921\n",
      "\n",
      "====> Epoch: 208 Train:\tAvg loss: 0.3886\n",
      "====> Epoch: 208 Valid:\tAvg loss: 0.3933\n",
      "\n",
      "====> Epoch: 209 Train:\tAvg loss: 0.3886\n",
      "====> Epoch: 209 Valid:\tAvg loss: 0.3920\n",
      "\n",
      "====> Epoch: 210 Train:\tAvg loss: 0.3886\n",
      "====> Epoch: 210 Valid:\tAvg loss: 0.3949\n",
      "\n",
      "====> Epoch: 211 Train:\tAvg loss: 0.3886\n",
      "====> Epoch: 211 Valid:\tAvg loss: 0.3887\n",
      "\n",
      "====> Epoch: 212 Train:\tAvg loss: 0.3886\n",
      "====> Epoch: 212 Valid:\tAvg loss: 0.3895\n",
      "\n",
      "====> Epoch: 213 Train:\tAvg loss: 0.3885\n",
      "====> Epoch: 213 Valid:\tAvg loss: 0.3920\n",
      "\n",
      "====> Epoch: 214 Train:\tAvg loss: 0.3884\n",
      "====> Epoch: 214 Valid:\tAvg loss: 0.3920\n",
      "\n",
      "====> Epoch: 215 Train:\tAvg loss: 0.3883\n",
      "====> Epoch: 215 Valid:\tAvg loss: 0.3923\n",
      "\n",
      "====> Epoch: 216 Train:\tAvg loss: 0.3885\n",
      "====> Epoch: 216 Valid:\tAvg loss: 0.3926\n",
      "\n",
      "====> Epoch: 217 Train:\tAvg loss: 0.3884\n",
      "====> Epoch: 217 Valid:\tAvg loss: 0.3919\n",
      "\n",
      "====> Epoch: 218 Train:\tAvg loss: 0.3884\n",
      "====> Epoch: 218 Valid:\tAvg loss: 0.3877\n",
      "\n",
      "====> Epoch: 219 Train:\tAvg loss: 0.3882\n",
      "====> Epoch: 219 Valid:\tAvg loss: 0.3939\n",
      "\n",
      "====> Epoch: 220 Train:\tAvg loss: 0.3883\n",
      "====> Epoch: 220 Valid:\tAvg loss: 0.3899\n",
      "\n",
      "====> Epoch: 221 Train:\tAvg loss: 0.3882\n",
      "====> Epoch: 221 Valid:\tAvg loss: 0.3918\n",
      "\n",
      "====> Epoch: 222 Train:\tAvg loss: 0.3882\n",
      "====> Epoch: 222 Valid:\tAvg loss: 0.3937\n",
      "\n",
      "====> Epoch: 223 Train:\tAvg loss: 0.3882\n",
      "====> Epoch: 223 Valid:\tAvg loss: 0.3950\n",
      "\n",
      "====> Epoch: 224 Train:\tAvg loss: 0.3882\n",
      "====> Epoch: 224 Valid:\tAvg loss: 0.3919\n",
      "\n",
      "====> Epoch: 225 Train:\tAvg loss: 0.3880\n",
      "====> Epoch: 225 Valid:\tAvg loss: 0.3909\n",
      "\n",
      "====> Epoch: 226 Train:\tAvg loss: 0.3882\n",
      "====> Epoch: 226 Valid:\tAvg loss: 0.3944\n",
      "\n",
      "====> Epoch: 227 Train:\tAvg loss: 0.3881\n",
      "====> Epoch: 227 Valid:\tAvg loss: 0.3900\n",
      "\n",
      "====> Epoch: 228 Train:\tAvg loss: 0.3880\n",
      "====> Epoch: 228 Valid:\tAvg loss: 0.3943\n",
      "\n",
      "====> Epoch: 229 Train:\tAvg loss: 0.3880\n",
      "====> Epoch: 229 Valid:\tAvg loss: 0.3877\n",
      "\n",
      "====> Epoch: 230 Train:\tAvg loss: 0.3879\n",
      "====> Epoch: 230 Valid:\tAvg loss: 0.3896\n",
      "\n",
      "====> Epoch: 231 Train:\tAvg loss: 0.3880\n",
      "====> Epoch: 231 Valid:\tAvg loss: 0.3920\n",
      "\n",
      "====> Epoch: 232 Train:\tAvg loss: 0.3879\n",
      "====> Epoch: 232 Valid:\tAvg loss: 0.3992\n",
      "\n",
      "====> Epoch: 233 Train:\tAvg loss: 0.3881\n",
      "====> Epoch: 233 Valid:\tAvg loss: 0.3933\n",
      "\n",
      "====> Epoch: 234 Train:\tAvg loss: 0.3878\n",
      "====> Epoch: 234 Valid:\tAvg loss: 0.3934\n",
      "\n",
      "====> Epoch: 235 Train:\tAvg loss: 0.3878\n",
      "====> Epoch: 235 Valid:\tAvg loss: 0.4033\n",
      "\n",
      "====> Epoch: 236 Train:\tAvg loss: 0.3878\n",
      "====> Epoch: 236 Valid:\tAvg loss: 0.3982\n",
      "\n",
      "====> Epoch: 237 Train:\tAvg loss: 0.3878\n",
      "====> Epoch: 237 Valid:\tAvg loss: 0.3933\n",
      "\n",
      "====> Epoch: 238 Train:\tAvg loss: 0.3877\n",
      "====> Epoch: 238 Valid:\tAvg loss: 0.3917\n",
      "\n",
      "====> Epoch: 239 Train:\tAvg loss: 0.3877\n",
      "====> Epoch: 239 Valid:\tAvg loss: 0.3908\n",
      "\n",
      "====> Epoch: 240 Train:\tAvg loss: 0.3877\n",
      "====> Epoch: 240 Valid:\tAvg loss: 0.3926\n",
      "\n",
      "====> Epoch: 241 Train:\tAvg loss: 0.3877\n",
      "====> Epoch: 241 Valid:\tAvg loss: 0.3935\n",
      "\n",
      "====> Epoch: 242 Train:\tAvg loss: 0.3877\n",
      "====> Epoch: 242 Valid:\tAvg loss: 0.3899\n",
      "\n",
      "====> Epoch: 243 Train:\tAvg loss: 0.3876\n",
      "====> Epoch: 243 Valid:\tAvg loss: 0.3891\n",
      "\n",
      "====> Epoch: 244 Train:\tAvg loss: 0.3875\n",
      "====> Epoch: 244 Valid:\tAvg loss: 0.3915\n",
      "\n",
      "====> Epoch: 245 Train:\tAvg loss: 0.3876\n",
      "====> Epoch: 245 Valid:\tAvg loss: 0.3919\n",
      "\n",
      "====> Epoch: 246 Train:\tAvg loss: 0.3876\n",
      "====> Epoch: 246 Valid:\tAvg loss: 0.3984\n",
      "\n",
      "====> Epoch: 247 Train:\tAvg loss: 0.3875\n",
      "====> Epoch: 247 Valid:\tAvg loss: 0.3905\n",
      "\n",
      "====> Epoch: 248 Train:\tAvg loss: 0.3875\n",
      "====> Epoch: 248 Valid:\tAvg loss: 0.3919\n",
      "\n",
