03.2_TrainNN.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train NN"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This notebook is used for testing and running model fitting pipelines defined in `train_nn.py`.\n",
    "\n",
    "It's run in the `train_nn` conda environment instead of the usual `nnn` environment."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib\n",
    "from matplotlib.ticker import (MultipleLocator, FormatStrFormatter,\n",
    "                               AutoMinorLocator)\n",
    "import seaborn as sns\n",
    "import json, os, pickle\n",
    "from collections import defaultdict\n",
    "import nupack\n",
    "import sklearn\n",
    "from scipy import stats\n",
    "\n",
    "\n",
    "from tqdm import tqdm\n",
    "import itertools\n",
    "\n",
    "from nnn import util, fileio, mupack, plotting, modeling\n",
    "import nnn.motif_fit as mf\n",
    "import wandb\n",
    "\n",
    "# suppress warnings \n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "from nnn import train_nn as tnn"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## *dna24*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Config for the 178 param model used to make *dna24*:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Config for the 178 param model used to make *dna24*\n",
    "\"\"\"\n",
    "fixed_pclass = ['hairpin_size', 'interior_size', 'bulge_size', 'hairpin_triloop', 'hairpin_tetraloop', 'terminal_mismatch', 'stack']\n",
    "config = dict(\n",
    "    use_train_set_ratio = 1.,\n",
    "    secondary_struct = 'target', # target uses the designed structures, mfe uses the predicted mfe structures\n",
    "    fit_method = 'svd',\n",
    "    feature_method = 'get_nupack_feature_list',\n",
    "    fit_intercept=False, \n",
    "    symmetry=False,\n",
    "    fix_some_coef=True,\n",
    "    fixed_pclass = fixed_pclass,\n",
    "    test_mode = 'test', # {'val', 'test'}\n",
    "    use_model_from = 'json', # {'lr_dict', 'json'}\n",
    "    )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run model training only:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\u001b[34m\u001b[1mwandb\u001b[0m: Currently logged in as: \u001b[33mkeyuxi\u001b[0m. Use \u001b[1m`wandb login --relogin`\u001b[0m to force relogin\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "wandb version 0.18.2 is available!  To upgrade, please run:\n",
       " $ pip install wandb --upgrade"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Tracking run with wandb version 0.15.12"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Run data is saved locally in <code>/Users/yuxi/workspace/nnn_paper/wandb/run-20240930_154355-lt9igbsd</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Syncing run <strong><a href='https://wandb.ai/keyuxi/nnn_paper/runs/lt9igbsd' target=\"_blank\">glorious-snowflake-1</a></strong> to <a href='https://wandb.ai/keyuxi/nnn_paper' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View project at <a href='https://wandb.ai/keyuxi/nnn_paper' target=\"_blank\">https://wandb.ai/keyuxi/nnn_paper</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run at <a href='https://wandb.ai/keyuxi/nnn_paper/runs/lt9igbsd' target=\"_blank\">https://wandb.ai/keyuxi/nnn_paper/runs/lt9igbsd</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "feats (30872, 185) Index(['bulge_size#1', 'bulge_size#2', 'hairpin_loop_mid#A'], dtype='object')\n",
      "fixed_feature_names: ['bulge_size#1', 'bulge_size#2', 'hairpin_size#3', 'hairpin_size#4', 'hairpin_size#7', 'hairpin_size#8', 'hairpin_size#9', 'hairpin_size#10', 'hairpin_size#11', 'hairpin_size#12', 'hairpin_size#13', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#23', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "rank 152\n",
      "shape of A_unknown (24689, 152)\n",
      "rank 152\n",
      "shape of A_unknown (24689, 152)\n",
      "\n",
      "Template file: ./models/dna04.json\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "Nothing happened.\n",
      "\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -3.684\n",
      "\tMean value of fitted new_dict is -13.355\n",
      "\tMean value of centered new_dict is -3.684\n",
      "\n",
      "group hairpin_size\n",
      "Nothing happened.\n",
      "\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -2.913\n",
      "\tMean value of fitted new_dict is -5.340\n",
      "\tMean value of centered new_dict is -2.913\n",
      "\n",
      "group interior_size\n",
      "Nothing happened.\n",
      "\n",
      "group stack\n",
      "\tMean value of old_dict is -8.275\n",
      "\tMean value of fitted new_dict is -8.275\n",
      "\tMean value of centered new_dict is -8.275\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 1.100\n",
      "\tMean value of fitted new_dict is -1.033\n",
      "\tMean value of centered new_dict is 1.100\n",
      "\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "Nothing happened.\n",
      "\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -0.715\n",
      "\tMean value of fitted new_dict is -3.964\n",
      "\tMean value of centered new_dict is -0.715\n",
      "\n",
      "group hairpin_size\n",
      "Nothing happened.\n",
      "\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -0.640\n",
      "\tMean value of fitted new_dict is -0.701\n",
      "\tMean value of centered new_dict is -0.640\n",
      "\n",
      "group interior_size\n",
      "Nothing happened.\n",
      "\n",
      "group stack\n",
      "\tMean value of old_dict is -1.406\n",
      "\tMean value of fitted new_dict is -1.406\n",
      "\tMean value of centered new_dict is -1.406\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 0.025\n",
      "\tMean value of fitted new_dict is 0.145\n",
      "\tMean value of centered new_dict is 0.025\n",
      "\n",
      "\n",
      "Centering dH of hairpin_triloop to ./models/dna04.json\n",
      "\tMean value of old_dict is -1.500\n",
      "\tMean value of fitted new_dict is 18.688\n",
      "\tMean value of centered new_dict is -1.500\n",
      "\n",
      "\n",
      "Centering dH of hairpin_tetraloop to ./models/dna04.json\n",
      "\tMean value of old_dict is -0.025\n",
      "\tMean value of fitted new_dict is 17.108\n",
      "\tMean value of centered new_dict is -0.025\n",
      "\n",
      "\n",
      "Centering dG of hairpin_triloop to ./models/dna04.json\n",
      "\tMean value of old_dict is -1.500\n",
      "\tMean value of fitted new_dict is 3.433\n",
      "\tMean value of centered new_dict is -1.500\n",
      "\n",
      "\n",
      "Centering dG of hairpin_tetraloop to ./models/dna04.json\n",
      "\tMean value of old_dict is -0.479\n",
      "\tMean value of fitted new_dict is 3.073\n",
      "\tMean value of centered new_dict is -0.479\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "lr_dict_1pass = tnn.train(config)\n",
    "dH_coef, dG_coef = lr_dict_1pass['dH'].coef_df, lr_dict_1pass['dG'].coef_df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run the entire model training and testing pipeline:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<wandb.sdk.lib.preinit.PreInitObject at 0x7f8581770040>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "wandb.config"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "Finishing last run (ID:lt9igbsd) before initializing another..."
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<style>\n",
       "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
       "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
       "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
       "    </style>\n",
       "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>▁</td></tr><tr><td>n_fixed_feat</td><td>▁</td></tr><tr><td>train_dG_mae</td><td>▁</td></tr><tr><td>train_dG_rsqr</td><td>▁</td></tr><tr><td>train_dH_mae</td><td>▁</td></tr><tr><td>train_dH_rsqr</td><td>▁</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>185</td></tr><tr><td>n_fixed_feat</td><td>33</td></tr><tr><td>param_set_file</td><td>./models/glorious-sn...</td></tr><tr><td>train_dG_mae</td><td>0.60934</td></tr><tr><td>train_dG_rsqr</td><td>0.62676</td></tr><tr><td>train_dH_mae</td><td>5.7068</td></tr><tr><td>train_dH_rsqr</td><td>0.43779</td></tr></table><br/></div></div>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run <strong style=\"color:#cdcd00\">glorious-snowflake-1</strong> at: <a href='https://wandb.ai/keyuxi/nnn_paper/runs/lt9igbsd' target=\"_blank\">https://wandb.ai/keyuxi/nnn_paper/runs/lt9igbsd</a><br/> View job at <a href='https://wandb.ai/keyuxi/nnn_paper/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjQ2NDQ3NzkyOA==/version_details/v0' target=\"_blank\">https://wandb.ai/keyuxi/nnn_paper/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjQ2NDQ3NzkyOA==/version_details/v0</a><br/>Synced 6 W&B file(s), 0 media file(s), 2 artifact file(s) and 0 other file(s)"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Find logs at: <code>./wandb/run-20240930_154355-lt9igbsd/logs</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Successfully finished last run (ID:lt9igbsd). Initializing new run:<br/>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "wandb version 0.18.2 is available!  To upgrade, please run:\n",
       " $ pip install wandb --upgrade"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Tracking run with wandb version 0.15.12"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Run data is saved locally in <code>/Users/yuxi/workspace/nnn_paper/wandb/run-20240930_155031-1pmo4zjj</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Syncing run <strong><a href='https://wandb.ai/keyuxi/NN/runs/1pmo4zjj' target=\"_blank\">true-smoke-147</a></strong> to <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View project at <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">https://wandb.ai/keyuxi/NN</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run at <a href='https://wandb.ai/keyuxi/NN/runs/1pmo4zjj' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/1pmo4zjj</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Finishing last run (ID:1pmo4zjj) before initializing another..."
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run <strong style=\"color:#cdcd00\">true-smoke-147</strong> at: <a href='https://wandb.ai/keyuxi/NN/runs/1pmo4zjj' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/1pmo4zjj</a><br/> View job at <a href='https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v9' target=\"_blank\">https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v9</a><br/>Synced 5 W&B file(s), 0 media file(s), 2 artifact file(s) and 0 other file(s)"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Find logs at: <code>./wandb/run-20240930_155031-1pmo4zjj/logs</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Successfully finished last run (ID:1pmo4zjj). Initializing new run:<br/>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "wandb version 0.18.2 is available!  To upgrade, please run:\n",
       " $ pip install wandb --upgrade"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Tracking run with wandb version 0.15.12"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Run data is saved locally in <code>/Users/yuxi/workspace/nnn_paper/wandb/run-20240930_155039-k2r8zvpq</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Syncing run <strong><a href='https://wandb.ai/keyuxi/nnn_paper/runs/k2r8zvpq' target=\"_blank\">elated-frost-2</a></strong> to <a href='https://wandb.ai/keyuxi/nnn_paper' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View project at <a href='https://wandb.ai/keyuxi/nnn_paper' target=\"_blank\">https://wandb.ai/keyuxi/nnn_paper</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run at <a href='https://wandb.ai/keyuxi/nnn_paper/runs/k2r8zvpq' target=\"_blank\">https://wandb.ai/keyuxi/nnn_paper/runs/k2r8zvpq</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "feats (30872, 185) Index(['bulge_size#1', 'bulge_size#2', 'hairpin_loop_mid#A'], dtype='object')\n",
      "fixed_feature_names: ['bulge_size#1', 'bulge_size#2', 'hairpin_size#3', 'hairpin_size#4', 'hairpin_size#7', 'hairpin_size#8', 'hairpin_size#9', 'hairpin_size#10', 'hairpin_size#11', 'hairpin_size#12', 'hairpin_size#13', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#23', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "rank 152\n",
      "shape of A_unknown (24689, 152)\n",
      "rank 152\n",
      "shape of A_unknown (24689, 152)\n",
      "\n",
      "Template file: ./models/dna04.json\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "Nothing happened.\n",
      "\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -3.684\n",
      "\tMean value of fitted new_dict is -13.355\n",
      "\tMean value of centered new_dict is -3.684\n",
      "\n",
      "group hairpin_size\n",
      "Nothing happened.\n",
      "\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -2.913\n",
      "\tMean value of fitted new_dict is -5.340\n",
      "\tMean value of centered new_dict is -2.913\n",
      "\n",
      "group interior_size\n",
      "Nothing happened.\n",
      "\n",
      "group stack\n",
      "\tMean value of old_dict is -8.275\n",
      "\tMean value of fitted new_dict is -8.275\n",
      "\tMean value of centered new_dict is -8.275\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 1.100\n",
      "\tMean value of fitted new_dict is -1.033\n",
      "\tMean value of centered new_dict is 1.100\n",
      "\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "Nothing happened.\n",
      "\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -0.715\n",
      "\tMean value of fitted new_dict is -3.964\n",
      "\tMean value of centered new_dict is -0.715\n",
      "\n",
      "group hairpin_size\n",
      "Nothing happened.\n",
      "\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -0.640\n",
      "\tMean value of fitted new_dict is -0.701\n",
      "\tMean value of centered new_dict is -0.640\n",
      "\n",
      "group interior_size\n",
      "Nothing happened.\n",
      "\n",
      "group stack\n",
      "\tMean value of old_dict is -1.406\n",
      "\tMean value of fitted new_dict is -1.406\n",
      "\tMean value of centered new_dict is -1.406\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 0.025\n",
      "\tMean value of fitted new_dict is 0.145\n",
      "\tMean value of centered new_dict is 0.025\n",
      "\n",
      "\n",
      "Centering dH of hairpin_triloop to ./models/dna04.json\n",
      "\tMean value of old_dict is -1.500\n",
      "\tMean value of fitted new_dict is 18.688\n",
      "\tMean value of centered new_dict is -1.500\n",
      "\n",
      "\n",
      "Centering dH of hairpin_tetraloop to ./models/dna04.json\n",
      "\tMean value of old_dict is -0.025\n",
      "\tMean value of fitted new_dict is 17.108\n",
      "\tMean value of centered new_dict is -0.025\n",
      "\n",
      "\n",
      "Centering dG of hairpin_triloop to ./models/dna04.json\n",
      "\tMean value of old_dict is -1.500\n",
      "\tMean value of fitted new_dict is 3.433\n",
      "\tMean value of centered new_dict is -1.500\n",
      "\n",
      "\n",
      "Centering dG of hairpin_tetraloop to ./models/dna04.json\n",
      "\tMean value of old_dict is -0.479\n",
      "\tMean value of fitted new_dict is 3.073\n",
      "\tMean value of centered new_dict is -0.479\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Traceback (most recent call last):\n",
      "  File \"/Users/yuxi/workspace/nnn_paper/nnn/train_nn.py\", line 24, in model_pipeline\n",
      "    test(config, json_file=json_file)\n",
      "  File \"/Users/yuxi/workspace/nnn_paper/nnn/train_nn.py\", line 283, in test\n",
      "    val_result_df_dict[dataset_name] = modeling.make_model_validation_df(\n",
      "  File \"/Users/yuxi/workspace/nnn_paper/nnn/modeling.py\", line 26, in make_model_validation_df\n",
      "    pred_df = get_model_prediction(df=val_data_df, **pred_kwargs)\n",
      "  File \"/Users/yuxi/workspace/nnn_paper/nnn/modeling.py\", line 83, in get_model_prediction\n",
      "    result_df = globals()[f'run_{model}'](seq_list, struct_list, sodium, model_param_file, model_kwargs)\n",
      "  File \"/Users/yuxi/workspace/nnn_paper/nnn/modeling.py\", line 199, in run_nupack\n",
      "    Tm = util.calculate_tm(seq, struct,\n",
      "  File \"/Users/yuxi/workspace/nnn_paper/nnn/util.py\", line 535, in calculate_tm\n",
      "    Q_K = calculate_distance_2_equilibrium(seq, DNA_conc, get_nupack_model(T), target_struct=target_struct)\n",
      "  File \"/Users/yuxi/workspace/nnn_paper/nnn/util.py\", line 525, in <lambda>\n",
      "    get_nupack_model = lambda T: nupack.Model(material=param_set, celsius=T, sodium=sodium)\n",
      "  File \"/Users/yuxi/opt/anaconda3/envs/nn_train/lib/python3.8/site-packages/nupack/rebind/render.py\", line 243, in wrap\n",
      "    return _old(*args, _fun_=_orig, **kwargs)\n",
      "  File \"/Users/yuxi/opt/anaconda3/envs/nn_train/lib/python3.8/site-packages/nupack/model.py\", line 154, in __init__\n",
      "    _fun_(self, Ensemble.get(ensemble), material, conditions, gu, return_type=cls)\n",
      "  File \"/Users/yuxi/opt/anaconda3/envs/nn_train/lib/python3.8/site-packages/nupack/rebind/render.py\", line 86, in __init__\n",
      "    self.move_from(_new(*args, return_type=return_type, signature=signature))\n",
      "KeyboardInterrupt\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[10], line 1\u001b[0m\n\u001b[0;32m----> 1\u001b[0m \u001b[43mtnn\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mmodel_pipeline\u001b[49m\u001b[43m(\u001b[49m\u001b[43mconfig\u001b[49m\u001b[43m)\u001b[49m\n",
      "File \u001b[0;32m~/workspace/nnn_paper/nnn/train_nn.py:24\u001b[0m, in \u001b[0;36mmodel_pipeline\u001b[0;34m(hyperparameters, tags)\u001b[0m\n\u001b[1;32m     22\u001b[0m     \u001b[38;5;28;01melif\u001b[39;00m config[\u001b[38;5;124m'\u001b[39m\u001b[38;5;124muse_model_from\u001b[39m\u001b[38;5;124m'\u001b[39m] \u001b[38;5;241m==\u001b[39m \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mjson\u001b[39m\u001b[38;5;124m'\u001b[39m:\n\u001b[1;32m     23\u001b[0m         json_file \u001b[38;5;241m=\u001b[39m \u001b[38;5;124m'\u001b[39m\u001b[38;5;124m./models/\u001b[39m\u001b[38;5;132;01m%s\u001b[39;00m\u001b[38;5;124m.json\u001b[39m\u001b[38;5;124m'\u001b[39m \u001b[38;5;241m%\u001b[39m wandb\u001b[38;5;241m.\u001b[39mrun\u001b[38;5;241m.\u001b[39mname\n\u001b[0;32m---> 24\u001b[0m         \u001b[43mtest\u001b[49m\u001b[43m(\u001b[49m\u001b[43mconfig\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mjson_file\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mjson_file\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m     27\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m lr_dict\n",
      "File \u001b[0;32m~/workspace/nnn_paper/nnn/train_nn.py:283\u001b[0m, in \u001b[0;36mtest\u001b[0;34m(config, lr_dict, json_file, debug, log_wandb, save_val_result_df, save_metric_json)\u001b[0m\n\u001b[1;32m    280\u001b[0m             val_kwargs \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mdict\u001b[39m()\n\u001b[1;32m    282\u001b[0m         \u001b[38;5;66;03m# Actually run\u001b[39;00m\n\u001b[0;32m--> 283\u001b[0m         val_result_df_dict[dataset_name] \u001b[38;5;241m=\u001b[39m \u001b[43mmodeling\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mmake_model_validation_df\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m    284\u001b[0m \u001b[43m            \u001b[49m\u001b[43mval_df\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    285\u001b[0m \u001b[43m            \u001b[49m\u001b[43mmodel\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43mnupack\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\n\u001b[1;32m    286\u001b[0m \u001b[43m            \u001b[49m\u001b[43mmodel_param_file\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mjson_file\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    287\u001b[0m \u001b[43m            \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mval_kwargs\u001b[49m\n\u001b[1;32m    288\u001b[0m \u001b[43m        \u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    290\u001b[0m val_result_df \u001b[38;5;241m=\u001b[39m pd\u001b[38;5;241m.\u001b[39mconcat(\n\u001b[1;32m    291\u001b[0m     val_result_df_dict,\n\u001b[1;32m    292\u001b[0m     axis\u001b[38;5;241m=\u001b[39m\u001b[38;5;241m0\u001b[39m,\n\u001b[1;32m    293\u001b[0m     join\u001b[38;5;241m=\u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mouter\u001b[39m\u001b[38;5;124m'\u001b[39m,\n\u001b[1;32m    294\u001b[0m     keys\u001b[38;5;241m=\u001b[39mval_result_df_dict\u001b[38;5;241m.\u001b[39mkeys()\n\u001b[1;32m    295\u001b[0m )\n\u001b[1;32m    297\u001b[0m \u001b[38;5;66;03m### Evaluate val_result_df ###\u001b[39;00m\n",
      "File \u001b[0;32m~/workspace/nnn_paper/nnn/modeling.py:26\u001b[0m, in \u001b[0;36mmake_model_validation_df\u001b[0;34m(val_data_df, **pred_kwargs)\u001b[0m\n\u001b[1;32m     15\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21mmake_model_validation_df\u001b[39m(val_data_df, \u001b[38;5;241m*\u001b[39m\u001b[38;5;241m*\u001b[39mpred_kwargs):\n\u001b[1;32m     16\u001b[0m \u001b[38;5;250m    \u001b[39m\u001b[38;5;124;03m\"\"\"\u001b[39;00m\n\u001b[1;32m     17\u001b[0m \u001b[38;5;124;03m    Args:\u001b[39;00m\n\u001b[1;32m     18\u001b[0m \u001b[38;5;124;03m        val_data_df - dataframe. Uses column names\u001b[39;00m\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m     24\u001b[0m \u001b[38;5;124;03m            Only has Index, RefSeq, TargetStruct, actual and predicted values\u001b[39;00m\n\u001b[1;32m     25\u001b[0m \u001b[38;5;124;03m    \"\"\"\u001b[39;00m\n\u001b[0;32m---> 26\u001b[0m     pred_df \u001b[38;5;241m=\u001b[39m \u001b[43mget_model_prediction\u001b[49m\u001b[43m(\u001b[49m\u001b[43mdf\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mval_data_df\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mpred_kwargs\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m     28\u001b[0m     pred_columns \u001b[38;5;241m=\u001b[39m [c\u001b[38;5;241m+\u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;124m_pred\u001b[39m\u001b[38;5;124m'\u001b[39m \u001b[38;5;28;01mfor\u001b[39;00m c \u001b[38;5;129;01min\u001b[39;00m pred_df\u001b[38;5;241m.\u001b[39mcolumns]\n\u001b[1;32m     29\u001b[0m     pred_df\u001b[38;5;241m.\u001b[39mcolumns \u001b[38;5;241m=\u001b[39m pred_columns\n",
      "File \u001b[0;32m~/workspace/nnn_paper/nnn/modeling.py:83\u001b[0m, in \u001b[0;36mget_model_prediction\u001b[0;34m(df, seq_list, struct_list, sodium, model, model_param_file, model_kwargs, append_df_suffix)\u001b[0m\n\u001b[1;32m     80\u001b[0m         \u001b[38;5;28;01massert\u001b[39;00m \u001b[38;5;28mlen\u001b[39m(seq_list) \u001b[38;5;241m==\u001b[39m struct_list\n\u001b[1;32m     82\u001b[0m     \u001b[38;5;66;03m# run model prediction\u001b[39;00m\n\u001b[0;32m---> 83\u001b[0m     result_df \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mglobals\u001b[39;49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;124;43mf\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43mrun_\u001b[39;49m\u001b[38;5;132;43;01m{\u001b[39;49;00m\u001b[43mmodel\u001b[49m\u001b[38;5;132;43;01m}\u001b[39;49;00m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m]\u001b[49m\u001b[43m(\u001b[49m\u001b[43mseq_list\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mstruct_list\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43msodium\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mmodel_param_file\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mmodel_kwargs\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m     85\u001b[0m \u001b[38;5;28;01melif\u001b[39;00m model \u001b[38;5;241m==\u001b[39m \u001b[38;5;124m'\u001b[39m\u001b[38;5;124munet\u001b[39m\u001b[38;5;124m'\u001b[39m:\n\u001b[1;32m     86\u001b[0m     \u001b[38;5;66;03m# TODO\u001b[39;00m\n\u001b[1;32m     87\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m result_df\n",
      "File \u001b[0;32m~/workspace/nnn_paper/nnn/modeling.py:199\u001b[0m, in \u001b[0;36mrun_nupack\u001b[0;34m(seq_list, struct_list, sodium, model_param_file, model_kwargs)\u001b[0m\n\u001b[1;32m    196\u001b[0m     \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[1;32m    197\u001b[0m         dna_conc \u001b[38;5;241m=\u001b[39m DNA_conc[i]\n\u001b[0;32m--> 199\u001b[0m     Tm \u001b[38;5;241m=\u001b[39m \u001b[43mutil\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mcalculate_tm\u001b[49m\u001b[43m(\u001b[49m\u001b[43mseq\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mstruct\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    200\u001b[0m \u001b[43m         \u001b[49m\u001b[43msodium\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;241;43m1.0\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mDNA_conc\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mdna_conc\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mparam_set\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mmodel_param_file\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    201\u001b[0m     row_result_dict \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mdict\u001b[39m(Tm\u001b[38;5;241m=\u001b[39mutil\u001b[38;5;241m.\u001b[39mget_Na_adjusted_Tm(Tm, \u001b[38;5;28;01mNone\u001b[39;00m, util\u001b[38;5;241m.\u001b[39mget_GC_content(\u001b[38;5;124m'\u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;241m.\u001b[39mjoin(seq)), Na\u001b[38;5;241m=\u001b[39mna))\n\u001b[1;32m    203\u001b[0m result_df\u001b[38;5;241m.\u001b[39miloc[i,:] \u001b[38;5;241m=\u001b[39m {k:v \u001b[38;5;28;01mfor\u001b[39;00m k,v \u001b[38;5;129;01min\u001b[39;00m row_result_dict\u001b[38;5;241m.\u001b[39mitems() \u001b[38;5;28;01mif\u001b[39;00m k \u001b[38;5;129;01min\u001b[39;00m PNAMES}\n",
      "File \u001b[0;32m~/workspace/nnn_paper/nnn/util.py:535\u001b[0m, in \u001b[0;36mcalculate_tm\u001b[0;34m(seq, target_struct, sodium, DNA_conc, param_set)\u001b[0m\n\u001b[1;32m    533\u001b[0m \u001b[38;5;28;01mwhile\u001b[39;00m \u001b[38;5;28;01mTrue\u001b[39;00m:\n\u001b[1;32m    534\u001b[0m     T \u001b[38;5;241m=\u001b[39m (lo \u001b[38;5;241m+\u001b[39m hi) \u001b[38;5;241m/\u001b[39m \u001b[38;5;241m2\u001b[39m\n\u001b[0;32m--> 535\u001b[0m     Q_K \u001b[38;5;241m=\u001b[39m calculate_distance_2_equilibrium(seq, DNA_conc, \u001b[43mget_nupack_model\u001b[49m\u001b[43m(\u001b[49m\u001b[43mT\u001b[49m\u001b[43m)\u001b[49m, target_struct\u001b[38;5;241m=\u001b[39mtarget_struct)\n\u001b[1;32m    537\u001b[0m     \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28mabs\u001b[39m(Q_K) \u001b[38;5;241m<\u001b[39m \u001b[38;5;241m1e-2\u001b[39m:\n\u001b[1;32m    538\u001b[0m         \u001b[38;5;28;01mreturn\u001b[39;00m T\n",
      "File \u001b[0;32m~/workspace/nnn_paper/nnn/util.py:525\u001b[0m, in \u001b[0;36mcalculate_tm.<locals>.<lambda>\u001b[0;34m(T)\u001b[0m\n\u001b[1;32m    522\u001b[0m \u001b[38;5;250m\u001b[39m\u001b[38;5;124;03m'''Simple Tm calculation using bisection to find where unpaired = paired population'''\u001b[39;00m\n\u001b[1;32m    524\u001b[0m lo, hi \u001b[38;5;241m=\u001b[39m \u001b[38;5;241m0\u001b[39m, \u001b[38;5;241m100\u001b[39m\n\u001b[0;32m--> 525\u001b[0m get_nupack_model \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;01mlambda\u001b[39;00m T: \u001b[43mnupack\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mModel\u001b[49m\u001b[43m(\u001b[49m\u001b[43mmaterial\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mparam_set\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mcelsius\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mT\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43msodium\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43msodium\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    527\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m calculate_distance_2_equilibrium(seq, DNA_conc, get_nupack_model(lo), target_struct\u001b[38;5;241m=\u001b[39mtarget_struct) \u001b[38;5;241m<\u001b[39m \u001b[38;5;241m0\u001b[39m:\n\u001b[1;32m    528\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m lo\n",
      "File \u001b[0;32m~/opt/anaconda3/envs/nn_train/lib/python3.8/site-packages/nupack/rebind/render.py:243\u001b[0m, in \u001b[0;36mrender_function.<locals>.wrap\u001b[0;34m(_orig, _bind, _old, gil, signature, *args, **kwargs)\u001b[0m\n\u001b[1;32m    241\u001b[0m args \u001b[38;5;241m=\u001b[39m (a \u001b[38;5;28;01mif\u001b[39;00m t \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m \u001b[38;5;129;01mor\u001b[39;00m a \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m \u001b[38;5;28;01melse\u001b[39;00m make_callback(a, t) \u001b[38;5;28;01mfor\u001b[39;00m a, t \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28mzip\u001b[39m(bound\u001b[38;5;241m.\u001b[39margs, types))\n\u001b[1;32m    242\u001b[0m kwargs \u001b[38;5;241m=\u001b[39m {k: (v \u001b[38;5;28;01mif\u001b[39;00m t \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m \u001b[38;5;129;01mor\u001b[39;00m v \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m \u001b[38;5;28;01melse\u001b[39;00m make_callback(v, t)) \u001b[38;5;28;01mfor\u001b[39;00m (k, v), t \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28mzip\u001b[39m(bound\u001b[38;5;241m.\u001b[39mkwargs\u001b[38;5;241m.\u001b[39mitems(), types[\u001b[38;5;28mlen\u001b[39m(bound\u001b[38;5;241m.\u001b[39margs):])}\n\u001b[0;32m--> 243\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43m_old\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43m_fun_\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43m_orig\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n",
      "File \u001b[0;32m~/opt/anaconda3/envs/nn_train/lib/python3.8/site-packages/nupack/model.py:154\u001b[0m, in \u001b[0;36mModel.__init__\u001b[0;34m(self, ensemble, material, wobble, kelvin, celsius, sodium, magnesium, bits, _fun_)\u001b[0m\n\u001b[1;32m    151\u001b[0m \u001b[38;5;28mcls\u001b[39m \u001b[38;5;241m=\u001b[39m match(k \u001b[38;5;28;01mfor\u001b[39;00m k, v \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_metadata_\u001b[38;5;241m.\u001b[39mitems() \u001b[38;5;28;01mif\u001b[39;00m v\u001b[38;5;241m.\u001b[39mcast(\u001b[38;5;28mint\u001b[39m) \u001b[38;5;241m==\u001b[39m bits)\n\u001b[1;32m    152\u001b[0m gu \u001b[38;5;241m=\u001b[39m \u001b[38;5;241m2\u001b[39m \u001b[38;5;28;01mif\u001b[39;00m wobble \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m \u001b[38;5;28;01melse\u001b[39;00m \u001b[38;5;28mint\u001b[39m(\u001b[38;5;28mbool\u001b[39m(wobble))\n\u001b[0;32m--> 154\u001b[0m \u001b[43m_fun_\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mEnsemble\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mget\u001b[49m\u001b[43m(\u001b[49m\u001b[43mensemble\u001b[49m\u001b[43m)\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mmaterial\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mconditions\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mgu\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mreturn_type\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43mcls\u001b[39;49m\u001b[43m)\u001b[49m\n",
      "File \u001b[0;32m~/opt/anaconda3/envs/nn_train/lib/python3.8/site-packages/nupack/rebind/render.py:86\u001b[0m, in \u001b[0;36mrender_init.<locals>.__init__\u001b[0;34m(self, _new, return_type, signature, *args)\u001b[0m\n\u001b[1;32m     85\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21m__init__\u001b[39m(\u001b[38;5;28mself\u001b[39m, \u001b[38;5;241m*\u001b[39margs, _new\u001b[38;5;241m=\u001b[39minit, return_type\u001b[38;5;241m=\u001b[39m\u001b[38;5;28;01mNone\u001b[39;00m, signature\u001b[38;5;241m=\u001b[39m\u001b[38;5;28;01mNone\u001b[39;00m):\n\u001b[0;32m---> 86\u001b[0m     \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mmove_from(\u001b[43m_new\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mreturn_type\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mreturn_type\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43msignature\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43msignature\u001b[49m\u001b[43m)\u001b[49m)\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAqUAAAFkCAYAAAAZhEK+AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/OQEPoAAAACXBIWXMAAA9hAAAPYQGoP6dpAAAbtElEQVR4nO3db2yVZ/nA8Yu5lLYrylSKGkbLMILOhRY2QTDoqsQ/pC7GmGLMJFU359wanJItm9uY2xyGOOuG0mwzvjToG9mYhP3xhduEIQRNDFsWz4AyitLqKoNyWoHn98J4ftYW9xRabs7x80mW7Nx9OFx3mFe+UkonZVmWBQAAJHRB6gEAAECUAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBI7oyi9JFHHonbb7991I/19vZGe3t7NDc3x/Lly2P37t1nNSBApbJLAf7fmKJ0aGgoOjs74/vf//5pn7njjjti7ty58cILL8R1110XN998c5w8efKsBwWoFHYpwEhjitJ777039uzZEytWrBj140ePHo1nn302brjhhqiqqoqrr746pkyZEtu3bx+XYQEqgV0KMNKFY3n4pptuimnTpsVDDz0Uf/7zn0d8vLu7Oy6++OKYMmVK6ayxsTEKhUIsWbJk1Pc8fPhw9Pb2jjg/cuRIFAqFeN/73heTJ08ey5gAb2hwcDBeffXV+NCHPhRvfetbz+nPPd671B4FUhnPXTqmKJ02bdp//fjAwMCIxVddXR3FYvG0P2bjxo2xfv36sYwBMG7WrVsXn/70p8/pzzneu9QeBVIbj106pih9IzU1NTE4ODjsrFgsRm1t7Wl/TFtbW7S0tIw4f/HFF+P222+PdevWxezZs8dzTIAoFAqxevXqmDFjRupRRhjrLrVHgVTGc5eOa5Q2NDREf39/HD16NOrq6iIiYu/evaf9c1MREfX19VFfX3/aj8+ePTsuu+yy8RwToOR8/LT2WHepPQqkNh67dFz/ntK6urpYsmRJPPjggzE0NBSPPfZY9Pf3xxVXXDGePw1ARbNLgf9FZx2lPT090dzcHD09PRHxz68q3bdvX3zwgx+MRx99NH70ox9FVVXVWQ8KUMnsUuB/3Rl9+v6mm24q/fu73vWuYX+p87Rp0+Lhhx8++8kAKpxdCvD/fJtRAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQ3JiidNeuXdHa2hpNTU3R3t4efX19I545ePBgrFy5MhYsWBCf+MQn4plnnhm3YQEqgV0KMFLuKC0Wi9HR0REdHR2xY8eOaGhoiLVr14547r777osrr7wydu7cGXfeeWd84xvfiGKxOK5DA5QruxRgdLmjdNu2bTF9+vRYtmxZVFVVxapVq2Lr1q0xMDAw7Lnu7u44depUnDp1KiZNmhQ1NTXjPjRAubJLAUZ3Yd4H9+/fH42NjaXXU6dOjdra2uju7o65c+eWzleuXBl33313dHV1RUREZ2dnVFdXn/Z9Dx8+HL29vSPOC4VC3tEAysZE7FJ7FKgEuaN0YGAgJk+ePOyspqZmxKeTTp06Fbfccku0tbXFc889F7feemtcfvnl8c53vnPU9924cWOsX7/+DEYHKD8TsUvtUaAS5I7SmpqaGBoaGnZ2/PjxqK2tLb3+y1/+Eg888EBs27YtLrjggmhpaYnm5uZ46qmn4otf/OKo79vW1hYtLS0jzguFQqxevTrveABlYSJ2qT0KVILcUTpr1qzYvHlz6XV/f38cO3YsZs6cWTrr6+uLf/zjH8N+3Jve9Ka48MLT/zT19fVRX18/lpkBytZE7FJ7FKgEub/QadGiRXHo0KHYsmVLDA0NRWdnZ7S0tAz7M07vfve746KLLoof//jHcerUqdi+fXvs2LEjli5dOiHDA5QbuxRgdLmjtLq6OjZs2BBdXV2xcOHCOHDgQKxZsyZ6enqiubk5enp6YvLkydHV1RXPP/98XHnllXHvvffGD37wg5gxY8ZE3gGgbNilAKPL/en7iIh58+bFpk2bRpzv3r279O+XXXZZ/OxnPzv7yQAqlF0KMJJvMwoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJjSlKd+3aFa2trdHU1BTt7e3R19c34plisRh33nlnLFmyJJYuXRq/+MUvxm1YgEpglwKMlDtKi8VidHR0REdHR+zYsSMaGhpi7dq1I5677777or+/P55++ul45JFH4nvf+17s27dvPGcGKFt2KcDockfptm3bYvr06bFs2bKoqqqKVatWxdatW2NgYKD0zNDQUDz++ONxxx13RE1NTcyZMyc2btwYb3/72ydkeIByY5cCjC53lO7fvz8aGxtLr6dOnRq1tbXR3d1dOtu3b1/U1dXF5s2b48Mf/nB8/OMfj5dffjnq6urGdWiAcmWXAozuwrwPDgwMxOTJk4ed1dTURLFYLL0+cuRI/O1vf4u9e/fG1q1bY8+ePXHttdfGnDlz4tJLLx31fQ8fPhy9vb0jzguFQt7RAMrGROxSexSoBLmjtKamJoaGhoadHT9+PGpra0uvq6qq4uTJk7Fq1aqorq6O+fPnx+LFi+P5558/bZRu3Lgx1q9ff4bjA5SXidil