      "====> Epoch: 249 Train:\tAvg loss: 0.3876\n",
      "====> Epoch: 249 Valid:\tAvg loss: 0.3914\n",
      "\n",
      "====> Epoch: 250 Train:\tAvg loss: 0.3875\n",
      "====> Epoch: 250 Valid:\tAvg loss: 0.3920\n",
      "\n",
      "====> Epoch: 251 Train:\tAvg loss: 0.3876\n",
      "====> Epoch: 251 Valid:\tAvg loss: 0.3901\n",
      "\n",
      "====> Epoch: 252 Train:\tAvg loss: 0.3874\n",
      "====> Epoch: 252 Valid:\tAvg loss: 0.3875\n",
      "\n",
      "====> Epoch: 253 Train:\tAvg loss: 0.3874\n",
      "====> Epoch: 253 Valid:\tAvg loss: 0.3993\n",
      "\n",
      "====> Epoch: 254 Train:\tAvg loss: 0.3873\n",
      "====> Epoch: 254 Valid:\tAvg loss: 0.3916\n",
      "\n",
      "====> Epoch: 255 Train:\tAvg loss: 0.3875\n",
      "====> Epoch: 255 Valid:\tAvg loss: 0.3889\n",
      "\n",
      "====> Epoch: 256 Train:\tAvg loss: 0.3875\n",
      "====> Epoch: 256 Valid:\tAvg loss: 0.3932\n",
      "\n",
      "====> Epoch: 257 Train:\tAvg loss: 0.3872\n",
      "====> Epoch: 257 Valid:\tAvg loss: 0.3900\n",
      "\n",
      "====> Epoch: 258 Train:\tAvg loss: 0.3872\n",
      "====> Epoch: 258 Valid:\tAvg loss: 0.3883\n",
      "\n",
      "====> Epoch: 259 Train:\tAvg loss: 0.3873\n",
      "====> Epoch: 259 Valid:\tAvg loss: 0.3931\n",
      "\n",
      "====> Epoch: 260 Train:\tAvg loss: 0.3872\n",
      "====> Epoch: 260 Valid:\tAvg loss: 0.3904\n",
      "\n",
      "====> Epoch: 261 Train:\tAvg loss: 0.3872\n",
      "====> Epoch: 261 Valid:\tAvg loss: 0.3903\n",
      "\n",
      "====> Epoch: 262 Train:\tAvg loss: 0.3873\n",
      "====> Epoch: 262 Valid:\tAvg loss: 0.3902\n",
      "\n",
      "====> Epoch: 263 Train:\tAvg loss: 0.3872\n",
      "====> Epoch: 263 Valid:\tAvg loss: 0.3942\n",
      "\n",
      "====> Epoch: 264 Train:\tAvg loss: 0.3872\n",
      "====> Epoch: 264 Valid:\tAvg loss: 0.3881\n",
      "\n",
      "====> Epoch: 265 Train:\tAvg loss: 0.3871\n",
      "====> Epoch: 265 Valid:\tAvg loss: 0.3886\n",
      "\n",
      "====> Epoch: 266 Train:\tAvg loss: 0.3872\n",
      "====> Epoch: 266 Valid:\tAvg loss: 0.3900\n",
      "\n",
      "====> Epoch: 267 Train:\tAvg loss: 0.3871\n",
      "====> Epoch: 267 Valid:\tAvg loss: 0.3894\n",
      "\n",
      "====> Epoch: 268 Train:\tAvg loss: 0.3872\n",
      "====> Epoch: 268 Valid:\tAvg loss: 0.3899\n",
      "\n",
      "====> Epoch: 269 Train:\tAvg loss: 0.3871\n",
      "====> Epoch: 269 Valid:\tAvg loss: 0.3946\n",
      "\n",
      "====> Epoch: 270 Train:\tAvg loss: 0.3872\n",
      "====> Epoch: 270 Valid:\tAvg loss: 0.3885\n",
      "\n",
      "====> Epoch: 271 Train:\tAvg loss: 0.3871\n",
      "====> Epoch: 271 Valid:\tAvg loss: 0.3888\n",
      "\n",
      "====> Epoch: 272 Train:\tAvg loss: 0.3871\n",
      "====> Epoch: 272 Valid:\tAvg loss: 0.3926\n",
      "\n",
      "====> Epoch: 273 Train:\tAvg loss: 0.3871\n",
      "====> Epoch: 273 Valid:\tAvg loss: 0.3917\n",
      "\n",
      "====> Epoch: 274 Train:\tAvg loss: 0.3870\n",
      "====> Epoch: 274 Valid:\tAvg loss: 0.3880\n",
      "\n",
      "====> Epoch: 275 Train:\tAvg loss: 0.3870\n",
      "====> Epoch: 275 Valid:\tAvg loss: 0.3930\n",
      "\n",
      "====> Epoch: 276 Train:\tAvg loss: 0.3871\n",
      "====> Epoch: 276 Valid:\tAvg loss: 0.3883\n",
      "\n",
      "====> Epoch: 277 Train:\tAvg loss: 0.3870\n",
      "====> Epoch: 277 Valid:\tAvg loss: 0.3909\n",
      "\n",
      "====> Epoch: 278 Train:\tAvg loss: 0.3869\n",
      "====> Epoch: 278 Valid:\tAvg loss: 0.3910\n",
      "\n",
      "====> Epoch: 279 Train:\tAvg loss: 0.3869\n",
      "====> Epoch: 279 Valid:\tAvg loss: 0.3901\n",
      "\n",
      "====> Epoch: 280 Train:\tAvg loss: 0.3870\n",
      "====> Epoch: 280 Valid:\tAvg loss: 0.3875\n",
      "\n",
      "====> Epoch: 281 Train:\tAvg loss: 0.3868\n",
      "====> Epoch: 281 Valid:\tAvg loss: 0.3939\n",
      "\n",
      "====> Epoch: 282 Train:\tAvg loss: 0.3868\n",
      "====> Epoch: 282 Valid:\tAvg loss: 0.3937\n",
      "\n",
      "====> Epoch: 283 Train:\tAvg loss: 0.3868\n",
      "====> Epoch: 283 Valid:\tAvg loss: 0.3925\n",
      "\n",
      "====> Epoch: 284 Train:\tAvg loss: 0.3869\n",
      "====> Epoch: 284 Valid:\tAvg loss: 0.3878\n",
      "\n",
      "====> Epoch: 285 Train:\tAvg loss: 0.3868\n",
      "====> Epoch: 285 Valid:\tAvg loss: 0.3896\n",
      "\n",
      "====> Epoch: 286 Train:\tAvg loss: 0.3867\n",
      "====> Epoch: 286 Valid:\tAvg loss: 0.3864\n",
      "\n",
      "====> Epoch: 287 Train:\tAvg loss: 0.3867\n",
      "====> Epoch: 287 Valid:\tAvg loss: 0.3939\n",
      "\n",
      "====> Epoch: 288 Train:\tAvg loss: 0.3869\n",
      "====> Epoch: 288 Valid:\tAvg loss: 0.3893\n",
      "\n",
      "====> Epoch: 289 Train:\tAvg loss: 0.3867\n",
      "====> Epoch: 289 Valid:\tAvg loss: 0.3874\n",
      "\n",
      "====> Epoch: 290 Train:\tAvg loss: 0.3868\n",