9ihQCXJH6axZs2Lz5s2l1/39/XHs2LGYOXNm6WzmzJkxadKkeP311+Otb31rREScOHEisiw77fu2tbVFS0vLiPNCoRCrV6/OOx5AWZiIXWqPApUg958pXbRoURw6dCi2bNkSQ0ND0dnZGS0tLVFdXV16ZurUqbF06dLo7OyMwcHB2LVrV2zfvj2uuuqq075vfX19XHbZZSP+mT179tndDOA8NBG71B4FKkHuKK2uro4NGzZEV1dXLFy4MA4cOBBr1qyJnp6eaG5ujp6enoiIWLduXUyaNCmWLl0aq1evjvvuuy8uueSSCbsAQDmxSwFGl/vT9xER8+bNi02bNo043717d+nf3/KWt8QDDzxw9pMBVCi7FGAk32YUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkN6Yo3bVrV7S2tkZTU1O0t7dHX1/faZ997bXXYvHixfHCCy+c9ZAAlcQuBRgpd5QWi8Xo6OiIjo6O2LFjRzQ0NMTatWtP+/y9994br7322rgMCVAp7FKA0eWO0m3btsX06dNj2bJlUVVVFatWrYqtW7fGwMDAiGd//etfx9GjR2PGjBnjOixAubNLAUZ3Yd4H9+/fH42NjaXXU6dOjdra2uju7o65c+eWzo8cORLr1q2Ln/70p3HNNde84fsePnw4ent7R5wXCoW8owGUjYnYpfYoUAlyR+nAwEBMnjx52FlNTU0Ui8VhZ/fff39cc8018Y53vCPX+27cuDHWr1+fdwyAsjYRu9QeBSpB7iitqamJoaGhYWfHjx+P2tra0utnn302uru747vf/W7uAdra2qKlpWXEeaFQiNWrV+d+H4ByMBG71B4FKkHuKJ01a1Zs3ry59Lq/vz+OHTsWM2fOLJ09+eSTsWfPnrjyyisjIuLYsWNx/fXXx3e+851obW0d9X3r6+ujvr7+TOcHKCsTsUvtUaAS5P5Cp0WLFsWhQ4diy5YtMTQ0FJ2dndHS0hLV1dWlZ+65557YvXt37Ny5M3bu3BkzZsyIrq6u0wYpwP8auxRgdLmjtLq6OjZs2BBdXV2xcOHCOHDgQKxZsyZ6enqiubk5enp6JnJOgIpglwKMLven7yMi5s2bF5s2bRpxvnv37lGff+qpp85sKoAKZpcCjOTbjAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByY4rSXbt2RWtrazQ1NUV7e3v09fWNeGbPnj2xYsWKWLBgQXzyk5+Mp59+etyGBagEdinASLmjtFgsRkdHR3R0dMSOHTuioaEh1q5dO+yZkydPxte//vX47Gc/G7/73e/irrvuiltuuSUOHjw47oMDlCO7FGB0uaN027ZtMX369Fi2bFlUVVXFqlWrYuvWrTEwMFB6pq+vL97//vfH5z73ubjgggti0aJF0dDQEC+++OKEDA9QbuxSgNHljtL9+/dHY2Nj6fXUqVOjtrY2uru7S2fTp0+Phx56qPS6p6cnCoVCzJkzZ3ymBShzdinA6C7M++DAwEBMnjx52FlNTU0Ui8VRn//73/8eN9xwQ7S1tcUll1xy2vc9fPhw9Pb2jjgvFAp5RwMoGxOxS+1RoBLkjtKampoYGhoadnb8+PGora0d8WxPT0985StfiXnz5sWtt976X99348aNsX79+rxjAJS1idil9ihQCXJH6axZs2Lz5s2l1/39/XHs2LGYOXPmsOdeeeWVWLlyZVx99dXxrW996w3ft62tLVpaWkacFwqFWL16dd7xAMrCROxSexSoBLmjdNGiRXHbbbfFli1b4qMf/Wh0dnZGS0tLVFdXl54ZHByM66+/Ptra2uLGG2/M9b719fVRX18/9skBytBE7FJ7FKgEub/Qqbq6OjZs2BBdXV2xcOHCOHDgQKxZsyZ6enqiubk5enp64plnnon9+/fHT37yk2hubi7986tf/Woi7wBQNuxSgNHl/p3SiIh58+bFpk2bRpzv3r07IiLe9a53xac+9anxmQygQtmlACP5NqMAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACC5MUXprl27orW1NZqamqK9vT36+vpGPNPb2xvt7e3R3Nwcy5cvj927d4/bsACVwC4FGCl3lBaLxejo6IiOjo7YsWNHNDQ0xNq1a0c8d8cdd8TcuXPjhRdeiOuuuy5uvvnmOHny5LgODVCu7FKA0eWO0m3btsX06dNj2bJlUVVVFatWrYqtW7fGwMBA6ZmjR4/Gs88+GzfccENUVVXF1VdfHVOmTInt27dPyPAA5cYuBRjdhXkf3L9/fzQ2NpZeT506NWpra6O7uzvmzp0bERHd3d1x8cUXx5QpU0rPNTY2RqFQiCVLloz6vocPH47e3t4R5y+++GJERBQKhbwjAuT2r90yODh4Tn/eidil9iiQynju0txROjAwEJMnTx52VlNTE8Vi8b8+U11dPeyZ/7Rx48ZYv379aT++evXqvCMCjNmePXti/vz55+znm4hdao8CqY3HLs0dpTU1NTE0NDTs7Pjx41FbWzvsmf8s5WKxOOyZ/9TW1hYtLS0jzv/whz/E3XffHd/97ndLv3tQSQqFQqxevTrWrVsXs2fPTj3OuKv0+0VU/h0r/X4vvfRS3Hbbbef8bhOxS+3RyvxvtNLvF1H5d6z0+0WM7y7NHaWzZs2KzZs3l1739/fHsWPHYubMmaWzhoaG6O/vj6NHj0ZdXV1EROzduzdWrFhx2vetr6+P+vr603587ty5cdlll+Uds+zMnj3b/cpcpd+x0u/35je/+Zz+fBOxS+3Ryv5vtNLvF1H5d6z0+0WMzy7N/YVOixYtikOHDsWWLVtiaGgoOjs7o6WlJaqrq0vP1NXVxZIlS+LBBx+MoaGheOyxx6K/vz+uuOKKsx4UoBLYpQCjyx2l1dXVsWHDhujq6oqFCxfGgQMHYs2aNdHT0xPNzc3R09MTERH33ntv7Nu3Lz74wQ/Go48+Gj/60Y+iqqpqwi4AUE7sUoDR5f70fUTEvHnzYtOmTSPO//0vdZ42bVo8/PDDZz8ZQIWySwFG8m1GAQBI7ryN0mnTpsWNN94Y06ZNSz3KhHC/8lfpd3S/8lfpd3S/8lfpd6z0+0WM7x0nZVmWjcNMAABwxs7b3ykFAOB/hygFACA5UQoAQHLJo3TXrl3R2toaTU1N0d7eHn19fSOe6e3tjfb29mhubo7ly5cP+2tTznd57rdnz55YsWJFLFiwID75yU/G008/nWDSM5Pnfv/y2muvxeLFi+OFF144hxOevTx3LBaLceedd8aSJUti6dKl8Ytf/CLBpGcmz/0OHjwYK1eujAULFsQnPvGJeOaZZxJMenYeeeSRuP3220f9WDnvmAh7NMIePd/Zo/ZoLllCx48fzxYvXpw9+eST2eDgYHbXXXdl3/zmN0c899WvfjVbu3ZtNjg4mP3yl7/MPvKRj2QnTpxIMPHY5LnfiRMnso985CPZz3/+8+zkyZPZtm3bsvnz52evvvpqoqnzy/vr9y8333xzNnfu3Gz79u3ncMqzk/eO3/72t7ObbropGxgYyF566aVswYIF2d69e8/9wGOU935f+9rXsoceeig7depU9vzzz2eXX355dvz48QQTj93g4GD2gx/8IJszZ0522223jfpMue6YLLNHs8wePd/Zo/9kj76xpFH661//OvvMZz5Tev3aa69l73//+7Njx46Vzl5//fXsfe97X3bkyJHSWWtra/bcc8+d01nPRJ77/fnPf85uvPHGYT/uM5/5TPbUU0+dsznPVJ77/cszzzyTXXfdddnHPvaxslqmee44ODiYzZs3Lzt8+HDp7E9/+lP2+uuvn9NZz0TeX8Ply5dnP/zhD7MTJ05kv/3tb7MPfOADZbNM77jjjuzaa6/N7rrrrlGXaTnvmCyzR7PMHj3f2aP/ZI++saSfvt+/f380NjaWXk+dOjVqa2uju7u7dNbd3R0XX3xxTJkypXTW2NgYhULhXI56RvLcb/r06fHQQw+VXvf09EShUIg5c+acy1HPSJ77RUQcOXIk1q1bF3ffffc5nvDs5bnjvn37oq6uLjZv3hwf/vCH4+Mf/3i8/PLLUVdXl2Discn7a7hy5cp4+OGH4/LLL48vf/nLcc899wz7Xu3ns5tuuikefvjheNvb3jbqx8t5x0TYoxH26PnOHv0ne/SNjenbjI63gYGBmDx58rCzmpqaKBaL//WZ6urqYc+cr/Lc79/9/e9/jxtuuCHa2trikksuORcjnpW897v//vvjmmuuiXe84x3ncrxxkeeOR44cib/97W+xd+/e2Lp1a+zZsyeuvfbamDNnTlx66aXneuQxyftreOrUqbjllluira0tnnvuubj11lvj8ssvj3e+853nctwz8kZ/oXM575gIe/Q/2aPnH3v0n+zRN5b0d0prampiaGho2Nnx48ejtrZ22DODg4PDnikWi8OeOV/lud+/9PT0xOc///l473vfG7feeuu5GvGs5Lnfs88+G93d3fH5z3/+XI83LvLcsaqqKk6ePBmrVq2K6urqmD9/fixevDief/75cz3umOW531/+8pd44IEH4gtf+EJUVVVFS0tLNDc3x1NPPXWux50Q5bxjIuzRf2ePnp/sUXs0r6RROmvWrNi3b1/pdX9/fxw7dixmzpxZOmtoaIj+/v44evRo6Wzv3r3n/f9zish3v4iIV155Jdra2qKlpSXuv//+uOCC5H8pQi557vfkk0/Gnj174sorr4wrrrgiXn311bj++uvj8ccfTzDx2OW548yZM2PSpEnx+uuvl85OnDgRWRl8s7Q89+vr64t//OMfw37cm970prjwwqSfaBk35bxjIuzRf7FHz1/2qD2aV9L/1S5atCgOHToUW7ZsiaGhoejs7IyWlpZhf8airq4ulixZEg8++GAMDQ3FY489Fv39/XHFFVcknDyfPPcbHByM66+/Ptra2uJb3/pWwmnHLs/97rnnnti9e3fs3Lkzdu7cGTNmzIiurq5obW1NOHl+ee44derUWLp0aXR2dsbg4GDs2rUrtm/fHldddVXCyfPJc793v/vdcdFFF8WPf/zjOHXqVGzfvj127NgRS5cuTTj5+CnnHRNhj0bYo+c7e9Qeze0svyDrrP3+97/PPv3pT2dNTU3Zl770peyvf/1rdvDgwaypqSk7ePBglmVZdvjw4ezaa6/N5s+fn7W2tmZ/+MMfEk+d3xvd74knnsje8573ZE1NTcP+eeKJJ1KPnkueX79/V25fNZpl+e7Y39+ffeMb38g+8IEPZFdddVXZ/PplWb77/fGPf8xWrFiRzZ8/P1u+fHn2m9/8JvHUY/fggw+Wvmq0knZMltmj9uj5zx61R/OYlGVl8HvjAABUtPL4QzcAAFQ0UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMn9HwU/ZlxKGdtiAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 800x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "tnn.model_pipeline(config)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Run models with different ratio of training data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.01      , 0.01668101, 0.02782559, 0.04641589, 0.07742637,\n",
       "       0.12915497, 0.21544347, 0.35938137, 0.59948425, 1.        ])"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Set the range of ratio of training data used here\n",
    "myrange = np.logspace(-2, 0,num=10)\n",
    "myrange"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "wandb version 0.16.0 is available!  To upgrade, please run:\n",
       " $ pip install wandb --upgrade"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Tracking run with wandb version 0.15.12"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Run data is saved locally in <code>/Users/yuxi/workspace/nnn_paper/wandb/run-20231202_124535-98qyivsu</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Syncing run <strong><a href='https://wandb.ai/keyuxi/NN/runs/98qyivsu' target=\"_blank\">soft-sun-87</a></strong> to <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View project at <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">https://wandb.ai/keyuxi/NN</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run at <a href='https://wandb.ai/keyuxi/NN/runs/98qyivsu' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/98qyivsu</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "feats (27732, 178) Index(['bulge_size#1', 'bulge_size#2', 'hairpin_loop_mid#A'], dtype='object')\n",
      "fixed_feature_names: ['bulge_size#1', 'bulge_size#2', 'hairpin_size#3', 'hairpin_size#4', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "                           dH\n",
      "bulge_size#1              0.0\n",
      "bulge_size#2              0.0\n",
      "bulge_size#3              0.0\n",
      "bulge_size#4              0.0\n",
      "bulge_size#5              0.0\n",
      "...                       ...\n",
      "hairpin_tetraloop#CCTTGG  0.0\n",
      "hairpin_tetraloop#GCTTGC  0.0\n",
      "hairpin_tetraloop#GCTTGT  0.0\n",
      "hairpin_tetraloop#TCTTGA  0.0\n",
      "hairpin_tetraloop#TCTTGG  0.0\n",
      "\n",
      "[352 rows x 1 columns]\n",
      "['bulge_size#1', 'bulge_size#2', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#3', 'hairpin_size#4', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "Warning: Rank of matrix A_unknown, 118, is smaller than the number of features 152!\n",
      "                           dG\n",
      "bulge_size#1              4.0\n",
      "bulge_size#2              2.9\n",
      "bulge_size#3              3.1\n",
      "bulge_size#4              3.2\n",
      "bulge_size#5              3.3\n",
      "...                       ...\n",
      "hairpin_tetraloop#CCTTGG -0.8\n",
      "hairpin_tetraloop#GCTTGC -0.8\n",
      "hairpin_tetraloop#GCTTGT -0.5\n",
      "hairpin_tetraloop#TCTTGA -0.8\n",
      "hairpin_tetraloop#TCTTGG -0.5\n",
      "\n",
      "[352 rows x 1 columns]\n",
      "['bulge_size#1', 'bulge_size#2', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#3', 'hairpin_size#4', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "Warning: Rank of matrix A_unknown, 106, is smaller than the number of features 152!\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -3.684\n",
      "\tMean value of fitted new_dict is 0.927\n",
      "\tMean value of centered new_dict is -3.684\n",
      "\n",
      "group hairpin_size\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -2.913\n",
      "\tMean value of fitted new_dict is -2.392\n",
      "\tMean value of centered new_dict is -2.913\n",
      "\n",
      "group interior_size\n",
      "group stack\n",
      "\tMean value of old_dict is -8.275\n",
      "\tMean value of fitted new_dict is -8.275\n",
      "\tMean value of centered new_dict is -8.275\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 1.100\n",
      "\tMean value of fitted new_dict is -0.084\n",
      "\tMean value of centered new_dict is 1.100\n",
      "\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -0.715\n",
      "\tMean value of fitted new_dict is -0.039\n",
      "\tMean value of centered new_dict is -0.715\n",
      "\n",
      "group hairpin_size\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -0.640\n",
      "\tMean value of fitted new_dict is -0.512\n",
      "\tMean value of centered new_dict is -0.640\n",
      "\n",
      "group interior_size\n",
      "group stack\n",
      "\tMean value of old_dict is -1.406\n",
      "\tMean value of fitted new_dict is -1.406\n",
      "\tMean value of centered new_dict is -1.406\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 0.025\n",
      "\tMean value of fitted new_dict is 0.073\n",
      "\tMean value of centered new_dict is 0.025\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<style>\n",
       "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
       "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
       "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
       "    </style>\n",
       "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>▁</td></tr><tr><td>n_fixed_feat</td><td>▁</td></tr><tr><td>train_dG_mae</td><td>▁</td></tr><tr><td>train_dG_rsqr</td><td>▁</td></tr><tr><td>train_dH_mae</td><td>▁</td></tr><tr><td>train_dH_rsqr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-all</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-ov</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-all</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-ov</td><td>▁</td></tr><tr><td>val_Tm_bias-all</td><td>▁</td></tr><tr><td>val_Tm_bias-arr</td><td>▁</td></tr><tr><td>val_Tm_bias-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_bias-ov</td><td>▁</td></tr><tr><td>val_Tm_corr-all</td><td>▁</td></tr><tr><td>val_Tm_corr-arr</td><td>▁</td></tr><tr><td>val_Tm_corr-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_corr-ov</td><td>▁</td></tr><tr><td>val_Tm_mae-all</td><td>▁</td></tr><tr><td>val_Tm_mae-arr</td><td>▁</td></tr><tr><td>val_Tm_mae-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_mae-ov</td><td>▁</td></tr><tr><td>val_Tm_rmse-all</td><td>▁</td></tr><tr><td>val_Tm_rmse-arr</td><td>▁</td></tr><tr><td>val_Tm_rmse-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_rmse-ov</td><td>▁</td></tr><tr><td>val_dG_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_dG_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_dG_bias-arr</td><td>▁</td></tr><tr><td>val_dG_corr-arr</td><td>▁</td></tr><tr><td>val_dG_mae-arr</td><td>▁</td></tr><tr><td>val_dG_rmse-arr</td><td>▁</td></tr><tr><td>val_dH_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_dH_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_dH_bias-arr</td><td>▁</td></tr><tr><td>val_dH_corr-arr</td><td>▁</td></tr><tr><td>val_dH_mae-arr</td><td>▁</td></tr><tr><td>val_dH_rmse-arr</td><td>▁</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>178</td></tr><tr><td>n_fixed_feat</td><td>26</td></tr><tr><td>param_set_file</td><td>./models/soft-sun-87...</td></tr><tr><td>train_dG_mae</td><td>0.53728</td></tr><tr><td>train_dG_rsqr</td><td>0.6701</td></tr><tr><td>train_dH_mae</td><td>3.99298</td></tr><tr><td>train_dH_rsqr</td><td>0.65522</td></tr><tr><td>val_Tm_adjusted_mae-all</td><td>17.57329</td></tr><tr><td>val_Tm_adjusted_mae-arr</td><td>27.07295</td></tr><tr><td>val_Tm_adjusted_mae-lit_uv</td><td>1.771</td></tr><tr><td>val_Tm_adjusted_mae-ov</td><td>6.06305</td></tr><tr><td>val_Tm_adjusted_rmse-all</td><td>32.43112</td></tr><tr><td>val_Tm_adjusted_rmse-arr</td><td>44.475</td></tr><tr><td>val_Tm_adjusted_rmse-lit_uv</td><td>2.24155</td></tr><tr><td>val_Tm_adjusted_rmse-ov</td><td>7.62737</td></tr><tr><td>val_Tm_bias-all</td><td>12.74044</td></tr><tr><td>val_Tm_bias-arr</td><td>21.98695</td></tr><tr><td>val_Tm_bias-lit_uv</td><td>6.90205</td></tr><tr><td>val_Tm_bias-ov</td><td>3.93939</td></tr><tr><td>val_Tm_corr-all</td><td>0.15973</td></tr><tr><td>val_Tm_corr-arr</td><td>0.16591</td></tr><tr><td>val_Tm_corr-lit_uv</td><td>0.98578</td></tr><tr><td>val_Tm_corr-ov</td><td>0.37089</td></tr><tr><td>val_Tm_mae-all</td><td>18.36993</td></tr><tr><td>val_Tm_mae-arr</td><td>31.11119</td></tr><tr><td>val_Tm_mae-lit_uv</td><td>6.90205</td></tr><tr><td>val_Tm_mae-ov</td><td>6.67158</td></tr><tr><td>val_Tm_rmse-all</td><td>34.84389</td></tr><tr><td>val_Tm_rmse-arr</td><td>49.61302</td></tr><tr><td>val_Tm_rmse-lit_uv</td><td>7.25692</td></tr><tr><td>val_Tm_rmse-ov</td><td>8.58461</td></tr><tr><td>val_dG_adjusted_mae-arr</td><td>0.90686</td></tr><tr><td>val_dG_adjusted_rmse-arr</td><td>1.2661</td></tr><tr><td>val_dG_bias-arr</td><td>-0.87765</td></tr><tr><td>val_dG_corr-arr</td><td>0.6276</td></tr><tr><td>val_dG_mae-arr</td><td>1.21375</td></tr><tr><td>val_dG_rmse-arr</td><td>1.54055</td></tr><tr><td>val_dH_adjusted_mae-arr</td><td>10.45547</td></tr><tr><td>val_dH_adjusted_rmse-arr</td><td>15.72871</td></tr><tr><td>val_dH_bias-arr</td><td>0.72176</td></tr><tr><td>val_dH_corr-arr</td><td>0.40826</td></tr><tr><td>val_dH_mae-arr</td><td>10.58504</td></tr><tr><td>val_dH_rmse-arr</td><td>15.74526</td></tr></table><br/></div></div>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run <strong style=\"color:#cdcd00\">soft-sun-87</strong> at: <a href='https://wandb.ai/keyuxi/NN/runs/98qyivsu' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/98qyivsu</a><br/> View job at <a href='https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v5' target=\"_blank\">https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v5</a><br/>Synced 6 W&B file(s), 0 media file(s), 2 artifact file(s) and 0 other file(s)"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Find logs at: <code>./wandb/run-20231202_124535-98qyivsu/logs</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "wandb version 0.16.0 is available!  To upgrade, please run:\n",
       " $ pip install wandb --upgrade"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Tracking run with wandb version 0.15.12"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Run data is saved locally in <code>/Users/yuxi/workspace/nnn_paper/wandb/run-20231202_125435-kb2eewci</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Syncing run <strong><a href='https://wandb.ai/keyuxi/NN/runs/kb2eewci' target=\"_blank\">efficient-vortex-88</a></strong> to <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View project at <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">https://wandb.ai/keyuxi/NN</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run at <a href='https://wandb.ai/keyuxi/NN/runs/kb2eewci' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/kb2eewci</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "feats (27732, 178) Index(['bulge_size#1', 'bulge_size#2', 'hairpin_loop_mid#A'], dtype='object')\n",
      "fixed_feature_names: ['bulge_size#1', 'bulge_size#2', 'hairpin_size#3', 'hairpin_size#4', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "                           dH\n",
      "bulge_size#1              0.0\n",
      "bulge_size#2              0.0\n",
      "bulge_size#3              0.0\n",
      "bulge_size#4              0.0\n",
      "bulge_size#5              0.0\n",
      "...                       ...\n",
      "hairpin_tetraloop#CCTTGG  0.0\n",
      "hairpin_tetraloop#GCTTGC  0.0\n",
      "hairpin_tetraloop#GCTTGT  0.0\n",
      "hairpin_tetraloop#TCTTGA  0.0\n",
      "hairpin_tetraloop#TCTTGG  0.0\n",
      "\n",
      "[352 rows x 1 columns]\n",
      "['bulge_size#1', 'bulge_size#2', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#3', 'hairpin_size#4', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "Warning: Rank of matrix A_unknown, 130, is smaller than the number of features 152!\n",
      "                           dG\n",
      "bulge_size#1              4.0\n",
      "bulge_size#2              2.9\n",
      "bulge_size#3              3.1\n",
      "bulge_size#4              3.2\n",
      "bulge_size#5              3.3\n",
      "...                       ...\n",
      "hairpin_tetraloop#CCTTGG -0.8\n",
      "hairpin_tetraloop#GCTTGC -0.8\n",
      "hairpin_tetraloop#GCTTGT -0.5\n",
      "hairpin_tetraloop#TCTTGA -0.8\n",
      "hairpin_tetraloop#TCTTGG -0.5\n",
      "\n",
      "[352 rows x 1 columns]\n",
      "['bulge_size#1', 'bulge_size#2', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#3', 'hairpin_size#4', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "Warning: Rank of matrix A_unknown, 134, is smaller than the number of features 152!\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -3.684\n",
      "\tMean value of fitted new_dict is 0.760\n",
      "\tMean value of centered new_dict is -3.684\n",
      "\n",
      "group hairpin_size\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -2.913\n",
      "\tMean value of fitted new_dict is -6.110\n",
      "\tMean value of centered new_dict is -2.913\n",
      "\n",
      "group interior_size\n",
      "group stack\n",
      "\tMean value of old_dict is -8.275\n",
      "\tMean value of fitted new_dict is -8.275\n",
      "\tMean value of centered new_dict is -8.275\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 1.100\n",
      "\tMean value of fitted new_dict is -1.215\n",
      "\tMean value of centered new_dict is 1.100\n",
      "\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -0.715\n",
      "\tMean value of fitted new_dict is -0.051\n",
      "\tMean value of centered new_dict is -0.715\n",
      "\n",
      "group hairpin_size\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -0.640\n",
      "\tMean value of fitted new_dict is -0.950\n",
      "\tMean value of centered new_dict is -0.640\n",
      "\n",
      "group interior_size\n",
      "group stack\n",
      "\tMean value of old_dict is -1.406\n",
      "\tMean value of fitted new_dict is -1.406\n",
      "\tMean value of centered new_dict is -1.406\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 0.025\n",
      "\tMean value of fitted new_dict is 0.235\n",
      "\tMean value of centered new_dict is 0.025\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<style>\n",
       "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
       "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
       "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
       "    </style>\n",
       "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>▁</td></tr><tr><td>n_fixed_feat</td><td>▁</td></tr><tr><td>train_dG_mae</td><td>▁</td></tr><tr><td>train_dG_rsqr</td><td>▁</td></tr><tr><td>train_dH_mae</td><td>▁</td></tr><tr><td>train_dH_rsqr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-all</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-ov</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-all</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-ov</td><td>▁</td></tr><tr><td>val_Tm_bias-all</td><td>▁</td></tr><tr><td>val_Tm_bias-arr</td><td>▁</td></tr><tr><td>val_Tm_bias-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_bias-ov</td><td>▁</td></tr><tr><td>val_Tm_corr-all</td><td>▁</td></tr><tr><td>val_Tm_corr-arr</td><td>▁</td></tr><tr><td>val_Tm_corr-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_corr-ov</td><td>▁</td></tr><tr><td>val_Tm_mae-all</td><td>▁</td></tr><tr><td>val_Tm_mae-arr</td><td>▁</td></tr><tr><td>val_Tm_mae-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_mae-ov</td><td>▁</td></tr><tr><td>val_Tm_rmse-all</td><td>▁</td></tr><tr><td>val_Tm_rmse-arr</td><td>▁</td></tr><tr><td>val_Tm_rmse-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_rmse-ov</td><td>▁</td></tr><tr><td>val_dG_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_dG_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_dG_bias-arr</td><td>▁</td></tr><tr><td>val_dG_corr-arr</td><td>▁</td></tr><tr><td>val_dG_mae-arr</td><td>▁</td></tr><tr><td>val_dG_rmse-arr</td><td>▁</td></tr><tr><td>val_dH_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_dH_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_dH_bias-arr</td><td>▁</td></tr><tr><td>val_dH_corr-arr</td><td>▁</td></tr><tr><td>val_dH_mae-arr</td><td>▁</td></tr><tr><td>val_dH_rmse-arr</td><td>▁</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>178</td></tr><tr><td>n_fixed_feat</td><td>26</td></tr><tr><td>param_set_file</td><td>./models/efficient-v...</td></tr><tr><td>train_dG_mae</td><td>0.51297</td></tr><tr><td>train_dG_rsqr</td><td>0.72104</td></tr><tr><td>train_dH_mae</td><td>5.00271</td></tr><tr><td>train_dH_rsqr</td><td>0.62842</td></tr><tr><td>val_Tm_adjusted_mae-all</td><td>25.51554</td></tr><tr><td>val_Tm_adjusted_mae-arr</td><td>39.51638</td></tr><tr><td>val_Tm_adjusted_mae-lit_uv</td><td>2.21476</td></tr><tr><td>val_Tm_adjusted_mae-ov</td><td>3.82612</td></tr><tr><td>val_Tm_adjusted_rmse-all</td><td>44.86251</td></tr><tr><td>val_Tm_adjusted_rmse-arr</td><td>60.88451</td></tr><tr><td>val_Tm_adjusted_rmse-lit_uv</td><td>2.73256</td></tr><tr><td>val_Tm_adjusted_rmse-ov</td><td>4.81984</td></tr><tr><td>val_Tm_bias-all</td><td>19.08471</td></tr><tr><td>val_Tm_bias-arr</td><td>35.38288</td></tr><tr><td>val_Tm_bias-lit_uv</td><td>6.94159</td></tr><tr><td>val_Tm_bias-ov</td><td>4.01525</td></tr><tr><td>val_Tm_corr-all</td><td>0.08955</td></tr><tr><td>val_Tm_corr-arr</td><td>0.14999</td></tr><tr><td>val_Tm_corr-lit_uv</td><td>0.98069</td></tr><tr><td>val_Tm_corr-ov</td><td>0.5313</td></tr><tr><td>val_Tm_mae-all</td><td>24.11516</td></tr><tr><td>val_Tm_mae-arr</td><td>44.88723</td></tr><tr><td>val_Tm_mae-lit_uv</td><td>6.94159</td></tr><tr><td>val_Tm_mae-ov</td><td>5.12184</td></tr><tr><td>val_Tm_rmse-all</td><td>48.75317</td></tr><tr><td>val_Tm_rmse-arr</td><td>70.41926</td></tr><tr><td>val_Tm_rmse-lit_uv</td><td>7.46007</td></tr><tr><td>val_Tm_rmse-ov</td><td>6.27321</td></tr><tr><td>val_dG_adjusted_mae-arr</td><td>0.83281</td></tr><tr><td>val_dG_adjusted_rmse-arr</td><td>1.10149</td></tr><tr><td>val_dG_bias-arr</td><td>-0.96835</td></tr><tr><td>val_dG_corr-arr</td><td>0.64136</td></tr><tr><td>val_dG_mae-arr</td><td>1.2065</td></tr><tr><td>val_dG_rmse-arr</td><td>1.46663</td></tr><tr><td>val_dH_adjusted_mae-arr</td><td>8.45906</td></tr><tr><td>val_dH_adjusted_rmse-arr</td><td>11.06465</td></tr><tr><td>val_dH_bias-arr</td><td>5.29384</td></tr><tr><td>val_dH_corr-arr</td><td>0.60513</td></tr><tr><td>val_dH_mae-arr</td><td>9.65824</td></tr><tr><td>val_dH_rmse-arr</td><td>12.26585</td></tr></table><br/></div></div>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run <strong style=\"color:#cdcd00\">efficient-vortex-88</strong> at: <a href='https://wandb.ai/keyuxi/NN/runs/kb2eewci' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/kb2eewci</a><br/> View job at <a href='https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v5' target=\"_blank\">https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v5</a><br/>Synced 6 W&B file(s), 0 media file(s), 0 artifact file(s) and 0 other file(s)"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Find logs at: <code>./wandb/run-20231202_125435-kb2eewci/logs</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "wandb version 0.16.0 is available!  To upgrade, please run:\n",
       " $ pip install wandb --upgrade"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Tracking run with wandb version 0.15.12"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Run data is saved locally in <code>/Users/yuxi/workspace/nnn_paper/wandb/run-20231202_130321-ambort9u</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Syncing run <strong><a href='https://wandb.ai/keyuxi/NN/runs/ambort9u' target=\"_blank\">vague-plasma-89</a></strong> to <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View project at <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">https://wandb.ai/keyuxi/NN</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run at <a href='https://wandb.ai/keyuxi/NN/runs/ambort9u' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/ambort9u</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "feats (27732, 178) Index(['bulge_size#1', 'bulge_size#2', 'hairpin_loop_mid#A'], dtype='object')\n",
      "fixed_feature_names: ['bulge_size#1', 'bulge_size#2', 'hairpin_size#3', 'hairpin_size#4', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "                           dH\n",
      "bulge_size#1              0.0\n",
      "bulge_size#2              0.0\n",
      "bulge_size#3              0.0\n",
      "bulge_size#4              0.0\n",
      "bulge_size#5              0.0\n",
      "...                       ...\n",
      "hairpin_tetraloop#CCTTGG  0.0\n",
      "hairpin_tetraloop#GCTTGC  0.0\n",
      "hairpin_tetraloop#GCTTGT  0.0\n",
      "hairpin_tetraloop#TCTTGA  0.0\n",
      "hairpin_tetraloop#TCTTGG  0.0\n",
      "\n",
      "[352 rows x 1 columns]\n",
      "['bulge_size#1', 'bulge_size#2', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#3', 'hairpin_size#4', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "Warning: Rank of matrix A_unknown, 149, is smaller than the number of features 152!\n",
      "                           dG\n",
      "bulge_size#1              4.0\n",
      "bulge_size#2              2.9\n",
      "bulge_size#3              3.1\n",
      "bulge_size#4              3.2\n",
      "bulge_size#5              3.3\n",
      "...                       ...\n",
      "hairpin_tetraloop#CCTTGG -0.8\n",
      "hairpin_tetraloop#GCTTGC -0.8\n",
      "hairpin_tetraloop#GCTTGT -0.5\n",
      "hairpin_tetraloop#TCTTGA -0.8\n",
      "hairpin_tetraloop#TCTTGG -0.5\n",
      "\n",
      "[352 rows x 1 columns]\n",
      "['bulge_size#1', 'bulge_size#2', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#3', 'hairpin_size#4', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "Warning: Rank of matrix A_unknown, 138, is smaller than the number of features 152!\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -3.684\n",
      "\tMean value of fitted new_dict is 1.627\n",
      "\tMean value of centered new_dict is -3.684\n",
      "\n",
      "group hairpin_size\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -2.913\n",
      "\tMean value of fitted new_dict is -7.374\n",
      "\tMean value of centered new_dict is -2.913\n",
      "\n",
      "group interior_size\n",
      "group stack\n",
      "\tMean value of old_dict is -8.275\n",
      "\tMean value of fitted new_dict is -8.275\n",
      "\tMean value of centered new_dict is -8.275\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 1.100\n",
      "\tMean value of fitted new_dict is -2.184\n",
      "\tMean value of centered new_dict is 1.100\n",
      "\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -0.715\n",
      "\tMean value of fitted new_dict is -0.002\n",
      "\tMean value of centered new_dict is -0.715\n",
      "\n",
      "group hairpin_size\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -0.640\n",
      "\tMean value of fitted new_dict is -1.066\n",
      "\tMean value of centered new_dict is -0.640\n",
      "\n",
      "group interior_size\n",
      "group stack\n",
      "\tMean value of old_dict is -1.406\n",
      "\tMean value of fitted new_dict is -1.406\n",
      "\tMean value of centered new_dict is -1.406\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 0.025\n",
      "\tMean value of fitted new_dict is 0.139\n",
      "\tMean value of centered new_dict is 0.025\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<style>\n",
       "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
       "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
       "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
       "    </style>\n",
       "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>▁</td></tr><tr><td>n_fixed_feat</td><td>▁</td></tr><tr><td>train_dG_mae</td><td>▁</td></tr><tr><td>train_dG_rsqr</td><td>▁</td></tr><tr><td>train_dH_mae</td><td>▁</td></tr><tr><td>train_dH_rsqr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-all</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-ov</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-all</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-ov</td><td>▁</td></tr><tr><td>val_Tm_bias-all</td><td>▁</td></tr><tr><td>val_Tm_bias-arr</td><td>▁</td></tr><tr><td>val_Tm_bias-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_bias-ov</td><td>▁</td></tr><tr><td>val_Tm_corr-all</td><td>▁</td></tr><tr><td>val_Tm_corr-arr</td><td>▁</td></tr><tr><td>val_Tm_corr-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_corr-ov</td><td>▁</td></tr><tr><td>val_Tm_mae-all</td><td>▁</td></tr><tr><td>val_Tm_mae-arr</td><td>▁</td></tr><tr><td>val_Tm_mae-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_mae-ov</td><td>▁</td></tr><tr><td>val_Tm_rmse-all</td><td>▁</td></tr><tr><td>val_Tm_rmse-arr</td><td>▁</td></tr><tr><td>val_Tm_rmse-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_rmse-ov</td><td>▁</td></tr><tr><td>val_dG_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_dG_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_dG_bias-arr</td><td>▁</td></tr><tr><td>val_dG_corr-arr</td><td>▁</td></tr><tr><td>val_dG_mae-arr</td><td>▁</td></tr><tr><td>val_dG_rmse-arr</td><td>▁</td></tr><tr><td>val_dH_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_dH_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_dH_bias-arr</td><td>▁</td></tr><tr><td>val_dH_corr-arr</td><td>▁</td></tr><tr><td>val_dH_mae-arr</td><td>▁</td></tr><tr><td>val_dH_rmse-arr</td><td>▁</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>178</td></tr><tr><td>n_fixed_feat</td><td>26</td></tr><tr><td>param_set_file</td><td>./models/vague-plasm...</td></tr><tr><td>train_dG_mae</td><td>0.58164</td></tr><tr><td>train_dG_rsqr</td><td>0.6573</td></tr><tr><td>train_dH_mae</td><td>5.68115</td></tr><tr><td>train_dH_rsqr</td><td>0.4848</td></tr><tr><td>val_Tm_adjusted_mae-all</td><td>29.61847</td></tr><tr><td>val_Tm_adjusted_mae-arr</td><td>49.74894</td></tr><tr><td>val_Tm_adjusted_mae-lit_uv</td><td>2.41035</td></tr><tr><td>val_Tm_adjusted_mae-ov</td><td>3.45002</td></tr><tr><td>val_Tm_adjusted_rmse-all</td><td>61.35697</td></tr><tr><td>val_Tm_adjusted_rmse-arr</td><td>87.62309</td></tr><tr><td>val_Tm_adjusted_rmse-lit_uv</td><td>2.99367</td></tr><tr><td>val_Tm_adjusted_rmse-ov</td><td>4.52749</td></tr><tr><td>val_Tm_bias-all</td><td>18.11417</td></tr><tr><td>val_Tm_bias-arr</td><td>34.6375</td></tr><tr><td>val_Tm_bias-lit_uv</td><td>6.92344</td></tr><tr><td>val_Tm_bias-ov</td><td>3.7556</td></tr><tr><td>val_Tm_corr-all</td><td>0.08383</td></tr><tr><td>val_Tm_corr-arr</td><td>0.12619</td></tr><tr><td>val_Tm_corr-lit_uv</td><td>0.97713</td></tr><tr><td>val_Tm_corr-ov</td><td>0.58566</td></tr><tr><td>val_Tm_mae-all</td><td>28.21475</td></tr><tr><td>val_Tm_mae-arr</td><td>55.66206</td></tr><tr><td>val_Tm_mae-lit_uv</td><td>6.92858</td></tr><tr><td>val_Tm_mae-ov</td><td>4.70147</td></tr><tr><td>val_Tm_rmse-all</td><td>63.975</td></tr><tr><td>val_Tm_rmse-arr</td><td>94.22082</td></tr><tr><td>val_Tm_rmse-lit_uv</td><td>7.54295</td></tr><tr><td>val_Tm_rmse-ov</td><td>5.88241</td></tr><tr><td>val_dG_adjusted_mae-arr</td><td>0.77942</td></tr><tr><td>val_dG_adjusted_rmse-arr</td><td>1.03189</td></tr><tr><td>val_dG_bias-arr</td><td>-0.92242</td></tr><tr><td>val_dG_corr-arr</td><td>0.71058</td></tr><tr><td>val_dG_mae-arr</td><td>1.14644</td></tr><tr><td>val_dG_rmse-arr</td><td>1.38407</td></tr><tr><td>val_dH_adjusted_mae-arr</td><td>7.823</td></tr><tr><td>val_dH_adjusted_rmse-arr</td><td>9.97299</td></tr><tr><td>val_dH_bias-arr</td><td>8.77898</td></tr><tr><td>val_dH_corr-arr</td><td>0.63506</td></tr><tr><td>val_dH_mae-arr</td><td>10.65668</td></tr><tr><td>val_dH_rmse-arr</td><td>13.2865</td></tr></table><br/></div></div>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run <strong style=\"color:#cdcd00\">vague-plasma-89</strong> at: <a href='https://wandb.ai/keyuxi/NN/runs/ambort9u' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/ambort9u</a><br/> View job at <a href='https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v5' target=\"_blank\">https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v5</a><br/>Synced 6 W&B file(s), 0 media file(s), 0 artifact file(s) and 0 other file(s)"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Find logs at: <code>./wandb/run-20231202_130321-ambort9u/logs</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "wandb version 0.16.0 is available!  To upgrade, please run:\n",
       " $ pip install wandb --upgrade"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Tracking run with wandb version 0.15.12"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Run data is saved locally in <code>/Users/yuxi/workspace/nnn_paper/wandb/run-20231202_131204-qajcaia8</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Syncing run <strong><a href='https://wandb.ai/keyuxi/NN/runs/qajcaia8' target=\"_blank\">ruby-oath-90</a></strong> to <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View project at <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">https://wandb.ai/keyuxi/NN</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run at <a href='https://wandb.ai/keyuxi/NN/runs/qajcaia8' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/qajcaia8</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "feats (27732, 178) Index(['bulge_size#1', 'bulge_size#2', 'hairpin_loop_mid#A'], dtype='object')\n",