      "====> Epoch: 290 Valid:\tAvg loss: 0.3890\n",
      "\n",
      "====> Epoch: 291 Train:\tAvg loss: 0.3868\n",
      "====> Epoch: 291 Valid:\tAvg loss: 0.3962\n",
      "\n",
      "====> Epoch: 292 Train:\tAvg loss: 0.3867\n",
      "====> Epoch: 292 Valid:\tAvg loss: 0.3982\n",
      "\n",
      "====> Epoch: 293 Train:\tAvg loss: 0.3866\n",
      "====> Epoch: 293 Valid:\tAvg loss: 0.3891\n",
      "\n",
      "====> Epoch: 294 Train:\tAvg loss: 0.3866\n",
      "====> Epoch: 294 Valid:\tAvg loss: 0.3885\n",
      "\n",
      "====> Epoch: 295 Train:\tAvg loss: 0.3866\n",
      "====> Epoch: 295 Valid:\tAvg loss: 0.3895\n",
      "\n",
      "====> Epoch: 296 Train:\tAvg loss: 0.3864\n",
      "====> Epoch: 296 Valid:\tAvg loss: 0.3897\n",
      "\n",
      "====> Epoch: 297 Train:\tAvg loss: 0.3866\n",
      "====> Epoch: 297 Valid:\tAvg loss: 0.3929\n",
      "\n",
      "====> Epoch: 298 Train:\tAvg loss: 0.3866\n",
      "====> Epoch: 298 Valid:\tAvg loss: 0.3903\n",
      "\n",
      "====> Epoch: 299 Train:\tAvg loss: 0.3865\n",
      "====> Epoch: 299 Valid:\tAvg loss: 0.3865\n",
      "\n",
      "====> Epoch: 300 Train:\tAvg loss: 0.3865\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====> Epoch: 300 Valid:\tAvg loss: 0.3909\n",
      "\n",
      "====> Epoch: 301 Train:\tAvg loss: 0.3866\n",
      "====> Epoch: 301 Valid:\tAvg loss: 0.3885\n",
      "\n",
      "====> Epoch: 302 Train:\tAvg loss: 0.3864\n",
      "====> Epoch: 302 Valid:\tAvg loss: 0.3893\n",
      "\n",
      "====> Epoch: 303 Train:\tAvg loss: 0.3864\n",
      "====> Epoch: 303 Valid:\tAvg loss: 0.3938\n",
      "\n",
      "====> Epoch: 304 Train:\tAvg loss: 0.3864\n",
      "====> Epoch: 304 Valid:\tAvg loss: 0.3903\n",
      "\n",
      "====> Epoch: 305 Train:\tAvg loss: 0.3864\n",
      "====> Epoch: 305 Valid:\tAvg loss: 0.3859\n",
      "\n",
      "====> Epoch: 306 Train:\tAvg loss: 0.3864\n",
      "====> Epoch: 306 Valid:\tAvg loss: 0.3930\n",
      "\n",
      "====> Epoch: 307 Train:\tAvg loss: 0.3864\n",
      "====> Epoch: 307 Valid:\tAvg loss: 0.3896\n",
      "\n",
      "====> Epoch: 308 Train:\tAvg loss: 0.3864\n",
      "====> Epoch: 308 Valid:\tAvg loss: 0.3942\n",
      "\n",
      "====> Epoch: 309 Train:\tAvg loss: 0.3862\n",
      "====> Epoch: 309 Valid:\tAvg loss: 0.3900\n",
      "\n",
      "====> Epoch: 310 Train:\tAvg loss: 0.3863\n",
      "====> Epoch: 310 Valid:\tAvg loss: 0.3945\n",
      "\n",
      "====> Epoch: 311 Train:\tAvg loss: 0.3862\n",
      "====> Epoch: 311 Valid:\tAvg loss: 0.3892\n",
      "\n",
      "====> Epoch: 312 Train:\tAvg loss: 0.3864\n",
      "====> Epoch: 312 Valid:\tAvg loss: 0.4000\n",
      "\n",
      "====> Epoch: 313 Train:\tAvg loss: 0.3863\n",
      "====> Epoch: 313 Valid:\tAvg loss: 0.3894\n",
      "\n",
      "====> Epoch: 314 Train:\tAvg loss: 0.3862\n",
      "====> Epoch: 314 Valid:\tAvg loss: 0.3954\n",
      "\n",
      "====> Epoch: 315 Train:\tAvg loss: 0.3863\n",
      "====> Epoch: 315 Valid:\tAvg loss: 0.3872\n",
      "\n",
      "====> Epoch: 316 Train:\tAvg loss: 0.3862\n",
      "====> Epoch: 316 Valid:\tAvg loss: 0.3884\n",
      "\n",
      "====> Epoch: 317 Train:\tAvg loss: 0.3862\n",
      "====> Epoch: 317 Valid:\tAvg loss: 0.3900\n",
      "\n",
      "====> Epoch: 318 Train:\tAvg loss: 0.3863\n",
      "====> Epoch: 318 Valid:\tAvg loss: 0.3960\n",
      "\n",
      "====> Epoch: 319 Train:\tAvg loss: 0.3862\n",
      "====> Epoch: 319 Valid:\tAvg loss: 0.3932\n",
      "\n",
      "====> Epoch: 320 Train:\tAvg loss: 0.3861\n",
      "====> Epoch: 320 Valid:\tAvg loss: 0.3912\n",
      "\n",
      "====> Epoch: 321 Train:\tAvg loss: 0.3861\n",
      "====> Epoch: 321 Valid:\tAvg loss: 0.3880\n",
      "\n",
      "====> Epoch: 322 Train:\tAvg loss: 0.3861\n",
      "====> Epoch: 322 Valid:\tAvg loss: 0.3874\n",
      "\n",
      "====> Epoch: 323 Train:\tAvg loss: 0.3862\n",
      "====> Epoch: 323 Valid:\tAvg loss: 0.3879\n",
      "\n",
      "====> Epoch: 324 Train:\tAvg loss: 0.3862\n",
      "====> Epoch: 324 Valid:\tAvg loss: 0.3898\n",
      "\n",
      "====> Epoch: 325 Train:\tAvg loss: 0.3861\n",
      "====> Epoch: 325 Valid:\tAvg loss: 0.3921\n",
      "\n",
      "====> Epoch: 326 Train:\tAvg loss: 0.3862\n",
      "====> Epoch: 326 Valid:\tAvg loss: 0.3926\n",
      "\n",
      "====> Epoch: 327 Train:\tAvg loss: 0.3860\n",
      "====> Epoch: 327 Valid:\tAvg loss: 0.3884\n",
      "\n",
      "====> Epoch: 328 Train:\tAvg loss: 0.3860\n",
      "====> Epoch: 328 Valid:\tAvg loss: 0.3945\n",
      "\n",
      "====> Epoch: 329 Train:\tAvg loss: 0.3860\n",
      "====> Epoch: 329 Valid:\tAvg loss: 0.3903\n",
      "\n",
      "====> Epoch: 330 Train:\tAvg loss: 0.3860\n",
      "====> Epoch: 330 Valid:\tAvg loss: 0.3922\n",
      "\n",
      "====> Epoch: 331 Train:\tAvg loss: 0.3861\n",