      "fixed_feature_names: ['bulge_size#1', 'bulge_size#2', 'hairpin_size#3', 'hairpin_size#4', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "                           dH\n",
      "bulge_size#1              0.0\n",
      "bulge_size#2              0.0\n",
      "bulge_size#3              0.0\n",
      "bulge_size#4              0.0\n",
      "bulge_size#5              0.0\n",
      "...                       ...\n",
      "hairpin_tetraloop#CCTTGG  0.0\n",
      "hairpin_tetraloop#GCTTGC  0.0\n",
      "hairpin_tetraloop#GCTTGT  0.0\n",
      "hairpin_tetraloop#TCTTGA  0.0\n",
      "hairpin_tetraloop#TCTTGG  0.0\n",
      "\n",
      "[352 rows x 1 columns]\n",
      "['bulge_size#1', 'bulge_size#2', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#3', 'hairpin_size#4', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "Warning: Rank of matrix A_unknown, 149, is smaller than the number of features 152!\n",
      "                           dG\n",
      "bulge_size#1              4.0\n",
      "bulge_size#2              2.9\n",
      "bulge_size#3              3.1\n",
      "bulge_size#4              3.2\n",
      "bulge_size#5              3.3\n",
      "...                       ...\n",
      "hairpin_tetraloop#CCTTGG -0.8\n",
      "hairpin_tetraloop#GCTTGC -0.8\n",
      "hairpin_tetraloop#GCTTGT -0.5\n",
      "hairpin_tetraloop#TCTTGA -0.8\n",
      "hairpin_tetraloop#TCTTGG -0.5\n",
      "\n",
      "[352 rows x 1 columns]\n",
      "['bulge_size#1', 'bulge_size#2', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#3', 'hairpin_size#4', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "Warning: Rank of matrix A_unknown, 151, is smaller than the number of features 152!\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -3.684\n",
      "\tMean value of fitted new_dict is 1.597\n",
      "\tMean value of centered new_dict is -3.684\n",
      "\n",
      "group hairpin_size\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -2.913\n",
      "\tMean value of fitted new_dict is -5.392\n",
      "\tMean value of centered new_dict is -2.912\n",
      "\n",
      "group interior_size\n",
      "group stack\n",
      "\tMean value of old_dict is -8.275\n",
      "\tMean value of fitted new_dict is -8.275\n",
      "\tMean value of centered new_dict is -8.275\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 1.100\n",
      "\tMean value of fitted new_dict is -1.365\n",
      "\tMean value of centered new_dict is 1.100\n",
      "\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -0.715\n",
      "\tMean value of fitted new_dict is -3.112\n",
      "\tMean value of centered new_dict is -0.715\n",
      "\n",
      "group hairpin_size\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -0.640\n",
      "\tMean value of fitted new_dict is -0.712\n",
      "\tMean value of centered new_dict is -0.640\n",
      "\n",
      "group interior_size\n",
      "group stack\n",
      "\tMean value of old_dict is -1.406\n",
      "\tMean value of fitted new_dict is -1.406\n",
      "\tMean value of centered new_dict is -1.406\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 0.025\n",
      "\tMean value of fitted new_dict is 0.075\n",
      "\tMean value of centered new_dict is 0.025\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<style>\n",
       "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
       "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
       "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
       "    </style>\n",
       "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>▁</td></tr><tr><td>n_fixed_feat</td><td>▁</td></tr><tr><td>train_dG_mae</td><td>▁</td></tr><tr><td>train_dG_rsqr</td><td>▁</td></tr><tr><td>train_dH_mae</td><td>▁</td></tr><tr><td>train_dH_rsqr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-all</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-ov</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-all</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-ov</td><td>▁</td></tr><tr><td>val_Tm_bias-all</td><td>▁</td></tr><tr><td>val_Tm_bias-arr</td><td>▁</td></tr><tr><td>val_Tm_bias-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_bias-ov</td><td>▁</td></tr><tr><td>val_Tm_corr-all</td><td>▁</td></tr><tr><td>val_Tm_corr-arr</td><td>▁</td></tr><tr><td>val_Tm_corr-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_corr-ov</td><td>▁</td></tr><tr><td>val_Tm_mae-all</td><td>▁</td></tr><tr><td>val_Tm_mae-arr</td><td>▁</td></tr><tr><td>val_Tm_mae-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_mae-ov</td><td>▁</td></tr><tr><td>val_Tm_rmse-all</td><td>▁</td></tr><tr><td>val_Tm_rmse-arr</td><td>▁</td></tr><tr><td>val_Tm_rmse-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_rmse-ov</td><td>▁</td></tr><tr><td>val_dG_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_dG_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_dG_bias-arr</td><td>▁</td></tr><tr><td>val_dG_corr-arr</td><td>▁</td></tr><tr><td>val_dG_mae-arr</td><td>▁</td></tr><tr><td>val_dG_rmse-arr</td><td>▁</td></tr><tr><td>val_dH_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_dH_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_dH_bias-arr</td><td>▁</td></tr><tr><td>val_dH_corr-arr</td><td>▁</td></tr><tr><td>val_dH_mae-arr</td><td>▁</td></tr><tr><td>val_dH_rmse-arr</td><td>▁</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>178</td></tr><tr><td>n_fixed_feat</td><td>26</td></tr><tr><td>param_set_file</td><td>./models/ruby-oath-9...</td></tr><tr><td>train_dG_mae</td><td>0.5936</td></tr><tr><td>train_dG_rsqr</td><td>0.62948</td></tr><tr><td>train_dH_mae</td><td>5.92639</td></tr><tr><td>train_dH_rsqr</td><td>0.45014</td></tr><tr><td>val_Tm_adjusted_mae-all</td><td>20.03418</td></tr><tr><td>val_Tm_adjusted_mae-arr</td><td>16.78739</td></tr><tr><td>val_Tm_adjusted_mae-lit_uv</td><td>2.28544</td></tr><tr><td>val_Tm_adjusted_mae-ov</td><td>2.0863</td></tr><tr><td>val_Tm_adjusted_rmse-all</td><td>26.76343</td></tr><tr><td>val_Tm_adjusted_rmse-arr</td><td>26.7711</td></tr><tr><td>val_Tm_adjusted_rmse-lit_uv</td><td>2.82208</td></tr><tr><td>val_Tm_adjusted_rmse-ov</td><td>2.67163</td></tr><tr><td>val_Tm_bias-all</td><td>-13.07373</td></tr><tr><td>val_Tm_bias-arr</td><td>-33.09966</td></tr><tr><td>val_Tm_bias-lit_uv</td><td>6.92196</td></tr><tr><td>val_Tm_bias-ov</td><td>5.1525</td></tr><tr><td>val_Tm_corr-all</td><td>0.6562</td></tr><tr><td>val_Tm_corr-arr</td><td>0.65295</td></tr><tr><td>val_Tm_corr-lit_uv</td><td>0.97953</td></tr><tr><td>val_Tm_corr-ov</td><td>0.80968</td></tr><tr><td>val_Tm_mae-all</td><td>18.9884</td></tr><tr><td>val_Tm_mae-arr</td><td>33.75058</td></tr><tr><td>val_Tm_mae-lit_uv</td><td>6.92315</td></tr><tr><td>val_Tm_mae-ov</td><td>5.21759</td></tr><tr><td>val_Tm_rmse-all</td><td>29.78596</td></tr><tr><td>val_Tm_rmse-arr</td><td>42.57087</td></tr><tr><td>val_Tm_rmse-lit_uv</td><td>7.47514</td></tr><tr><td>val_Tm_rmse-ov</td><td>5.80395</td></tr><tr><td>val_dG_adjusted_mae-arr</td><td>0.62564</td></tr><tr><td>val_dG_adjusted_rmse-arr</td><td>0.84211</td></tr><tr><td>val_dG_bias-arr</td><td>2.16941</td></tr><tr><td>val_dG_corr-arr</td><td>0.78059</td></tr><tr><td>val_dG_mae-arr</td><td>2.17547</td></tr><tr><td>val_dG_rmse-arr</td><td>2.32712</td></tr><tr><td>val_dH_adjusted_mae-arr</td><td>6.84006</td></tr><tr><td>val_dH_adjusted_rmse-arr</td><td>8.59049</td></tr><tr><td>val_dH_bias-arr</td><td>4.68728</td></tr><tr><td>val_dH_corr-arr</td><td>0.68113</td></tr><tr><td>val_dH_mae-arr</td><td>7.68641</td></tr><tr><td>val_dH_rmse-arr</td><td>9.78606</td></tr></table><br/></div></div>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run <strong style=\"color:#cdcd00\">ruby-oath-90</strong> at: <a href='https://wandb.ai/keyuxi/NN/runs/qajcaia8' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/qajcaia8</a><br/> View job at <a href='https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v5' target=\"_blank\">https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v5</a><br/>Synced 6 W&B file(s), 0 media file(s), 0 artifact file(s) and 0 other file(s)"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Find logs at: <code>./wandb/run-20231202_131204-qajcaia8/logs</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "wandb version 0.16.0 is available!  To upgrade, please run:\n",
       " $ pip install wandb --upgrade"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Tracking run with wandb version 0.15.12"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Run data is saved locally in <code>/Users/yuxi/workspace/nnn_paper/wandb/run-20231202_132047-3rz0fuwk</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Syncing run <strong><a href='https://wandb.ai/keyuxi/NN/runs/3rz0fuwk' target=\"_blank\">ruby-shape-91</a></strong> to <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View project at <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">https://wandb.ai/keyuxi/NN</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run at <a href='https://wandb.ai/keyuxi/NN/runs/3rz0fuwk' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/3rz0fuwk</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "feats (27732, 178) Index(['bulge_size#1', 'bulge_size#2', 'hairpin_loop_mid#A'], dtype='object')\n",
      "fixed_feature_names: ['bulge_size#1', 'bulge_size#2', 'hairpin_size#3', 'hairpin_size#4', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "                           dH\n",
      "bulge_size#1              0.0\n",
      "bulge_size#2              0.0\n",
      "bulge_size#3              0.0\n",
      "bulge_size#4              0.0\n",
      "bulge_size#5              0.0\n",
      "...                       ...\n",
      "hairpin_tetraloop#CCTTGG  0.0\n",
      "hairpin_tetraloop#GCTTGC  0.0\n",
      "hairpin_tetraloop#GCTTGT  0.0\n",
      "hairpin_tetraloop#TCTTGA  0.0\n",
      "hairpin_tetraloop#TCTTGG  0.0\n",
      "\n",
      "[352 rows x 1 columns]\n",
      "['bulge_size#1', 'bulge_size#2', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#3', 'hairpin_size#4', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "                           dG\n",
      "bulge_size#1              4.0\n",
      "bulge_size#2              2.9\n",
      "bulge_size#3              3.1\n",
      "bulge_size#4              3.2\n",
      "bulge_size#5              3.3\n",
      "...                       ...\n",
      "hairpin_tetraloop#CCTTGG -0.8\n",
      "hairpin_tetraloop#GCTTGC -0.8\n",
      "hairpin_tetraloop#GCTTGT -0.5\n",
      "hairpin_tetraloop#TCTTGA -0.8\n",
      "hairpin_tetraloop#TCTTGG -0.5\n",
      "\n",
      "[352 rows x 1 columns]\n",
      "['bulge_size#1', 'bulge_size#2', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#3', 'hairpin_size#4', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "Warning: Rank of matrix A_unknown, 151, is smaller than the number of features 152!\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -3.684\n",
      "\tMean value of fitted new_dict is -16.136\n",
      "\tMean value of centered new_dict is -3.684\n",
      "\n",
      "group hairpin_size\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -2.913\n",
      "\tMean value of fitted new_dict is -5.883\n",
      "\tMean value of centered new_dict is -2.912\n",
      "\n",
      "group interior_size\n",
      "group stack\n",
      "\tMean value of old_dict is -8.275\n",
      "\tMean value of fitted new_dict is -8.275\n",
      "\tMean value of centered new_dict is -8.275\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 1.100\n",
      "\tMean value of fitted new_dict is -0.426\n",
      "\tMean value of centered new_dict is 1.100\n",
      "\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -0.715\n",
      "\tMean value of fitted new_dict is -0.042\n",
      "\tMean value of centered new_dict is -0.715\n",
      "\n",
      "group hairpin_size\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -0.640\n",
      "\tMean value of fitted new_dict is -0.707\n",
      "\tMean value of centered new_dict is -0.640\n",
      "\n",
      "group interior_size\n",
      "group stack\n",
      "\tMean value of old_dict is -1.406\n",
      "\tMean value of fitted new_dict is -1.406\n",
      "\tMean value of centered new_dict is -1.406\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 0.025\n",
      "\tMean value of fitted new_dict is 0.253\n",
      "\tMean value of centered new_dict is 0.025\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<style>\n",
       "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
       "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
       "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
       "    </style>\n",
       "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>▁</td></tr><tr><td>n_fixed_feat</td><td>▁</td></tr><tr><td>train_dG_mae</td><td>▁</td></tr><tr><td>train_dG_rsqr</td><td>▁</td></tr><tr><td>train_dH_mae</td><td>▁</td></tr><tr><td>train_dH_rsqr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-all</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-ov</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-all</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-ov</td><td>▁</td></tr><tr><td>val_Tm_bias-all</td><td>▁</td></tr><tr><td>val_Tm_bias-arr</td><td>▁</td></tr><tr><td>val_Tm_bias-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_bias-ov</td><td>▁</td></tr><tr><td>val_Tm_corr-all</td><td>▁</td></tr><tr><td>val_Tm_corr-arr</td><td>▁</td></tr><tr><td>val_Tm_corr-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_corr-ov</td><td>▁</td></tr><tr><td>val_Tm_mae-all</td><td>▁</td></tr><tr><td>val_Tm_mae-arr</td><td>▁</td></tr><tr><td>val_Tm_mae-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_mae-ov</td><td>▁</td></tr><tr><td>val_Tm_rmse-all</td><td>▁</td></tr><tr><td>val_Tm_rmse-arr</td><td>▁</td></tr><tr><td>val_Tm_rmse-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_rmse-ov</td><td>▁</td></tr><tr><td>val_dG_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_dG_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_dG_bias-arr</td><td>▁</td></tr><tr><td>val_dG_corr-arr</td><td>▁</td></tr><tr><td>val_dG_mae-arr</td><td>▁</td></tr><tr><td>val_dG_rmse-arr</td><td>▁</td></tr><tr><td>val_dH_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_dH_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_dH_bias-arr</td><td>▁</td></tr><tr><td>val_dH_corr-arr</td><td>▁</td></tr><tr><td>val_dH_mae-arr</td><td>▁</td></tr><tr><td>val_dH_rmse-arr</td><td>▁</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>178</td></tr><tr><td>n_fixed_feat</td><td>26</td></tr><tr><td>param_set_file</td><td>./models/ruby-shape-...</td></tr><tr><td>train_dG_mae</td><td>0.623</td></tr><tr><td>train_dG_rsqr</td><td>0.5917</td></tr><tr><td>train_dH_mae</td><td>5.84149</td></tr><tr><td>train_dH_rsqr</td><td>0.49831</td></tr><tr><td>val_Tm_adjusted_mae-all</td><td>47.36485</td></tr><tr><td>val_Tm_adjusted_mae-arr</td><td>104.82778</td></tr><tr><td>val_Tm_adjusted_mae-lit_uv</td><td>2.08145</td></tr><tr><td>val_Tm_adjusted_mae-ov</td><td>2.08979</td></tr><tr><td>val_Tm_adjusted_rmse-all</td><td>88.44467</td></tr><tr><td>val_Tm_adjusted_rmse-arr</td><td>133.39181</td></tr><tr><td>val_Tm_adjusted_rmse-lit_uv</td><td>2.58342</td></tr><tr><td>val_Tm_adjusted_rmse-ov</td><td>2.73233</td></tr><tr><td>val_Tm_bias-all</td><td>5.48556</td></tr><tr><td>val_Tm_bias-arr</td><td>6.40793</td></tr><tr><td>val_Tm_bias-lit_uv</td><td>6.90919</td></tr><tr><td>val_Tm_bias-ov</td><td>4.4937</td></tr><tr><td>val_Tm_corr-all</td><td>0.20742</td></tr><tr><td>val_Tm_corr-arr</td><td>0.21982</td></tr><tr><td>val_Tm_corr-lit_uv</td><td>0.98218</td></tr><tr><td>val_Tm_corr-ov</td><td>0.80406</td></tr><tr><td>val_Tm_mae-all</td><td>49.09728</td></tr><tr><td>val_Tm_mae-arr</td><td>105.54777</td></tr><tr><td>val_Tm_mae-lit_uv</td><td>6.90919</td></tr><tr><td>val_Tm_mae-ov</td><td>4.60544</td></tr><tr><td>val_Tm_rmse-all</td><td>88.61463</td></tr><tr><td>val_Tm_rmse-arr</td><td>133.54564</td></tr><tr><td>val_Tm_rmse-lit_uv</td><td>7.37638</td></tr><tr><td>val_Tm_rmse-ov</td><td>5.25918</td></tr><tr><td>val_dG_adjusted_mae-arr</td><td>0.62092</td></tr><tr><td>val_dG_adjusted_rmse-arr</td><td>0.81814</td></tr><tr><td>val_dG_bias-arr</td><td>-1.39269</td></tr><tr><td>val_dG_corr-arr</td><td>0.78176</td></tr><tr><td>val_dG_mae-arr</td><td>1.44351</td></tr><tr><td>val_dG_rmse-arr</td><td>1.61522</td></tr><tr><td>val_dH_adjusted_mae-arr</td><td>6.54269</td></tr><tr><td>val_dH_adjusted_rmse-arr</td><td>8.37318</td></tr><tr><td>val_dH_bias-arr</td><td>21.53976</td></tr><tr><td>val_dH_corr-arr</td><td>0.68968</td></tr><tr><td>val_dH_mae-arr</td><td>21.56348</td></tr><tr><td>val_dH_rmse-arr</td><td>23.10999</td></tr></table><br/></div></div>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run <strong style=\"color:#cdcd00\">ruby-shape-91</strong> at: <a href='https://wandb.ai/keyuxi/NN/runs/3rz0fuwk' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/3rz0fuwk</a><br/> View job at <a href='https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v5' target=\"_blank\">https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v5</a><br/>Synced 6 W&B file(s), 0 media file(s), 0 artifact file(s) and 0 other file(s)"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Find logs at: <code>./wandb/run-20231202_132047-3rz0fuwk/logs</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "wandb version 0.16.0 is available!  To upgrade, please run:\n",
       " $ pip install wandb --upgrade"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Tracking run with wandb version 0.15.12"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Run data is saved locally in <code>/Users/yuxi/workspace/nnn_paper/wandb/run-20231202_132930-roj7xjia</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Syncing run <strong><a href='https://wandb.ai/keyuxi/NN/runs/roj7xjia' target=\"_blank\">fine-dream-92</a></strong> to <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View project at <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">https://wandb.ai/keyuxi/NN</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run at <a href='https://wandb.ai/keyuxi/NN/runs/roj7xjia' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/roj7xjia</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "feats (27732, 178) Index(['bulge_size#1', 'bulge_size#2', 'hairpin_loop_mid#A'], dtype='object')\n",
      "fixed_feature_names: ['bulge_size#1', 'bulge_size#2', 'hairpin_size#3', 'hairpin_size#4', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "                           dH\n",
      "bulge_size#1              0.0\n",
      "bulge_size#2              0.0\n",
      "bulge_size#3              0.0\n",
      "bulge_size#4              0.0\n",
      "bulge_size#5              0.0\n",
      "...                       ...\n",
      "hairpin_tetraloop#CCTTGG  0.0\n",
      "hairpin_tetraloop#GCTTGC  0.0\n",
      "hairpin_tetraloop#GCTTGT  0.0\n",
      "hairpin_tetraloop#TCTTGA  0.0\n",
      "hairpin_tetraloop#TCTTGG  0.0\n",
      "\n",
      "[352 rows x 1 columns]\n",
      "['bulge_size#1', 'bulge_size#2', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#3', 'hairpin_size#4', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "Warning: Rank of matrix A_unknown, 151, is smaller than the number of features 152!\n",
      "                           dG\n",
      "bulge_size#1              4.0\n",
      "bulge_size#2              2.9\n",
      "bulge_size#3              3.1\n",
      "bulge_size#4              3.2\n",
      "bulge_size#5              3.3\n",
      "...                       ...\n",
      "hairpin_tetraloop#CCTTGG -0.8\n",
      "hairpin_tetraloop#GCTTGC -0.8\n",
      "hairpin_tetraloop#GCTTGT -0.5\n",
      "hairpin_tetraloop#TCTTGA -0.8\n",
      "hairpin_tetraloop#TCTTGG -0.5\n",
      "\n",
      "[352 rows x 1 columns]\n",
      "['bulge_size#1', 'bulge_size#2', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#3', 'hairpin_size#4', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "Warning: Rank of matrix A_unknown, 151, is smaller than the number of features 152!\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -3.684\n",
      "\tMean value of fitted new_dict is 1.632\n",
      "\tMean value of centered new_dict is -3.684\n",
      "\n",
      "group hairpin_size\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -2.913\n",
      "\tMean value of fitted new_dict is -5.909\n",
      "\tMean value of centered new_dict is -2.912\n",
      "\n",
      "group interior_size\n",
      "group stack\n",
      "\tMean value of old_dict is -8.275\n",
      "\tMean value of fitted new_dict is -8.275\n",
      "\tMean value of centered new_dict is -8.275\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 1.100\n",
      "\tMean value of fitted new_dict is -1.226\n",
      "\tMean value of centered new_dict is 1.100\n",
      "\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -0.715\n",
      "\tMean value of fitted new_dict is -0.006\n",
      "\tMean value of centered new_dict is -0.715\n",
      "\n",
      "group hairpin_size\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -0.640\n",
      "\tMean value of fitted new_dict is -0.708\n",
      "\tMean value of centered new_dict is -0.640\n",
      "\n",
      "group interior_size\n",
      "group stack\n",
      "\tMean value of old_dict is -1.406\n",
      "\tMean value of fitted new_dict is -1.406\n",
      "\tMean value of centered new_dict is -1.406\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 0.025\n",
      "\tMean value of fitted new_dict is 0.150\n",
      "\tMean value of centered new_dict is 0.025\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<style>\n",
       "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
       "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
       "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
       "    </style>\n",
       "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>▁</td></tr><tr><td>n_fixed_feat</td><td>▁</td></tr><tr><td>train_dG_mae</td><td>▁</td></tr><tr><td>train_dG_rsqr</td><td>▁</td></tr><tr><td>train_dH_mae</td><td>▁</td></tr><tr><td>train_dH_rsqr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-all</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-ov</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-all</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-ov</td><td>▁</td></tr><tr><td>val_Tm_bias-all</td><td>▁</td></tr><tr><td>val_Tm_bias-arr</td><td>▁</td></tr><tr><td>val_Tm_bias-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_bias-ov</td><td>▁</td></tr><tr><td>val_Tm_corr-all</td><td>▁</td></tr><tr><td>val_Tm_corr-arr</td><td>▁</td></tr><tr><td>val_Tm_corr-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_corr-ov</td><td>▁</td></tr><tr><td>val_Tm_mae-all</td><td>▁</td></tr><tr><td>val_Tm_mae-arr</td><td>▁</td></tr><tr><td>val_Tm_mae-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_mae-ov</td><td>▁</td></tr><tr><td>val_Tm_rmse-all</td><td>▁</td></tr><tr><td>val_Tm_rmse-arr</td><td>▁</td></tr><tr><td>val_Tm_rmse-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_rmse-ov</td><td>▁</td></tr><tr><td>val_dG_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_dG_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_dG_bias-arr</td><td>▁</td></tr><tr><td>val_dG_corr-arr</td><td>▁</td></tr><tr><td>val_dG_mae-arr</td><td>▁</td></tr><tr><td>val_dG_rmse-arr</td><td>▁</td></tr><tr><td>val_dH_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_dH_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_dH_bias-arr</td><td>▁</td></tr><tr><td>val_dH_corr-arr</td><td>▁</td></tr><tr><td>val_dH_mae-arr</td><td>▁</td></tr><tr><td>val_dH_rmse-arr</td><td>▁</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>178</td></tr><tr><td>n_fixed_feat</td><td>26</td></tr><tr><td>param_set_file</td><td>./models/fine-dream-...</td></tr><tr><td>train_dG_mae</td><td>0.61028</td></tr><tr><td>train_dG_rsqr</td><td>0.60435</td></tr><tr><td>train_dH_mae</td><td>6.2497</td></tr><tr><td>train_dH_rsqr</td><td>0.43033</td></tr><tr><td>val_Tm_adjusted_mae-all</td><td>16.72361</td></tr><tr><td>val_Tm_adjusted_mae-arr</td><td>23.43088</td></tr><tr><td>val_Tm_adjusted_mae-lit_uv</td><td>2.17537</td></tr><tr><td>val_Tm_adjusted_mae-ov</td><td>1.73121</td></tr><tr><td>val_Tm_adjusted_rmse-all</td><td>25.68218</td></tr><tr><td>val_Tm_adjusted_rmse-arr</td><td>32.86392</td></tr><tr><td>val_Tm_adjusted_rmse-lit_uv</td><td>2.69207</td></tr><tr><td>val_Tm_adjusted_rmse-ov</td><td>2.27246</td></tr><tr><td>val_Tm_bias-all</td><td>16.37152</td></tr><tr><td>val_Tm_bias-arr</td><td>28.70206</td></tr><tr><td>val_Tm_bias-lit_uv</td><td>6.9312</td></tr><tr><td>val_Tm_bias-ov</td><td>4.81577</td></tr><tr><td>val_Tm_corr-all</td><td>0.31305</td></tr><tr><td>val_Tm_corr-arr</td><td>0.42264</td></tr><tr><td>val_Tm_corr-lit_uv</td><td>0.98122</td></tr><tr><td>val_Tm_corr-ov</td><td>0.86289</td></tr><tr><td>val_Tm_mae-all</td><td>17.47327</td></tr><tr><td>val_Tm_mae-arr</td><td>30.96747</td></tr><tr><td>val_Tm_mae-lit_uv</td><td>6.9312</td></tr><tr><td>val_Tm_mae-ov</td><td>4.85342</td></tr><tr><td>val_Tm_rmse-all</td><td>30.45655</td></tr><tr><td>val_Tm_rmse-arr</td><td>43.63308</td></tr><tr><td>val_Tm_rmse-lit_uv</td><td>7.43564</td></tr><tr><td>val_Tm_rmse-ov</td><td>5.32501</td></tr><tr><td>val_dG_adjusted_mae-arr</td><td>0.58773</td></tr><tr><td>val_dG_adjusted_rmse-arr</td><td>0.77332</td></tr><tr><td>val_dG_bias-arr</td><td>-0.93296</td></tr><tr><td>val_dG_corr-arr</td><td>0.8004</td></tr><tr><td>val_dG_mae-arr</td><td>1.01952</td></tr><tr><td>val_dG_rmse-arr</td><td>1.21179</td></tr><tr><td>val_dH_adjusted_mae-arr</td><td>6.55787</td></tr><tr><td>val_dH_adjusted_rmse-arr</td><td>8.38244</td></tr><tr><td>val_dH_bias-arr</td><td>5.11144</td></tr><tr><td>val_dH_corr-arr</td><td>0.69345</td></tr><tr><td>val_dH_mae-arr</td><td>7.62138</td></tr><tr><td>val_dH_rmse-arr</td><td>9.81795</td></tr></table><br/></div></div>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run <strong style=\"color:#cdcd00\">fine-dream-92</strong> at: <a href='https://wandb.ai/keyuxi/NN/runs/roj7xjia' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/roj7xjia</a><br/> View job at <a href='https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v5' target=\"_blank\">https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v5</a><br/>Synced 6 W&B file(s), 0 media file(s), 0 artifact file(s) and 0 other file(s)"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Find logs at: <code>./wandb/run-20231202_132930-roj7xjia/logs</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "wandb version 0.16.0 is available!  To upgrade, please run:\n",
       " $ pip install wandb --upgrade"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Tracking run with wandb version 0.15.12"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Run data is saved locally in <code>/Users/yuxi/workspace/nnn_paper/wandb/run-20231202_133826-79emcnhe</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Syncing run <strong><a href='https://wandb.ai/keyuxi/NN/runs/79emcnhe' target=\"_blank\">giddy-monkey-93</a></strong> to <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View project at <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">https://wandb.ai/keyuxi/NN</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run at <a href='https://wandb.ai/keyuxi/NN/runs/79emcnhe' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/79emcnhe</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "feats (27732, 178) Index(['bulge_size#1', 'bulge_size#2', 'hairpin_loop_mid#A'], dtype='object')\n",
      "fixed_feature_names: ['bulge_size#1', 'bulge_size#2', 'hairpin_size#3', 'hairpin_size#4', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "                           dH\n",
      "bulge_size#1              0.0\n",
      "bulge_size#2              0.0\n",
      "bulge_size#3              0.0\n",
      "bulge_size#4              0.0\n",
      "bulge_size#5              0.0\n",
      "...                       ...\n",
      "hairpin_tetraloop#CCTTGG  0.0\n",
      "hairpin_tetraloop#GCTTGC  0.0\n",
      "hairpin_tetraloop#GCTTGT  0.0\n",
      "hairpin_tetraloop#TCTTGA  0.0\n",
      "hairpin_tetraloop#TCTTGG  0.0\n",
      "\n",
      "[352 rows x 1 columns]\n",
      "['bulge_size#1', 'bulge_size#2', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#3', 'hairpin_size#4', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "                           dG\n",
      "bulge_size#1              4.0\n",
      "bulge_size#2              2.9\n",
      "bulge_size#3              3.1\n",
      "bulge_size#4              3.2\n",
      "bulge_size#5              3.3\n",
      "...                       ...\n",
      "hairpin_tetraloop#CCTTGG -0.8\n",
      "hairpin_tetraloop#GCTTGC -0.8\n",
      "hairpin_tetraloop#GCTTGT -0.5\n",
      "hairpin_tetraloop#TCTTGA -0.8\n",
      "hairpin_tetraloop#TCTTGG -0.5\n",
      "\n",
      "[352 rows x 1 columns]\n",
      "['bulge_size#1', 'bulge_size#2', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#3', 'hairpin_size#4', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "Warning: Rank of matrix A_unknown, 151, is smaller than the number of features 152!\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -3.684\n",
      "\tMean value of fitted new_dict is -12.304\n",
      "\tMean value of centered new_dict is -3.684\n",
      "\n",
      "group hairpin_size\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -2.913\n",
      "\tMean value of fitted new_dict is -5.285\n",
      "\tMean value of centered new_dict is -2.912\n",
      "\n",
      "group interior_size\n",
      "group stack\n",
      "\tMean value of old_dict is -8.275\n",
      "\tMean value of fitted new_dict is -8.275\n",
      "\tMean value of centered new_dict is -8.275\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 1.100\n",
      "\tMean value of fitted new_dict is -0.598\n",
      "\tMean value of centered new_dict is 1.100\n",
      "\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -0.715\n",
      "\tMean value of fitted new_dict is -0.003\n",
      "\tMean value of centered new_dict is -0.715\n",
      "\n",
      "group hairpin_size\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -0.640\n",
      "\tMean value of fitted new_dict is -0.740\n",
      "\tMean value of centered new_dict is -0.640\n",
      "\n",
      "group interior_size\n",
      "group stack\n",
      "\tMean value of old_dict is -1.406\n",
      "\tMean value of fitted new_dict is -1.406\n",
      "\tMean value of centered new_dict is -1.406\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 0.025\n",
      "\tMean value of fitted new_dict is 0.158\n",
      "\tMean value of centered new_dict is 0.025\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<style>\n",
       "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
       "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
       "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
       "    </style>\n",
       "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>▁</td></tr><tr><td>n_fixed_feat</td><td>▁</td></tr><tr><td>train_dG_mae</td><td>▁</td></tr><tr><td>train_dG_rsqr</td><td>▁</td></tr><tr><td>train_dH_mae</td><td>▁</td></tr><tr><td>train_dH_rsqr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-all</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-ov</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-all</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-ov</td><td>▁</td></tr><tr><td>val_Tm_bias-all</td><td>▁</td></tr><tr><td>val_Tm_bias-arr</td><td>▁</td></tr><tr><td>val_Tm_bias-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_bias-ov</td><td>▁</td></tr><tr><td>val_Tm_corr-all</td><td>▁</td></tr><tr><td>val_Tm_corr-arr</td><td>▁</td></tr><tr><td>val_Tm_corr-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_corr-ov</td><td>▁</td></tr><tr><td>val_Tm_mae-all</td><td>▁</td></tr><tr><td>val_Tm_mae-arr</td><td>▁</td></tr><tr><td>val_Tm_mae-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_mae-ov</td><td>▁</td></tr><tr><td>val_Tm_rmse-all</td><td>▁</td></tr><tr><td>val_Tm_rmse-arr</td><td>▁</td></tr><tr><td>val_Tm_rmse-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_rmse-ov</td><td>▁</td></tr><tr><td>val_dG_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_dG_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_dG_bias-arr</td><td>▁</td></tr><tr><td>val_dG_corr-arr</td><td>▁</td></tr><tr><td>val_dG_mae-arr</td><td>▁</td></tr><tr><td>val_dG_rmse-arr</td><td>▁</td></tr><tr><td>val_dH_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_dH_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_dH_bias-arr</td><td>▁</td></tr><tr><td>val_dH_corr-arr</td><td>▁</td></tr><tr><td>val_dH_mae-arr</td><td>▁</td></tr><tr><td>val_dH_rmse-arr</td><td>▁</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>178</td></tr><tr><td>n_fixed_feat</td><td>26</td></tr><tr><td>param_set_file</td><td>./models/giddy-monke...</td></tr><tr><td>train_dG_mae</td><td>0.61478</td></tr><tr><td>train_dG_rsqr</td><td>0.61283</td></tr><tr><td>train_dH_mae</td><td>6.17829</td></tr><tr><td>train_dH_rsqr</td><td>0.44881</td></tr><tr><td>val_Tm_adjusted_mae-all</td><td>45.09545</td></tr><tr><td>val_Tm_adjusted_mae-arr</td><td>96.48488</td></tr><tr><td>val_Tm_adjusted_mae-lit_uv</td><td>2.11822</td></tr><tr><td>val_Tm_adjusted_mae-ov</td><td>1.81493</td></tr><tr><td>val_Tm_adjusted_rmse-all</td><td>87.49842</td></tr><tr><td>val_Tm_adjusted_rmse-arr</td><td>129.49198</td></tr><tr><td>val_Tm_adjusted_rmse-lit_uv</td><td>2.61291</td></tr><tr><td>val_Tm_adjusted_rmse-ov</td><td>2.36302</td></tr><tr><td>val_Tm_bias-all</td><td>2.13126</td></tr><tr><td>val_Tm_bias-arr</td><td>-1.41904</td></tr><tr><td>val_Tm_bias-lit_uv</td><td>6.91674</td></tr><tr><td>val_Tm_bias-ov</td><td>4.87704</td></tr><tr><td>val_Tm_corr-all</td><td>0.2654</td></tr><tr><td>val_Tm_corr-arr</td><td>0.27736</td></tr><tr><td>val_Tm_corr-lit_uv</td><td>0.98181</td></tr><tr><td>val_Tm_corr-ov</td><td>0.85228</td></tr><tr><td>val_Tm_mae-all</td><td>46.50572</td></tr><tr><td>val_Tm_mae-arr</td><td>95.91596</td></tr><tr><td>val_Tm_mae-lit_uv</td><td>6.91674</td></tr><tr><td>val_Tm_mae-ov</td><td>4.90621</td></tr><tr><td>val_Tm_rmse-all</td><td>87.52437</td></tr><tr><td>val_Tm_rmse-arr</td><td>129.49975</td></tr><tr><td>val_Tm_rmse-lit_uv</td><td>7.39382</td></tr><tr><td>val_Tm_rmse-ov</td><td>5.41935</td></tr><tr><td>val_dG_adjusted_mae-arr</td><td>0.57452</td></tr><tr><td>val_dG_adjusted_rmse-arr</td><td>0.76635</td></tr><tr><td>val_dG_bias-arr</td><td>-1.16149</td></tr><tr><td>val_dG_corr-arr</td><td>0.80434</td></tr><tr><td>val_dG_mae-arr</td><td>1.21921</td></tr><tr><td>val_dG_rmse-arr</td><td>1.39152</td></tr><tr><td>val_dH_adjusted_mae-arr</td><td>6.28291</td></tr><tr><td>val_dH_adjusted_rmse-arr</td><td>7.97455</td></tr><tr><td>val_dH_bias-arr</td><td>17.04904</td></tr><tr><td>val_dH_corr-arr</td><td>0.71042</td></tr><tr><td>val_dH_mae-arr</td><td>17.105</td></tr><tr><td>val_dH_rmse-arr</td><td>18.82188</td></tr></table><br/></div></div>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run <strong style=\"color:#cdcd00\">giddy-monkey-93</strong> at: <a href='https://wandb.ai/keyuxi/NN/runs/79emcnhe' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/79emcnhe</a><br/> View job at <a href='https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v5' target=\"_blank\">https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v5</a><br/>Synced 6 W&B file(s), 0 media file(s), 0 artifact file(s) and 0 other file(s)"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Find logs at: <code>./wandb/run-20231202_133826-79emcnhe/logs</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "wandb version 0.16.0 is available!  To upgrade, please run:\n",