      "====> Epoch: 331 Valid:\tAvg loss: 0.3900\n",
      "\n",
      "====> Epoch: 332 Train:\tAvg loss: 0.3861\n",
      "====> Epoch: 332 Valid:\tAvg loss: 0.3899\n",
      "\n",
      "====> Epoch: 333 Train:\tAvg loss: 0.3859\n",
      "====> Epoch: 333 Valid:\tAvg loss: 0.3906\n",
      "\n",
      "====> Epoch: 334 Train:\tAvg loss: 0.3860\n",
      "====> Epoch: 334 Valid:\tAvg loss: 0.3882\n",
      "\n",
      "====> Epoch: 335 Train:\tAvg loss: 0.3859\n",
      "====> Epoch: 335 Valid:\tAvg loss: 0.3854\n",
      "\n",
      "====> Epoch: 336 Train:\tAvg loss: 0.3860\n",
      "====> Epoch: 336 Valid:\tAvg loss: 0.3966\n",
      "\n",
      "====> Epoch: 337 Train:\tAvg loss: 0.3859\n",
      "====> Epoch: 337 Valid:\tAvg loss: 0.3882\n",
      "\n",
      "====> Epoch: 338 Train:\tAvg loss: 0.3859\n",
      "====> Epoch: 338 Valid:\tAvg loss: 0.3879\n",
      "\n",
      "====> Epoch: 339 Train:\tAvg loss: 0.3859\n",
      "====> Epoch: 339 Valid:\tAvg loss: 0.3898\n",
      "\n",
      "====> Epoch: 340 Train:\tAvg loss: 0.3859\n",
      "====> Epoch: 340 Valid:\tAvg loss: 0.3891\n",
      "\n",
      "====> Epoch: 341 Train:\tAvg loss: 0.3858\n",
      "====> Epoch: 341 Valid:\tAvg loss: 0.3892\n",
      "\n",
      "====> Epoch: 342 Train:\tAvg loss: 0.3857\n",
      "====> Epoch: 342 Valid:\tAvg loss: 0.3924\n",
      "\n",
      "====> Epoch: 343 Train:\tAvg loss: 0.3858\n",
      "====> Epoch: 343 Valid:\tAvg loss: 0.3897\n",
      "\n",
      "====> Epoch: 344 Train:\tAvg loss: 0.3858\n",
      "====> Epoch: 344 Valid:\tAvg loss: 0.3882\n",
      "\n",
      "====> Epoch: 345 Train:\tAvg loss: 0.3859\n",
      "====> Epoch: 345 Valid:\tAvg loss: 0.3881\n",
      "\n",
      "====> Epoch: 346 Train:\tAvg loss: 0.3858\n",
      "====> Epoch: 346 Valid:\tAvg loss: 0.3898\n",
      "\n",
      "====> Epoch: 347 Train:\tAvg loss: 0.3858\n",
      "====> Epoch: 347 Valid:\tAvg loss: 0.3900\n",
      "\n",
      "====> Epoch: 348 Train:\tAvg loss: 0.3856\n",
      "====> Epoch: 348 Valid:\tAvg loss: 0.3950\n",
      "\n",
      "====> Epoch: 349 Train:\tAvg loss: 0.3858\n",
      "====> Epoch: 349 Valid:\tAvg loss: 0.3931\n",
      "\n",
      "====> Epoch: 350 Train:\tAvg loss: 0.3857\n",
      "====> Epoch: 350 Valid:\tAvg loss: 0.3922\n",
      "\n",
      "====> Epoch: 351 Train:\tAvg loss: 0.3855\n",
      "====> Epoch: 351 Valid:\tAvg loss: 0.3857\n",
      "\n",
      "====> Epoch: 352 Train:\tAvg loss: 0.3857\n",
      "====> Epoch: 352 Valid:\tAvg loss: 0.3872\n",
      "\n",
      "====> Epoch: 353 Train:\tAvg loss: 0.3858\n",
      "====> Epoch: 353 Valid:\tAvg loss: 0.3879\n",
      "\n",
      "====> Epoch: 354 Train:\tAvg loss: 0.3855\n",
      "====> Epoch: 354 Valid:\tAvg loss: 0.3903\n",
      "\n",
      "====> Epoch: 355 Train:\tAvg loss: 0.3857\n",
      "====> Epoch: 355 Valid:\tAvg loss: 0.3886\n",
      "\n",
      "====> Epoch: 356 Train:\tAvg loss: 0.3855\n",
      "====> Epoch: 356 Valid:\tAvg loss: 0.3878\n",
      "\n",
      "====> Epoch: 357 Train:\tAvg loss: 0.3857\n",
      "====> Epoch: 357 Valid:\tAvg loss: 0.3888\n",
      "\n",
      "====> Epoch: 358 Train:\tAvg loss: 0.3856\n",
      "====> Epoch: 358 Valid:\tAvg loss: 0.3875\n",
      "\n",
      "====> Epoch: 359 Train:\tAvg loss: 0.3855\n",
      "====> Epoch: 359 Valid:\tAvg loss: 0.3880\n",
      "\n",
      "====> Epoch: 360 Train:\tAvg loss: 0.3856\n",
      "====> Epoch: 360 Valid:\tAvg loss: 0.3872\n",
      "\n",
      "====> Epoch: 361 Train:\tAvg loss: 0.3855\n",
      "====> Epoch: 361 Valid:\tAvg loss: 0.3866\n",
      "\n",
      "====> Epoch: 362 Train:\tAvg loss: 0.3856\n",
      "====> Epoch: 362 Valid:\tAvg loss: 0.3892\n",
      "\n",
      "====> Epoch: 363 Train:\tAvg loss: 0.3857\n",
      "====> Epoch: 363 Valid:\tAvg loss: 0.3892\n",
      "\n",
      "====> Epoch: 364 Train:\tAvg loss: 0.3856\n",
      "====> Epoch: 364 Valid:\tAvg loss: 0.3858\n",
      "\n",
      "====> Epoch: 365 Train:\tAvg loss: 0.3855\n",
      "====> Epoch: 365 Valid:\tAvg loss: 0.3894\n",
      "\n",
      "====> Epoch: 366 Train:\tAvg loss: 0.3856\n",
      "====> Epoch: 366 Valid:\tAvg loss: 0.3898\n",
      "\n",
      "====> Epoch: 367 Train:\tAvg loss: 0.3855\n",
      "====> Epoch: 367 Valid:\tAvg loss: 0.3926\n",
      "\n",
      "====> Epoch: 368 Train:\tAvg loss: 0.3855\n",
      "====> Epoch: 368 Valid:\tAvg loss: 0.3904\n",
      "\n",
      "====> Epoch: 369 Train:\tAvg loss: 0.3855\n",
      "====> Epoch: 369 Valid:\tAvg loss: 0.3916\n",
      "\n",
      "====> Epoch: 370 Train:\tAvg loss: 0.3854\n",
      "====> Epoch: 370 Valid:\tAvg loss: 0.3876\n",
      "\n",
      "====> Epoch: 371 Train:\tAvg loss: 0.3856\n",
      "====> Epoch: 371 Valid:\tAvg loss: 0.3883\n",
      "\n",
      "====> Epoch: 372 Train:\tAvg loss: 0.3854\n",
      "====> Epoch: 372 Valid:\tAvg loss: 0.3913\n",