       " $ pip install wandb --upgrade"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Tracking run with wandb version 0.15.12"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Run data is saved locally in <code>/Users/yuxi/workspace/nnn_paper/wandb/run-20231202_134718-c7f40ctz</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Syncing run <strong><a href='https://wandb.ai/keyuxi/NN/runs/c7f40ctz' target=\"_blank\">glowing-capybara-114</a></strong> to <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View project at <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">https://wandb.ai/keyuxi/NN</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run at <a href='https://wandb.ai/keyuxi/NN/runs/c7f40ctz' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/c7f40ctz</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "feats (27732, 178) Index(['bulge_size#1', 'bulge_size#2', 'hairpin_loop_mid#A'], dtype='object')\n",
      "fixed_feature_names: ['bulge_size#1', 'bulge_size#2', 'hairpin_size#3', 'hairpin_size#4', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "                           dH\n",
      "bulge_size#1              0.0\n",
      "bulge_size#2              0.0\n",
      "bulge_size#3              0.0\n",
      "bulge_size#4              0.0\n",
      "bulge_size#5              0.0\n",
      "...                       ...\n",
      "hairpin_tetraloop#CCTTGG  0.0\n",
      "hairpin_tetraloop#GCTTGC  0.0\n",
      "hairpin_tetraloop#GCTTGT  0.0\n",
      "hairpin_tetraloop#TCTTGA  0.0\n",
      "hairpin_tetraloop#TCTTGG  0.0\n",
      "\n",
      "[352 rows x 1 columns]\n",
      "['bulge_size#1', 'bulge_size#2', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#3', 'hairpin_size#4', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "                           dG\n",
      "bulge_size#1              4.0\n",
      "bulge_size#2              2.9\n",
      "bulge_size#3              3.1\n",
      "bulge_size#4              3.2\n",
      "bulge_size#5              3.3\n",
      "...                       ...\n",
      "hairpin_tetraloop#CCTTGG -0.8\n",
      "hairpin_tetraloop#GCTTGC -0.8\n",
      "hairpin_tetraloop#GCTTGT -0.5\n",
      "hairpin_tetraloop#TCTTGA -0.8\n",
      "hairpin_tetraloop#TCTTGG -0.5\n",
      "\n",
      "[352 rows x 1 columns]\n",
      "['bulge_size#1', 'bulge_size#2', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#3', 'hairpin_size#4', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -3.684\n",
      "\tMean value of fitted new_dict is -5.907\n",
      "\tMean value of centered new_dict is -3.684\n",
      "\n",
      "group hairpin_size\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -2.913\n",
      "\tMean value of fitted new_dict is -5.344\n",
      "\tMean value of centered new_dict is -2.912\n",
      "\n",
      "group interior_size\n",
      "group stack\n",
      "\tMean value of old_dict is -8.275\n",
      "\tMean value of fitted new_dict is -8.275\n",
      "\tMean value of centered new_dict is -8.275\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 1.100\n",
      "\tMean value of fitted new_dict is -0.796\n",
      "\tMean value of centered new_dict is 1.100\n",
      "\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -0.715\n",
      "\tMean value of fitted new_dict is -4.193\n",
      "\tMean value of centered new_dict is -0.715\n",
      "\n",
      "group hairpin_size\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -0.640\n",
      "\tMean value of fitted new_dict is -0.714\n",
      "\tMean value of centered new_dict is -0.640\n",
      "\n",
      "group interior_size\n",
      "group stack\n",
      "\tMean value of old_dict is -1.406\n",
      "\tMean value of fitted new_dict is -1.406\n",
      "\tMean value of centered new_dict is -1.406\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 0.025\n",
      "\tMean value of fitted new_dict is 0.179\n",
      "\tMean value of centered new_dict is 0.025\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<style>\n",
       "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
       "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
       "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
       "    </style>\n",
       "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>▁</td></tr><tr><td>n_fixed_feat</td><td>▁</td></tr><tr><td>train_dG_mae</td><td>▁</td></tr><tr><td>train_dG_rsqr</td><td>▁</td></tr><tr><td>train_dH_mae</td><td>▁</td></tr><tr><td>train_dH_rsqr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-all</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-ov</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-all</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-ov</td><td>▁</td></tr><tr><td>val_Tm_bias-all</td><td>▁</td></tr><tr><td>val_Tm_bias-arr</td><td>▁</td></tr><tr><td>val_Tm_bias-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_bias-ov</td><td>▁</td></tr><tr><td>val_Tm_corr-all</td><td>▁</td></tr><tr><td>val_Tm_corr-arr</td><td>▁</td></tr><tr><td>val_Tm_corr-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_corr-ov</td><td>▁</td></tr><tr><td>val_Tm_mae-all</td><td>▁</td></tr><tr><td>val_Tm_mae-arr</td><td>▁</td></tr><tr><td>val_Tm_mae-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_mae-ov</td><td>▁</td></tr><tr><td>val_Tm_rmse-all</td><td>▁</td></tr><tr><td>val_Tm_rmse-arr</td><td>▁</td></tr><tr><td>val_Tm_rmse-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_rmse-ov</td><td>▁</td></tr><tr><td>val_dG_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_dG_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_dG_bias-arr</td><td>▁</td></tr><tr><td>val_dG_corr-arr</td><td>▁</td></tr><tr><td>val_dG_mae-arr</td><td>▁</td></tr><tr><td>val_dG_rmse-arr</td><td>▁</td></tr><tr><td>val_dH_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_dH_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_dH_bias-arr</td><td>▁</td></tr><tr><td>val_dH_corr-arr</td><td>▁</td></tr><tr><td>val_dH_mae-arr</td><td>▁</td></tr><tr><td>val_dH_rmse-arr</td><td>▁</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>178</td></tr><tr><td>n_fixed_feat</td><td>26</td></tr><tr><td>param_set_file</td><td>./models/glowing-cap...</td></tr><tr><td>train_dG_mae</td><td>0.62435</td></tr><tr><td>train_dG_rsqr</td><td>0.60464</td></tr><tr><td>train_dH_mae</td><td>6.39664</td></tr><tr><td>train_dH_rsqr</td><td>0.41905</td></tr><tr><td>val_Tm_adjusted_mae-all</td><td>42.20823</td></tr><tr><td>val_Tm_adjusted_mae-arr</td><td>46.9475</td></tr><tr><td>val_Tm_adjusted_mae-lit_uv</td><td>2.12597</td></tr><tr><td>val_Tm_adjusted_mae-ov</td><td>1.73118</td></tr><tr><td>val_Tm_adjusted_rmse-all</td><td>58.78882</td></tr><tr><td>val_Tm_adjusted_rmse-arr</td><td>62.10235</td></tr><tr><td>val_Tm_adjusted_rmse-lit_uv</td><td>2.62855</td></tr><tr><td>val_Tm_adjusted_rmse-ov</td><td>2.25631</td></tr><tr><td>val_Tm_bias-all</td><td>-33.44246</td></tr><tr><td>val_Tm_bias-arr</td><td>-75.44642</td></tr><tr><td>val_Tm_bias-lit_uv</td><td>6.91515</td></tr><tr><td>val_Tm_bias-ov</td><td>4.68249</td></tr><tr><td>val_Tm_corr-all</td><td>0.57284</td></tr><tr><td>val_Tm_corr-arr</td><td>0.57201</td></tr><tr><td>val_Tm_corr-lit_uv</td><td>0.98192</td></tr><tr><td>val_Tm_corr-ov</td><td>0.86401</td></tr><tr><td>val_Tm_mae-all</td><td>39.06107</td></tr><tr><td>val_Tm_mae-arr</td><td>76.38319</td></tr><tr><td>val_Tm_mae-lit_uv</td><td>6.91515</td></tr><tr><td>val_Tm_mae-ov</td><td>4.7201</td></tr><tr><td>val_Tm_rmse-all</td><td>67.63522</td></tr><tr><td>val_Tm_rmse-arr</td><td>97.71829</td></tr><tr><td>val_Tm_rmse-lit_uv</td><td>7.39787</td></tr><tr><td>val_Tm_rmse-ov</td><td>5.19775</td></tr><tr><td>val_dG_adjusted_mae-arr</td><td>0.56553</td></tr><tr><td>val_dG_adjusted_rmse-arr</td><td>0.75418</td></tr><tr><td>val_dG_bias-arr</td><td>3.10385</td></tr><tr><td>val_dG_corr-arr</td><td>0.81181</td></tr><tr><td>val_dG_mae-arr</td><td>3.10403</td></tr><tr><td>val_dG_rmse-arr</td><td>3.19417</td></tr><tr><td>val_dH_adjusted_mae-arr</td><td>6.32659</td></tr><tr><td>val_dH_adjusted_rmse-arr</td><td>8.09115</td></tr><tr><td>val_dH_bias-arr</td><td>11.77707</td></tr><tr><td>val_dH_corr-arr</td><td>0.70588</td></tr><tr><td>val_dH_mae-arr</td><td>12.13929</td></tr><tr><td>val_dH_rmse-arr</td><td>14.28867</td></tr></table><br/></div></div>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run <strong style=\"color:#cdcd00\">glowing-capybara-114</strong> at: <a href='https://wandb.ai/keyuxi/NN/runs/c7f40ctz' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/c7f40ctz</a><br/> View job at <a href='https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v5' target=\"_blank\">https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v5</a><br/>Synced 6 W&B file(s), 0 media file(s), 0 artifact file(s) and 0 other file(s)"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Find logs at: <code>./wandb/run-20231202_134718-c7f40ctz/logs</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "wandb version 0.16.0 is available!  To upgrade, please run:\n",
       " $ pip install wandb --upgrade"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Tracking run with wandb version 0.15.12"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Run data is saved locally in <code>/Users/yuxi/workspace/nnn_paper/wandb/run-20231202_135603-lru8kx01</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Syncing run <strong><a href='https://wandb.ai/keyuxi/NN/runs/lru8kx01' target=\"_blank\">smooth-cloud-115</a></strong> to <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View project at <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">https://wandb.ai/keyuxi/NN</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run at <a href='https://wandb.ai/keyuxi/NN/runs/lru8kx01' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/lru8kx01</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "feats (27732, 178) Index(['bulge_size#1', 'bulge_size#2', 'hairpin_loop_mid#A'], dtype='object')\n",
      "fixed_feature_names: ['bulge_size#1', 'bulge_size#2', 'hairpin_size#3', 'hairpin_size#4', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "                           dH\n",
      "bulge_size#1              0.0\n",
      "bulge_size#2              0.0\n",
      "bulge_size#3              0.0\n",
      "bulge_size#4              0.0\n",
      "bulge_size#5              0.0\n",
      "...                       ...\n",
      "hairpin_tetraloop#CCTTGG  0.0\n",
      "hairpin_tetraloop#GCTTGC  0.0\n",
      "hairpin_tetraloop#GCTTGT  0.0\n",
      "hairpin_tetraloop#TCTTGA  0.0\n",
      "hairpin_tetraloop#TCTTGG  0.0\n",
      "\n",
      "[352 rows x 1 columns]\n",
      "['bulge_size#1', 'bulge_size#2', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#3', 'hairpin_size#4', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "                           dG\n",
      "bulge_size#1              4.0\n",
      "bulge_size#2              2.9\n",
      "bulge_size#3              3.1\n",
      "bulge_size#4              3.2\n",
      "bulge_size#5              3.3\n",
      "...                       ...\n",
      "hairpin_tetraloop#CCTTGG -0.8\n",
      "hairpin_tetraloop#GCTTGC -0.8\n",
      "hairpin_tetraloop#GCTTGT -0.5\n",
      "hairpin_tetraloop#TCTTGA -0.8\n",
      "hairpin_tetraloop#TCTTGG -0.5\n",
      "\n",
      "[352 rows x 1 columns]\n",
      "['bulge_size#1', 'bulge_size#2', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#3', 'hairpin_size#4', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -3.684\n",
      "\tMean value of fitted new_dict is -11.518\n",
      "\tMean value of centered new_dict is -3.684\n",
      "\n",
      "group hairpin_size\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -2.913\n",
      "\tMean value of fitted new_dict is -5.289\n",
      "\tMean value of centered new_dict is -2.913\n",
      "\n",
      "group interior_size\n",
      "group stack\n",
      "\tMean value of old_dict is -8.275\n",
      "\tMean value of fitted new_dict is -8.275\n",
      "\tMean value of centered new_dict is -8.275\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 1.100\n",
      "\tMean value of fitted new_dict is -0.825\n",
      "\tMean value of centered new_dict is 1.100\n",
      "\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -0.715\n",
      "\tMean value of fitted new_dict is -4.157\n",
      "\tMean value of centered new_dict is -0.715\n",
      "\n",
      "group hairpin_size\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -0.640\n",
      "\tMean value of fitted new_dict is -0.708\n",
      "\tMean value of centered new_dict is -0.640\n",
      "\n",
      "group interior_size\n",
      "group stack\n",
      "\tMean value of old_dict is -1.406\n",
      "\tMean value of fitted new_dict is -1.406\n",
      "\tMean value of centered new_dict is -1.406\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 0.025\n",
      "\tMean value of fitted new_dict is 0.176\n",
      "\tMean value of centered new_dict is 0.025\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<style>\n",
       "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
       "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
       "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
       "    </style>\n",
       "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>▁</td></tr><tr><td>n_fixed_feat</td><td>▁</td></tr><tr><td>train_dG_mae</td><td>▁</td></tr><tr><td>train_dG_rsqr</td><td>▁</td></tr><tr><td>train_dH_mae</td><td>▁</td></tr><tr><td>train_dH_rsqr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-all</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-ov</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-all</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-ov</td><td>▁</td></tr><tr><td>val_Tm_bias-all</td><td>▁</td></tr><tr><td>val_Tm_bias-arr</td><td>▁</td></tr><tr><td>val_Tm_bias-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_bias-ov</td><td>▁</td></tr><tr><td>val_Tm_corr-all</td><td>▁</td></tr><tr><td>val_Tm_corr-arr</td><td>▁</td></tr><tr><td>val_Tm_corr-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_corr-ov</td><td>▁</td></tr><tr><td>val_Tm_mae-all</td><td>▁</td></tr><tr><td>val_Tm_mae-arr</td><td>▁</td></tr><tr><td>val_Tm_mae-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_mae-ov</td><td>▁</td></tr><tr><td>val_Tm_rmse-all</td><td>▁</td></tr><tr><td>val_Tm_rmse-arr</td><td>▁</td></tr><tr><td>val_Tm_rmse-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_rmse-ov</td><td>▁</td></tr><tr><td>val_dG_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_dG_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_dG_bias-arr</td><td>▁</td></tr><tr><td>val_dG_corr-arr</td><td>▁</td></tr><tr><td>val_dG_mae-arr</td><td>▁</td></tr><tr><td>val_dG_rmse-arr</td><td>▁</td></tr><tr><td>val_dH_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_dH_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_dH_bias-arr</td><td>▁</td></tr><tr><td>val_dH_corr-arr</td><td>▁</td></tr><tr><td>val_dH_mae-arr</td><td>▁</td></tr><tr><td>val_dH_rmse-arr</td><td>▁</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>178</td></tr><tr><td>n_fixed_feat</td><td>26</td></tr><tr><td>param_set_file</td><td>./models/smooth-clou...</td></tr><tr><td>train_dG_mae</td><td>0.61903</td></tr><tr><td>train_dG_rsqr</td><td>0.61976</td></tr><tr><td>train_dH_mae</td><td>6.32634</td></tr><tr><td>train_dH_rsqr</td><td>0.42596</td></tr><tr><td>val_Tm_adjusted_mae-all</td><td>46.31002</td></tr><tr><td>val_Tm_adjusted_mae-arr</td><td>63.07217</td></tr><tr><td>val_Tm_adjusted_mae-lit_uv</td><td>2.11705</td></tr><tr><td>val_Tm_adjusted_mae-ov</td><td>1.79312</td></tr><tr><td>val_Tm_adjusted_rmse-all</td><td>77.57277</td></tr><tr><td>val_Tm_adjusted_rmse-arr</td><td>103.67653</td></tr><tr><td>val_Tm_adjusted_rmse-lit_uv</td><td>2.61598</td></tr><tr><td>val_Tm_adjusted_rmse-ov</td><td>2.31434</td></tr><tr><td>val_Tm_bias-all</td><td>-27.60997</td></tr><tr><td>val_Tm_bias-arr</td><td>-71.11577</td></tr><tr><td>val_Tm_bias-lit_uv</td><td>6.92076</td></tr><tr><td>val_Tm_bias-ov</td><td>4.67132</td></tr><tr><td>val_Tm_corr-all</td><td>0.38068</td></tr><tr><td>val_Tm_corr-arr</td><td>0.34975</td></tr><tr><td>val_Tm_corr-lit_uv</td><td>0.98194</td></tr><tr><td>val_Tm_corr-ov</td><td>0.85575</td></tr><tr><td>val_Tm_mae-all</td><td>40.25737</td></tr><tr><td>val_Tm_mae-arr</td><td>87.45695</td></tr><tr><td>val_Tm_mae-lit_uv</td><td>6.92076</td></tr><tr><td>val_Tm_mae-ov</td><td>4.71807</td></tr><tr><td>val_Tm_rmse-all</td><td>82.33981</td></tr><tr><td>val_Tm_rmse-arr</td><td>125.72301</td></tr><tr><td>val_Tm_rmse-lit_uv</td><td>7.39867</td></tr><tr><td>val_Tm_rmse-ov</td><td>5.21319</td></tr><tr><td>val_dG_adjusted_mae-arr</td><td>0.56393</td></tr><tr><td>val_dG_adjusted_rmse-arr</td><td>0.75065</td></tr><tr><td>val_dG_bias-arr</td><td>2.88816</td></tr><tr><td>val_dG_corr-arr</td><td>0.80779</td></tr><tr><td>val_dG_mae-arr</td><td>2.88864</td></tr><tr><td>val_dG_rmse-arr</td><td>2.98412</td></tr><tr><td>val_dH_adjusted_mae-arr</td><td>6.27711</td></tr><tr><td>val_dH_adjusted_rmse-arr</td><td>7.99594</td></tr><tr><td>val_dH_bias-arr</td><td>17.08792</td></tr><tr><td>val_dH_corr-arr</td><td>0.71185</td></tr><tr><td>val_dH_mae-arr</td><td>17.13718</td></tr><tr><td>val_dH_rmse-arr</td><td>18.86617</td></tr></table><br/></div></div>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run <strong style=\"color:#cdcd00\">smooth-cloud-115</strong> at: <a href='https://wandb.ai/keyuxi/NN/runs/lru8kx01' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/lru8kx01</a><br/> View job at <a href='https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v5' target=\"_blank\">https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v5</a><br/>Synced 6 W&B file(s), 0 media file(s), 0 artifact file(s) and 0 other file(s)"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Find logs at: <code>./wandb/run-20231202_135603-lru8kx01/logs</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "wandb version 0.16.0 is available!  To upgrade, please run:\n",