      "\n",
      "====> Epoch: 373 Train:\tAvg loss: 0.3854\n",
      "====> Epoch: 373 Valid:\tAvg loss: 0.3858\n",
      "\n",
      "====> Epoch: 374 Train:\tAvg loss: 0.3854\n",
      "====> Epoch: 374 Valid:\tAvg loss: 0.3895\n",
      "\n",
      "====> Epoch: 375 Train:\tAvg loss: 0.3854\n",
      "====> Epoch: 375 Valid:\tAvg loss: 0.3928\n",
      "\n",
      "====> Epoch: 376 Train:\tAvg loss: 0.3855\n",
      "====> Epoch: 376 Valid:\tAvg loss: 0.3876\n",
      "\n",
      "====> Epoch: 377 Train:\tAvg loss: 0.3853\n",
      "====> Epoch: 377 Valid:\tAvg loss: 0.3898\n",
      "\n",
      "====> Epoch: 378 Train:\tAvg loss: 0.3853\n",
      "====> Epoch: 378 Valid:\tAvg loss: 0.3869\n",
      "\n",
      "====> Epoch: 379 Train:\tAvg loss: 0.3853\n",
      "====> Epoch: 379 Valid:\tAvg loss: 0.3974\n",
      "\n",
      "====> Epoch: 380 Train:\tAvg loss: 0.3854\n",
      "====> Epoch: 380 Valid:\tAvg loss: 0.3862\n",
      "\n",
      "====> Epoch: 381 Train:\tAvg loss: 0.3852\n",
      "====> Epoch: 381 Valid:\tAvg loss: 0.3916\n",
      "\n",
      "====> Epoch: 382 Train:\tAvg loss: 0.3854\n",
      "====> Epoch: 382 Valid:\tAvg loss: 0.3879\n",
      "\n",
      "====> Epoch: 383 Train:\tAvg loss: 0.3852\n",
      "====> Epoch: 383 Valid:\tAvg loss: 0.3889\n",
      "\n",
      "====> Epoch: 384 Train:\tAvg loss: 0.3853\n",
      "====> Epoch: 384 Valid:\tAvg loss: 0.3943\n",
      "\n",
      "====> Epoch: 385 Train:\tAvg loss: 0.3853\n",
      "====> Epoch: 385 Valid:\tAvg loss: 0.3883\n",
      "\n",
      "====> Epoch: 386 Train:\tAvg loss: 0.3853\n",
      "====> Epoch: 386 Valid:\tAvg loss: 0.3883\n",
      "\n",
      "====> Epoch: 387 Train:\tAvg loss: 0.3854\n",
      "====> Epoch: 387 Valid:\tAvg loss: 0.3933\n",
      "\n",
      "====> Epoch: 388 Train:\tAvg loss: 0.3851\n",
      "====> Epoch: 388 Valid:\tAvg loss: 0.3873\n",
      "\n",
      "====> Epoch: 389 Train:\tAvg loss: 0.3852\n",
      "====> Epoch: 389 Valid:\tAvg loss: 0.3914\n",
      "\n",
      "====> Epoch: 390 Train:\tAvg loss: 0.3852\n",
      "====> Epoch: 390 Valid:\tAvg loss: 0.3905\n",
      "\n",
      "====> Epoch: 391 Train:\tAvg loss: 0.3851\n",
      "====> Epoch: 391 Valid:\tAvg loss: 0.3911\n",
      "\n",
      "====> Epoch: 392 Train:\tAvg loss: 0.3852\n",
      "====> Epoch: 392 Valid:\tAvg loss: 0.3922\n",
      "\n",
      "====> Epoch: 393 Train:\tAvg loss: 0.3852\n",
      "====> Epoch: 393 Valid:\tAvg loss: 0.3909\n",
      "\n",
      "====> Epoch: 394 Train:\tAvg loss: 0.3851\n",
      "====> Epoch: 394 Valid:\tAvg loss: 0.3913\n",
      "\n",
      "====> Epoch: 395 Train:\tAvg loss: 0.3851\n",
      "====> Epoch: 395 Valid:\tAvg loss: 0.3867\n",
      "\n",
      "====> Epoch: 396 Train:\tAvg loss: 0.3851\n",
      "====> Epoch: 396 Valid:\tAvg loss: 0.3860\n",
      "\n",
      "====> Epoch: 397 Train:\tAvg loss: 0.3850\n",
      "====> Epoch: 397 Valid:\tAvg loss: 0.3884\n",
      "\n",
      "====> Epoch: 398 Train:\tAvg loss: 0.3851\n",
      "====> Epoch: 398 Valid:\tAvg loss: 0.3894\n",
      "\n",
      "====> Epoch: 399 Train:\tAvg loss: 0.3851\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====> Epoch: 399 Valid:\tAvg loss: 0.3932\n",
      "\n",
      "====> Epoch: 400 Train:\tAvg loss: 0.3851\n",
      "====> Epoch: 400 Valid:\tAvg loss: 0.3899\n",
      "\n",
      "====> Epoch: 401 Train:\tAvg loss: 0.3850\n",
      "====> Epoch: 401 Valid:\tAvg loss: 0.3883\n",
      "\n",
      "====> Epoch: 402 Train:\tAvg loss: 0.3848\n",
      "====> Epoch: 402 Valid:\tAvg loss: 0.3877\n",
      "\n",
      "====> Epoch: 403 Train:\tAvg loss: 0.3851\n",
      "====> Epoch: 403 Valid:\tAvg loss: 0.3873\n",
      "\n",
      "====> Epoch: 404 Train:\tAvg loss: 0.3849\n",
      "====> Epoch: 404 Valid:\tAvg loss: 0.3879\n",
      "\n",
      "====> Epoch: 405 Train:\tAvg loss: 0.3850\n",
      "====> Epoch: 405 Valid:\tAvg loss: 0.3892\n",
      "\n",
      "====> Epoch: 406 Train:\tAvg loss: 0.3850\n",
      "====> Epoch: 406 Valid:\tAvg loss: 0.3894\n",
      "\n",
      "====> Epoch: 407 Train:\tAvg loss: 0.3850\n",
      "====> Epoch: 407 Valid:\tAvg loss: 0.3913\n",
      "\n",
      "====> Epoch: 408 Train:\tAvg loss: 0.3848\n",
      "====> Epoch: 408 Valid:\tAvg loss: 0.3911\n",
      "\n",
      "====> Epoch: 409 Train:\tAvg loss: 0.3849\n",
      "====> Epoch: 409 Valid:\tAvg loss: 0.3879\n",
      "\n",
      "====> Epoch: 410 Train:\tAvg loss: 0.3848\n",
      "====> Epoch: 410 Valid:\tAvg loss: 0.3944\n",
      "\n",
      "====> Epoch: 411 Train:\tAvg loss: 0.3848\n",
      "====> Epoch: 411 Valid:\tAvg loss: 0.3909\n",
      "\n",
      "====> Epoch: 412 Train:\tAvg loss: 0.3849\n",
      "====> Epoch: 412 Valid:\tAvg loss: 0.3884\n",
      "\n",
      "====> Epoch: 413 Train:\tAvg loss: 0.3849\n",
      "====> Epoch: 413 Valid:\tAvg loss: 0.3850\n",
      "\n",
      "====> Epoch: 414 Train:\tAvg loss: 0.3847\n",
      "====> Epoch: 414 Valid:\tAvg loss: 0.3866\n",
      "\n",
      "====> Epoch: 415 Train:\tAvg loss: 0.3848\n",
      "====> Epoch: 415 Valid:\tAvg loss: 0.3895\n",
      "\n",
      "====> Epoch: 416 Train:\tAvg loss: 0.3847\n",
      "====> Epoch: 416 Valid:\tAvg loss: 0.3917\n",
      "\n",
      "====> Epoch: 417 Train:\tAvg loss: 0.3848\n",
      "====> Epoch: 417 Valid:\tAvg loss: 0.3875\n",
      "\n",
      "====> Epoch: 418 Train:\tAvg loss: 0.3847\n",
      "====> Epoch: 418 Valid:\tAvg loss: 0.3841\n",
      "\n",
      "====> Epoch: 419 Train:\tAvg loss: 0.3847\n",
      "====> Epoch: 419 Valid:\tAvg loss: 0.3926\n",
      "\n",
      "====> Epoch: 420 Train:\tAvg loss: 0.3846\n",
      "====> Epoch: 420 Valid:\tAvg loss: 0.3919\n",
      "\n",
      "====> Epoch: 421 Train:\tAvg loss: 0.3846\n",
      "====> Epoch: 421 Valid:\tAvg loss: 0.3902\n",
      "\n",
      "====> Epoch: 422 Train:\tAvg loss: 0.3847\n",
      "====> Epoch: 422 Valid:\tAvg loss: 0.3865\n",
      "\n",
      "====> Epoch: 423 Train:\tAvg loss: 0.3847\n",
      "====> Epoch: 423 Valid:\tAvg loss: 0.3869\n",
      "\n",
      "====> Epoch: 424 Train:\tAvg loss: 0.3845\n",
      "====> Epoch: 424 Valid:\tAvg loss: 0.3842\n",
      "\n",
      "====> Epoch: 425 Train:\tAvg loss: 0.3847\n",
      "====> Epoch: 425 Valid:\tAvg loss: 0.3863\n",
      "\n",
      "====> Epoch: 426 Train:\tAvg loss: 0.3845\n",
      "====> Epoch: 426 Valid:\tAvg loss: 0.3907\n",
      "\n",
      "====> Epoch: 427 Train:\tAvg loss: 0.3846\n",
      "====> Epoch: 427 Valid:\tAvg loss: 0.3903\n",
      "\n",
      "====> Epoch: 428 Train:\tAvg loss: 0.3846\n",
      "====> Epoch: 428 Valid:\tAvg loss: 0.3894\n",
      "\n",
      "====> Epoch: 429 Train:\tAvg loss: 0.3846\n",
      "====> Epoch: 429 Valid:\tAvg loss: 0.3872\n",
      "\n",
      "====> Epoch: 430 Train:\tAvg loss: 0.3846\n",
      "====> Epoch: 430 Valid:\tAvg loss: 0.3871\n",
      "\n",
      "====> Epoch: 431 Train:\tAvg loss: 0.3845\n",
      "====> Epoch: 431 Valid:\tAvg loss: 0.3903\n",
      "\n",
      "====> Epoch: 432 Train:\tAvg loss: 0.3846\n",
      "====> Epoch: 432 Valid:\tAvg loss: 0.3922\n",
      "\n",
      "====> Epoch: 433 Train:\tAvg loss: 0.3845\n",
      "====> Epoch: 433 Valid:\tAvg loss: 0.3885\n",
      "\n",
      "====> Epoch: 434 Train:\tAvg loss: 0.3844\n",
      "====> Epoch: 434 Valid:\tAvg loss: 0.3872\n",
      "\n",
      "====> Epoch: 435 Train:\tAvg loss: 0.3846\n",
      "====> Epoch: 435 Valid:\tAvg loss: 0.3852\n",
      "\n",
      "====> Epoch: 436 Train:\tAvg loss: 0.3846\n",
      "====> Epoch: 436 Valid:\tAvg loss: 0.3903\n",
      "\n",
      "====> Epoch: 437 Train:\tAvg loss: 0.3845\n",
      "====> Epoch: 437 Valid:\tAvg loss: 0.3916\n",
      "\n",
      "====> Epoch: 438 Train:\tAvg loss: 0.3844\n",
      "====> Epoch: 438 Valid:\tAvg loss: 0.3873\n",
      "\n",
      "====> Epoch: 439 Train:\tAvg loss: 0.3845\n",
      "====> Epoch: 439 Valid:\tAvg loss: 0.3894\n",
      "\n",
      "====> Epoch: 440 Train:\tAvg loss: 0.3844\n",
      "====> Epoch: 440 Valid:\tAvg loss: 0.3862\n",
      "\n",
      "====> Epoch: 441 Train:\tAvg loss: 0.3845\n",
      "====> Epoch: 441 Valid:\tAvg loss: 0.3890\n",
      "\n",
      "====> Epoch: 442 Train:\tAvg loss: 0.3845\n",
      "====> Epoch: 442 Valid:\tAvg loss: 0.3944\n",
      "\n",
      "====> Epoch: 443 Train:\tAvg loss: 0.3844\n",
      "====> Epoch: 443 Valid:\tAvg loss: 0.3876\n",
      "\n",
      "====> Epoch: 444 Train:\tAvg loss: 0.3844\n",
      "====> Epoch: 444 Valid:\tAvg loss: 0.3909\n",
      "\n",
      "====> Epoch: 445 Train:\tAvg loss: 0.3845\n",
      "====> Epoch: 445 Valid:\tAvg loss: 0.3856\n",
      "\n",
      "====> Epoch: 446 Train:\tAvg loss: 0.3844\n",
      "====> Epoch: 446 Valid:\tAvg loss: 0.3911\n",
      "\n",
      "====> Epoch: 447 Train:\tAvg loss: 0.3843\n",
      "====> Epoch: 447 Valid:\tAvg loss: 0.3874\n",
      "\n",
      "====> Epoch: 448 Train:\tAvg loss: 0.3845\n",
      "====> Epoch: 448 Valid:\tAvg loss: 0.3909\n",
      "\n",
      "====> Epoch: 449 Train:\tAvg loss: 0.3842\n",
      "====> Epoch: 449 Valid:\tAvg loss: 0.3937\n",
      "\n",
      "====> Epoch: 450 Train:\tAvg loss: 0.3843\n",
      "====> Epoch: 450 Valid:\tAvg loss: 0.3869\n",
      "\n",
      "====> Epoch: 451 Train:\tAvg loss: 0.3843\n",
      "====> Epoch: 451 Valid:\tAvg loss: 0.3904\n",
      "\n",
      "====> Epoch: 452 Train:\tAvg loss: 0.3844\n",
      "====> Epoch: 452 Valid:\tAvg loss: 0.3858\n",
      "\n",
      "====> Epoch: 453 Train:\tAvg loss: 0.3841\n",
      "====> Epoch: 453 Valid:\tAvg loss: 0.3864\n",
      "\n",
      "====> Epoch: 454 Train:\tAvg loss: 0.3843\n",
      "====> Epoch: 454 Valid:\tAvg loss: 0.3916\n",
      "\n",
      "====> Epoch: 455 Train:\tAvg loss: 0.3842\n",
      "====> Epoch: 455 Valid:\tAvg loss: 0.3912\n",
      "\n",
      "====> Epoch: 456 Train:\tAvg loss: 0.3843\n",
      "====> Epoch: 456 Valid:\tAvg loss: 0.3868\n",
      "\n",
      "====> Epoch: 457 Train:\tAvg loss: 0.3842\n",
      "====> Epoch: 457 Valid:\tAvg loss: 0.3897\n",
      "\n",
      "====> Epoch: 458 Train:\tAvg loss: 0.3842\n",
      "====> Epoch: 458 Valid:\tAvg loss: 0.3884\n",
      "\n",
      "====> Epoch: 459 Train:\tAvg loss: 0.3843\n",
      "====> Epoch: 459 Valid:\tAvg loss: 0.3873\n",
      "\n",
      "====> Epoch: 460 Train:\tAvg loss: 0.3842\n",
      "====> Epoch: 460 Valid:\tAvg loss: 0.3873\n",
      "\n",
      "====> Epoch: 461 Train:\tAvg loss: 0.3842\n",
      "====> Epoch: 461 Valid:\tAvg loss: 0.3924\n",
      "\n",
      "====> Epoch: 462 Train:\tAvg loss: 0.3842\n",
      "====> Epoch: 462 Valid:\tAvg loss: 0.3936\n",
      "\n",
      "====> Epoch: 463 Train:\tAvg loss: 0.3842\n",
      "====> Epoch: 463 Valid:\tAvg loss: 0.3871\n",
      "\n",
      "====> Epoch: 464 Train:\tAvg loss: 0.3842\n",
      "====> Epoch: 464 Valid:\tAvg loss: 0.3865\n",
      "\n",
      "====> Epoch: 465 Train:\tAvg loss: 0.3842\n",
      "====> Epoch: 465 Valid:\tAvg loss: 0.3873\n",
      "\n",
      "====> Epoch: 466 Train:\tAvg loss: 0.3840\n",
      "====> Epoch: 466 Valid:\tAvg loss: 0.3895\n",
      "\n",
      "====> Epoch: 467 Train:\tAvg loss: 0.3841\n",
      "====> Epoch: 467 Valid:\tAvg loss: 0.3882\n",
      "\n",
      "====> Epoch: 468 Train:\tAvg loss: 0.3841\n",
      "====> Epoch: 468 Valid:\tAvg loss: 0.3877\n",
      "\n",
      "====> Epoch: 469 Train:\tAvg loss: 0.3841\n",
      "====> Epoch: 469 Valid:\tAvg loss: 0.3909\n",
      "\n",
      "====> Epoch: 470 Train:\tAvg loss: 0.3843\n",
      "====> Epoch: 470 Valid:\tAvg loss: 0.3860\n",
      "\n",
      "====> Epoch: 471 Train:\tAvg loss: 0.3840\n",
      "====> Epoch: 471 Valid:\tAvg loss: 0.3883\n",
      "\n",
      "====> Epoch: 472 Train:\tAvg loss: 0.3840\n",
      "====> Epoch: 472 Valid:\tAvg loss: 0.3857\n",
      "\n",
      "====> Epoch: 473 Train:\tAvg loss: 0.3842\n",
      "====> Epoch: 473 Valid:\tAvg loss: 0.3873\n",
      "\n",
      "====> Epoch: 474 Train:\tAvg loss: 0.3840\n",
      "====> Epoch: 474 Valid:\tAvg loss: 0.3844\n",
      "\n",
      "====> Epoch: 475 Train:\tAvg loss: 0.3840\n",
      "====> Epoch: 475 Valid:\tAvg loss: 0.3876\n",
      "\n",
      "====> Epoch: 476 Train:\tAvg loss: 0.3839\n",
      "====> Epoch: 476 Valid:\tAvg loss: 0.3872\n",
      "\n",
      "====> Epoch: 477 Train:\tAvg loss: 0.3841\n",
      "====> Epoch: 477 Valid:\tAvg loss: 0.3862\n",
      "\n",
      "====> Epoch: 478 Train:\tAvg loss: 0.3841\n",
      "====> Epoch: 478 Valid:\tAvg loss: 0.3883\n",
      "\n",
      "====> Epoch: 479 Train:\tAvg loss: 0.3840\n",
      "====> Epoch: 479 Valid:\tAvg loss: 0.3887\n",
      "\n",
      "====> Epoch: 480 Train:\tAvg loss: 0.3840\n",
      "====> Epoch: 480 Valid:\tAvg loss: 0.3860\n",
      "\n",
      "====> Epoch: 481 Train:\tAvg loss: 0.3841\n",
      "====> Epoch: 481 Valid:\tAvg loss: 0.3866\n",
      "\n",
      "====> Epoch: 482 Train:\tAvg loss: 0.3840\n",
      "====> Epoch: 482 Valid:\tAvg loss: 0.3909\n",
      "\n",
      "====> Epoch: 483 Train:\tAvg loss: 0.3839\n",
      "====> Epoch: 483 Valid:\tAvg loss: 0.3909\n",
      "\n",
      "====> Epoch: 484 Train:\tAvg loss: 0.3840\n",
      "====> Epoch: 484 Valid:\tAvg loss: 0.3926\n",
      "\n",
      "====> Epoch: 485 Train:\tAvg loss: 0.3839\n",
      "====> Epoch: 485 Valid:\tAvg loss: 0.3929\n",
      "\n",
      "====> Epoch: 486 Train:\tAvg loss: 0.3839\n",
      "====> Epoch: 486 Valid:\tAvg loss: 0.3880\n",
      "\n",
      "====> Epoch: 487 Train:\tAvg loss: 0.3841\n",
      "====> Epoch: 487 Valid:\tAvg loss: 0.3882\n",
      "\n",
      "====> Epoch: 488 Train:\tAvg loss: 0.3840\n",
      "====> Epoch: 488 Valid:\tAvg loss: 0.3877\n",
      "\n",
      "====> Epoch: 489 Train:\tAvg loss: 0.3838\n",
      "====> Epoch: 489 Valid:\tAvg loss: 0.3881\n",
      "\n",
      "====> Epoch: 490 Train:\tAvg loss: 0.3839\n",
      "====> Epoch: 490 Valid:\tAvg loss: 0.3886\n",
      "\n",
      "====> Epoch: 491 Train:\tAvg loss: 0.3838\n",
      "====> Epoch: 491 Valid:\tAvg loss: 0.3913\n",
      "\n",
      "====> Epoch: 492 Train:\tAvg loss: 0.3838\n",
      "====> Epoch: 492 Valid:\tAvg loss: 0.3866\n",
      "\n",
      "====> Epoch: 493 Train:\tAvg loss: 0.3838\n",
      "====> Epoch: 493 Valid:\tAvg loss: 0.3857\n",
      "\n",
      "====> Epoch: 494 Train:\tAvg loss: 0.3837\n",
      "====> Epoch: 494 Valid:\tAvg loss: 0.3898\n",
      "\n",
      "====> Epoch: 495 Train:\tAvg loss: 0.3837\n",
      "====> Epoch: 495 Valid:\tAvg loss: 0.3858\n",
      "\n",
      "====> Epoch: 496 Train:\tAvg loss: 0.3838\n",
      "====> Epoch: 496 Valid:\tAvg loss: 0.3857\n",
      "\n",
      "====> Epoch: 497 Train:\tAvg loss: 0.3837\n",
      "====> Epoch: 497 Valid:\tAvg loss: 0.3855\n",
      "\n",
      "====> Epoch: 498 Train:\tAvg loss: 0.3838\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "====> Epoch: 498 Valid:\tAvg loss: 0.3883\n",
      "\n",
      "====> Epoch: 499 Train:\tAvg loss: 0.3838\n",
      "====> Epoch: 499 Valid:\tAvg loss: 0.3880\n",
      "\n",
      "====> Epoch: 500 Train:\tAvg loss: 0.3837\n",
      "====> Epoch: 500 Valid:\tAvg loss: 0.3874\n",
      "\n"
     ]
    }
   ],
   "source": [
    "trainer.fit(X_subset, scalars, results_dir / \"autoencoder\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "id": "da9b249b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pd.DataFrame(trainer.loss_curve_).plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "id": "fd1015da",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.38374280276814476"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = np.load(results_dir / \"X.npy\")\n",
    "z, loss = trainer.predict(X)\n",
    "loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "id": "b4db9957",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_scatter(z[:, :3], e_rmsd[-1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "id": "d2ef24ed",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[0m\u001b[01;34mfrontera\u001b[0m/  \u001b[01;34mrun_final\u001b[0m/  \u001b[01;34msmall_data_production-cross2\u001b[0m/  small_positions_7egq.npy\r\n"
     ]
    }
   ],
   "source": [
    "ls data\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "id": "4901d4da",
   "metadata": {},
   "outputs": [],
   "source": [
    "labels = [] # 0\n",
    "for p in data_files:\n",
    "    len_data = len(np.load(p))\n",
    "    if \"small_positions\" in str(p):\n",
    "        labels.extend([0] * len_data) # post\n",
    "    elif \"run_final\" in str(p) or \"frontera\" in str(p):\n",
    "        labels.extend([1] * len_data) # pre\n",
    "    elif \"small_data_production-cross2\" in str(p):\n",
    "        labels.extend([2] * len_data) # ddmd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "id": "c4b6afba",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "866708"
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "id": "4cd5c2b5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[<AxesSubplot:title={'center':'0'}>]], dtype=object)"
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pd.DataFrame(labels).hist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "id": "2a16ca03",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.save(results_dir / \"simulation_labels.npy\", labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "id": "fb3f7c70",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_scatter(z[:, :3], labels)\n",
    "show_scatter(z[:, 3:6], labels)\n",
    "show_scatter(z[:, 6:9], labels)\n",
    "show_scatter(z[:, 9:12], labels)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "id": "9ab4975f",
   "metadata": {},
   "outputs": [],
   "source": [
    "embeddings = {\"pre\": [], \"post\": [], \"ddmd\": []}\n",
    "for i, label in enumerate(labels):\n",
    "    if label == 0:\n",
    "        embeddings[\"post\"].append(z[i])\n",
    "    elif label == 1:\n",
    "        embeddings[\"pre\"].append(z[i])\n",
    "    elif label == 2:\n",
    "        embeddings[\"ddmd\"].append(z[i])\n",
    "embeddings[\"pre\"] = np.array(embeddings[\"pre\"])\n",
    "embeddings[\"post\"] = np.array(embeddings[\"post\"])\n",
    "embeddings[\"ddmd\"] = np.array(embeddings[\"ddmd\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "id": "07feb90e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(401185, 16)"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "embeddings[\"pre\"].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "id": "a31b34d3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(260723, 16)"
      ]
     },
     "execution_count": 103,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "embeddings[\"post\"].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "id": "f5277155",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(204800, 16)"
      ]
     },
     "execution_count": 104,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "embeddings[\"ddmd\"].shape # Each frame is a picosecond for DDMD run, meaning we have 0.2048 microseconds total"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "id": "37837a24",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.save(results_dir / \"pre_embeddings.npy\", embeddings[\"pre\"])\n",
    "np.save(results_dir / \"post_embeddings.npy\", embeddings[\"post\"])\n",
    "np.save(results_dir / \"ddmd_embeddings.npy\", embeddings[\"ddmd\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "32f57f37",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}