       " $ pip install wandb --upgrade"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Tracking run with wandb version 0.15.12"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Run data is saved locally in <code>/Users/yuxi/workspace/nnn_paper/wandb/run-20231202_140446-gt93mvx9</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Syncing run <strong><a href='https://wandb.ai/keyuxi/NN/runs/gt93mvx9' target=\"_blank\">swift-lake-118</a></strong> to <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View project at <a href='https://wandb.ai/keyuxi/NN' target=\"_blank\">https://wandb.ai/keyuxi/NN</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run at <a href='https://wandb.ai/keyuxi/NN/runs/gt93mvx9' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/gt93mvx9</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "feats (27732, 178) Index(['bulge_size#1', 'bulge_size#2', 'hairpin_loop_mid#A'], dtype='object')\n",
      "fixed_feature_names: ['bulge_size#1', 'bulge_size#2', 'hairpin_size#3', 'hairpin_size#4', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "                           dH\n",
      "bulge_size#1              0.0\n",
      "bulge_size#2              0.0\n",
      "bulge_size#3              0.0\n",
      "bulge_size#4              0.0\n",
      "bulge_size#5              0.0\n",
      "...                       ...\n",
      "hairpin_tetraloop#CCTTGG  0.0\n",
      "hairpin_tetraloop#GCTTGC  0.0\n",
      "hairpin_tetraloop#GCTTGT  0.0\n",
      "hairpin_tetraloop#TCTTGA  0.0\n",
      "hairpin_tetraloop#TCTTGG  0.0\n",
      "\n",
      "[352 rows x 1 columns]\n",
      "['bulge_size#1', 'bulge_size#2', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#3', 'hairpin_size#4', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "                           dG\n",
      "bulge_size#1              4.0\n",
      "bulge_size#2              2.9\n",
      "bulge_size#3              3.1\n",
      "bulge_size#4              3.2\n",
      "bulge_size#5              3.3\n",
      "...                       ...\n",
      "hairpin_tetraloop#CCTTGG -0.8\n",
      "hairpin_tetraloop#GCTTGC -0.8\n",
      "hairpin_tetraloop#GCTTGT -0.5\n",
      "hairpin_tetraloop#TCTTGA -0.8\n",
      "hairpin_tetraloop#TCTTGG -0.5\n",
      "\n",
      "[352 rows x 1 columns]\n",
      "['bulge_size#1', 'bulge_size#2', 'hairpin_size#11', 'hairpin_size#14', 'hairpin_size#17', 'hairpin_size#20', 'hairpin_size#3', 'hairpin_size#4', 'interior_size#2', 'interior_size#4', 'stack#AATT', 'stack#ACGT', 'stack#AGCT', 'stack#ATAT', 'stack#CATG', 'stack#CCGG', 'stack#CGCG', 'stack#CTAG', 'stack#GATC', 'stack#GCGC', 'stack#GGCC', 'stack#GTAC', 'stack#TATA', 'stack#TCGA', 'stack#TGCA', 'stack#TTAA']\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -3.684\n",
      "\tMean value of fitted new_dict is -12.986\n",
      "\tMean value of centered new_dict is -3.684\n",
      "\n",
      "group hairpin_size\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -2.913\n",
      "\tMean value of fitted new_dict is -5.266\n",
      "\tMean value of centered new_dict is -2.913\n",
      "\n",
      "group interior_size\n",
      "group stack\n",
      "\tMean value of old_dict is -8.275\n",
      "\tMean value of fitted new_dict is -8.275\n",
      "\tMean value of centered new_dict is -8.275\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 1.100\n",
      "\tMean value of fitted new_dict is -0.687\n",
      "\tMean value of centered new_dict is 1.100\n",
      "\n",
      "Centering new parameters...\n",
      "group bulge_size\n",
      "group hairpin_mismatch\n",
      "\tMean value of old_dict is -0.715\n",
      "\tMean value of fitted new_dict is -4.034\n",
      "\tMean value of centered new_dict is -0.715\n",
      "\n",
      "group hairpin_size\n",
      "group interior_mismatch\n",
      "\tMean value of old_dict is -0.640\n",
      "\tMean value of fitted new_dict is -0.724\n",
      "\tMean value of centered new_dict is -0.640\n",
      "\n",
      "group interior_size\n",
      "group stack\n",
      "\tMean value of old_dict is -1.406\n",
      "\tMean value of fitted new_dict is -1.406\n",
      "\tMean value of centered new_dict is -1.406\n",
      "\n",
      "group terminal_penalty\n",
      "\tMean value of old_dict is 0.025\n",
      "\tMean value of fitted new_dict is 0.170\n",
      "\tMean value of centered new_dict is 0.025\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<style>\n",
       "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
       "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
       "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
       "    </style>\n",
       "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>▁</td></tr><tr><td>n_fixed_feat</td><td>▁</td></tr><tr><td>train_dG_mae</td><td>▁</td></tr><tr><td>train_dG_rsqr</td><td>▁</td></tr><tr><td>train_dH_mae</td><td>▁</td></tr><tr><td>train_dH_rsqr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-all</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_adjusted_mae-ov</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-all</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_adjusted_rmse-ov</td><td>▁</td></tr><tr><td>val_Tm_bias-all</td><td>▁</td></tr><tr><td>val_Tm_bias-arr</td><td>▁</td></tr><tr><td>val_Tm_bias-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_bias-ov</td><td>▁</td></tr><tr><td>val_Tm_corr-all</td><td>▁</td></tr><tr><td>val_Tm_corr-arr</td><td>▁</td></tr><tr><td>val_Tm_corr-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_corr-ov</td><td>▁</td></tr><tr><td>val_Tm_mae-all</td><td>▁</td></tr><tr><td>val_Tm_mae-arr</td><td>▁</td></tr><tr><td>val_Tm_mae-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_mae-ov</td><td>▁</td></tr><tr><td>val_Tm_rmse-all</td><td>▁</td></tr><tr><td>val_Tm_rmse-arr</td><td>▁</td></tr><tr><td>val_Tm_rmse-lit_uv</td><td>▁</td></tr><tr><td>val_Tm_rmse-ov</td><td>▁</td></tr><tr><td>val_dG_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_dG_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_dG_bias-arr</td><td>▁</td></tr><tr><td>val_dG_corr-arr</td><td>▁</td></tr><tr><td>val_dG_mae-arr</td><td>▁</td></tr><tr><td>val_dG_rmse-arr</td><td>▁</td></tr><tr><td>val_dH_adjusted_mae-arr</td><td>▁</td></tr><tr><td>val_dH_adjusted_rmse-arr</td><td>▁</td></tr><tr><td>val_dH_bias-arr</td><td>▁</td></tr><tr><td>val_dH_corr-arr</td><td>▁</td></tr><tr><td>val_dH_mae-arr</td><td>▁</td></tr><tr><td>val_dH_rmse-arr</td><td>▁</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>n_feat</td><td>178</td></tr><tr><td>n_fixed_feat</td><td>26</td></tr><tr><td>param_set_file</td><td>./models/swift-lake-...</td></tr><tr><td>train_dG_mae</td><td>0.62251</td></tr><tr><td>train_dG_rsqr</td><td>0.61267</td></tr><tr><td>train_dH_mae</td><td>6.36952</td></tr><tr><td>train_dH_rsqr</td><td>0.42186</td></tr><tr><td>val_Tm_adjusted_mae-all</td><td>41.73711</td></tr><tr><td>val_Tm_adjusted_mae-arr</td><td>60.36296</td></tr><tr><td>val_Tm_adjusted_mae-lit_uv</td><td>2.11421</td></tr><tr><td>val_Tm_adjusted_mae-ov</td><td>1.76578</td></tr><tr><td>val_Tm_adjusted_rmse-all</td><td>72.18559</td></tr><tr><td>val_Tm_adjusted_rmse-arr</td><td>104.05513</td></tr><tr><td>val_Tm_adjusted_rmse-lit_uv</td><td>2.61512</td></tr><tr><td>val_Tm_adjusted_rmse-ov</td><td>2.29492</td></tr><tr><td>val_Tm_bias-all</td><td>-16.83271</td></tr><tr><td>val_Tm_bias-arr</td><td>-46.75773</td></tr><tr><td>val_Tm_bias-lit_uv</td><td>6.91914</td></tr><tr><td>val_Tm_bias-ov</td><td>4.66014</td></tr><tr><td>val_Tm_corr-all</td><td>0.33816</td></tr><tr><td>val_Tm_corr-arr</td><td>0.30792</td></tr><tr><td>val_Tm_corr-lit_uv</td><td>0.98196</td></tr><tr><td>val_Tm_corr-ov</td><td>0.85889</td></tr><tr><td>val_Tm_mae-all</td><td>37.22837</td></tr><tr><td>val_Tm_mae-arr</td><td>81.64919</td></tr><tr><td>val_Tm_mae-lit_uv</td><td>6.91914</td></tr><tr><td>val_Tm_mae-ov</td><td>4.704</td></tr><tr><td>val_Tm_rmse-all</td><td>74.12219</td></tr><tr><td>val_Tm_rmse-arr</td><td>114.07785</td></tr><tr><td>val_Tm_rmse-lit_uv</td><td>7.39685</td></tr><tr><td>val_Tm_rmse-ov</td><td>5.19457</td></tr><tr><td>val_dG_adjusted_mae-arr</td><td>0.56022</td></tr><tr><td>val_dG_adjusted_rmse-arr</td><td>0.7493</td></tr><tr><td>val_dG_bias-arr</td><td>2.76177</td></tr><tr><td>val_dG_corr-arr</td><td>0.809</td></tr><tr><td>val_dG_mae-arr</td><td>2.76234</td></tr><tr><td>val_dG_rmse-arr</td><td>2.86161</td></tr><tr><td>val_dH_adjusted_mae-arr</td><td>6.27013</td></tr><tr><td>val_dH_adjusted_rmse-arr</td><td>7.97789</td></tr><tr><td>val_dH_bias-arr</td><td>18.42177</td></tr><tr><td>val_dH_corr-arr</td><td>0.7107</td></tr><tr><td>val_dH_mae-arr</td><td>18.44259</td></tr><tr><td>val_dH_rmse-arr</td><td>20.07507</td></tr></table><br/></div></div>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run <strong style=\"color:#cdcd00\">swift-lake-118</strong> at: <a href='https://wandb.ai/keyuxi/NN/runs/gt93mvx9' target=\"_blank\">https://wandb.ai/keyuxi/NN/runs/gt93mvx9</a><br/> View job at <a href='https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v5' target=\"_blank\">https://wandb.ai/keyuxi/NN/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjExMTcyMjg1NQ==/version_details/v5</a><br/>Synced 6 W&B file(s), 0 media file(s), 0 artifact file(s) and 0 other file(s)"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Find logs at: <code>./wandb/run-20231202_140446-gt93mvx9/logs</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAqUAAAFkCAYAAAAZhEK+AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/OQEPoAAAACXBIWXMAAA9hAAAPYQGoP6dpAAAbtElEQVR4nO3db2yVZ/nA8Yu5lLYrylSKGkbLMILOhRY2QTDoqsQ/pC7GmGLMJFU359wanJItm9uY2xyGOOuG0mwzvjToG9mYhP3xhduEIQRNDFsWz4AyitLqKoNyWoHn98J4ftYW9xRabs7x80mW7Nx9OFx3mFe+UkonZVmWBQAAJHRB6gEAAECUAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBI7oyi9JFHHonbb7991I/19vZGe3t7NDc3x/Lly2P37t1nNSBApbJLAf7fmKJ0aGgoOjs74/vf//5pn7njjjti7ty58cILL8R1110XN998c5w8efKsBwWoFHYpwEhjitJ777039uzZEytWrBj140ePHo1nn302brjhhqiqqoqrr746pkyZEtu3bx+XYQEqgV0KMNKFY3n4pptuimnTpsVDDz0Uf/7zn0d8vLu7Oy6++OKYMmVK6ayxsTEKhUIsWbJk1Pc8fPhw9Pb2jjg/cuRIFAqFeN/73heTJ08ey5gAb2hwcDBeffXV+NCHPhRvfetbz+nPPd671B4FUhnPXTqmKJ02bdp//fjAwMCIxVddXR3FYvG0P2bjxo2xfv36sYwBMG7WrVsXn/70p8/pzzneu9QeBVIbj106pih9IzU1NTE4ODjsrFgsRm1t7Wl/TFtbW7S0tIw4f/HFF+P222+PdevWxezZs8dzTIAoFAqxevXqmDFjRupRRhjrLrVHgVTGc5eOa5Q2NDREf39/HD16NOrq6iIiYu/evaf9c1MREfX19VFfX3/aj8+ePTsuu+yy8RwToOR8/LT2WHepPQqkNh67dFz/ntK6urpYsmRJPPjggzE0NBSPPfZY9Pf3xxVXXDGePw1ARbNLgf9FZx2lPT090dzcHD09PRHxz68q3bdvX3zwgx+MRx99NH70ox9FVVXVWQ8KUMnsUuB/3Rl9+v6mm24q/fu73vWuYX+p87Rp0+Lhhx8++8kAKpxdCvD/fJtRAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQ3JiidNeuXdHa2hpNTU3R3t4efX19I545ePBgrFy5MhYsWBCf+MQn4plnnhm3YQEqgV0KMFLuKC0Wi9HR0REdHR2xY8eOaGhoiLVr14547r777osrr7wydu7cGXfeeWd84xvfiGKxOK5DA5QruxRgdLmjdNu2bTF9+vRYtmxZVFVVxapVq2Lr1q0xMDAw7Lnu7u44depUnDp1KiZNmhQ1NTXjPjRAubJLAUZ3Yd4H9+/fH42NjaXXU6dOjdra2uju7o65c+eWzleuXBl33313dHV1RUREZ2dnVFdXn/Z9Dx8+HL29vSPOC4VC3tEAysZE7FJ7FKgEuaN0YGAgJk+ePOyspqZmxKeTTp06Fbfccku0tbXFc889F7feemtcfvnl8c53vnPU9924cWOsX7/+DEYHKD8TsUvtUaAS5I7SmpqaGBoaGnZ2/PjxqK2tLb3+y1/+Eg888EBs27YtLrjggmhpaYnm5uZ46qmn4otf/OKo79vW1hYtLS0jzguFQqxevTrveABlYSJ2qT0KVILcUTpr1qzYvHlz6XV/f38cO3YsZs6cWTrr6+uLf/zjH8N+3Jve9Ka48MLT/zT19fVRX18/lpkBytZE7FJ7FKgEub/QadGiRXHo0KHYsmVLDA0NRWdnZ7S0tAz7M07vfve746KLLoof//jHcerUqdi+fXvs2LEjli5dOiHDA5QbuxRgdLmjtLq6OjZs2BBdXV2xcOHCOHDgQKxZsyZ6enqiubk5enp6YvLkydHV1RXPP/98XHnllXHvvffGD37wg5gxY8ZE3gGgbNilAKPL/en7iIh58+bFpk2bRpzv3r279O+XXXZZ/OxnPzv7yQAqlF0KMJJvMwoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJjSlKd+3aFa2trdHU1BTt7e3R19c34plisRh33nlnLFmyJJYuXRq/+MUvxm1YgEpglwKMlDtKi8VidHR0REdHR+zYsSMaGhpi7dq1I5677777or+/P55++ul45JFH4nvf+17s27dvPGcGKFt2KcDockfptm3bYvr06bFs2bKoqqqKVatWxdatW2NgYKD0zNDQUDz++ONxxx13RE1NTcyZMyc2btwYb3/72ydkeIByY5cCjC53lO7fvz8aGxtLr6dOnRq1tbXR3d1dOtu3b1/U1dXF5s2b48Mf/nB8/OMfj5dffjnq6urGdWiAcmWXAozuwrwPDgwMxOTJk4ed1dTURLFYLL0+cuRI/O1vf4u9e/fG1q1bY8+ePXHttdfGnDlz4tJLLx31fQ8fPhy9vb0jzguFQt7RAMrGROxSexSoBLmjtKamJoaGhoadHT9+PGpra0uvq6qq4uTJk7Fq1aqorq6O+fPnx+LFi+P5558/bZRu3Lgx1q9ff4bjA5SXidil9ihQCXJH6axZs2Lz5s2l1/39/XHs2LGYOXNm6WzmzJkxadKkeP311+Otb31rREScOHEisiw77fu2tbVFS0vLiPNCoRCrV6/OOx5AWZiIXWqPApUg958pXbRoURw6dCi2bNkSQ0ND0dnZGS0tLVFdXV16ZurUqbF06dLo7OyMwcHB2LVrV2zfvj2uuuqq075vfX19XHbZZSP+mT179tndDOA8NBG71B4FKkHuKK2uro4NGzZEV1dXLFy4MA4cOBBr1qyJnp6eaG5ujp6enoiIWLduXUyaNCmWLl0aq1evjvvuuy8uueSSCbsAQDmxSwFGl/vT9xER8+bNi02bNo043717d+nf3/KWt8QDDzxw9pMBVCi7FGAk32YUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkN6Yo3bVrV7S2tkZTU1O0t7dHX1/faZ997bXXYvHixfHCCy+c9ZAAlcQuBRgpd5QWi8Xo6OiIjo6O2LFjRzQ0NMTatWtP+/y9994br7322rgMCVAp7FKA0eWO0m3btsX06dNj2bJlUVVVFatWrYqtW7fGwMDAiGd//etfx9GjR2PGjBnjOixAubNLAUZ3Yd4H9+/fH42NjaXXU6dOjdra2uju7o65c+eWzo8cORLr1q2Ln/70p3HNNde84fsePnw4ent7R5wXCoW8owGUjYnYpfYoUAlyR+nAwEBMnjx52FlNTU0Ui8VhZ/fff39cc8018Y53vCPX+27cuDHWr1+fdwyAsjYRu9QeBSpB7iitqamJoaGhYWfHjx+P2tra0utnn302uru747vf/W7uAdra2qKlpWXEeaFQiNWrV+d+H4ByMBG71B4FKkHuKJ01a1Zs3ry59Lq/vz+OHTsWM2fOLJ09+eSTsWfPnrjyyisjIuLYsWNx/fXXx3e+851obW0d9X3r6+ujvr7+TOcHKCsTsUvtUaAS5P5Cp0WLFsWhQ4diy5YtMTQ0FJ2dndHS0hLV1dWlZ+65557YvXt37Ny5M3bu3BkzZsyIrq6u0wYpwP8auxRgdLmjtLq6OjZs2BBdXV2xcOHCOHDgQKxZsyZ6enqiubk5enp6JnJOgIpglwKMLven7yMi5s2bF5s2bRpxvnv37lGff+qpp85sKoAKZpcCjOTbjAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByY4rSXbt2RWtrazQ1NUV7e3v09fWNeGbPnj2xYsWKWLBgQXzyk5+Mp59+etyGBagEdinASLmjtFgsRkdHR3R0dMSOHTuioaEh1q5dO+yZkydPxte//vX47Gc/G7/73e/irrvuiltuuSUOHjw47oMDlCO7FGB0uaN027ZtMX369Fi2bFlUVVXFqlWrYuvWrTEwMFB6pq+vL97//vfH5z73ubjgggti0aJF0dDQEC+++OKEDA9QbuxSgNHljtL9+/dHY2Nj6fXUqVOjtrY2uru7S2fTp0+Phx56qPS6p6cnCoVCzJkzZ3ymBShzdinA6C7M++DAwEBMnjx52FlNTU0Ui8VRn//73/8eN9xwQ7S1tcUll1xy2vc9fPhw9Pb2jjgvFAp5RwMoGxOxS+1RoBLkjtKampoYGhoadnb8+PGora0d8WxPT0985StfiXnz5sWtt976X99348aNsX79+rxjAJS1idil9ihQCXJH6axZs2Lz5s2l1/39/XHs2LGYOXPmsOdeeeWVWLlyZVx99dXxrW996w3ft62tLVpaWkacFwqFWL16dd7xAMrCROxSexSoBLmjdNGiRXHbbbfFli1b4qMf/Wh0dnZGS0tLVFdXl54ZHByM66+/Ptra2uLGG2/M9b719fVRX18/9skBytBE7FJ7FKgEub/Qqbq6OjZs2BBdXV2xcOHCOHDgQKxZsyZ6enqiubk5enp64plnnon9+/fHT37yk2hubi7986tf/Woi7wBQNuxSgNHl/p3SiIh58+bFpk2bRpzv3r07IiLe9a53xac+9anxmQygQtmlACP5NqMAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACC5MUXprl27orW1NZqamqK9vT36+vpGPNPb2xvt7e3R3Nwcy5cvj927d4/bsACVwC4FGCl3lBaLxejo6IiOjo7YsWNHNDQ0xNq1a0c8d8cdd8TcuXPjhRdeiOuuuy5uvvnmOHny5LgODVCu7FKA0eWO0m3btsX06dNj2bJlUVVVFatWrYqtW7fGwMBA6ZmjR4/Gs88+GzfccENUVVXF1VdfHVOmTInt27dPyPAA5cYuBRjdhXkf3L9/fzQ2NpZeT506NWpra6O7uzvmzp0bERHd3d1x8cUXx5QpU0rPNTY2RqFQiCVLloz6vocPH47e3t4R5y+++GJERBQKhbwjAuT2r90yODh4Tn/eidil9iiQynju0txROjAwEJMnTx52VlNTE8Vi8b8+U11dPeyZ/7Rx48ZYv379aT++evXqvCMCjNmePXti/vz55+znm4hdao8CqY3HLs0dpTU1NTE0NDTs7Pjx41FbWzvsmf8s5WKxOOyZ/9TW1hYtLS0jzv/whz/E3XffHd/97ndLv3tQSQqFQqxevTrWrVsXs2fPTj3OuKv0+0VU/h0r/X4vvfRS3Hbbbef8bhOxS+3RyvxvtNLvF1H5d6z0+0WM7y7NHaWzZs2KzZs3l1739/fHsWPHYubMmaWzhoaG6O/vj6NHj0ZdXV1EROzduzdWrFhx2vetr6+P+vr603587ty5cdlll+Uds+zMnj3b/cpcpd+x0u/35je/+Zz+fBOxS+3Ryv5vtNLvF1H5d6z0+0WMzy7N/YVOixYtikOHDsWWLVtiaGgoOjs7o6WlJaqrq0vP1NXVxZIlS+LBBx+MoaGheOyxx6K/vz+uuOKKsx4UoBLYpQCjyx2l1dXVsWHDhujq6oqFCxfGgQMHYs2aNdHT0xPNzc3R09MTERH33ntv7Nu3Lz74wQ/Go48+Gj/60Y+iqqpqwi4AUE7sUoDR5f70fUTEvHnzYtOmTSPO//0vdZ42bVo8/PDDZz8ZQIWySwFG8m1GAQBI7ryN0mnTpsWNN94Y06ZNSz3KhHC/8lfpd3S/8lfpd3S/8lfpd6z0+0WM7x0nZVmWjcNMAABwxs7b3ykFAOB/hygFACA5UQoAQHLJo3TXrl3R2toaTU1N0d7eHn19fSOe6e3tjfb29mhubo7ly5cP+2tTznd57rdnz55YsWJFLFiwID75yU/G008/nWDSM5Pnfv/y2muvxeLFi+OFF144hxOevTx3LBaLceedd8aSJUti6dKl8Ytf/CLBpGcmz/0OHjwYK1eujAULFsQnPvGJeOaZZxJMenYeeeSRuP3220f9WDnvmAh7NMIePd/Zo/ZoLllCx48fzxYvXpw9+eST2eDgYHbXXXdl3/zmN0c899WvfjVbu3ZtNjg4mP3yl7/MPvKRj2QnTpxIMPHY5LnfiRMnso985CPZz3/+8+zkyZPZtm3bsvnz52evvvpqoqnzy/vr9y8333xzNnfu3Gz79u3ncMqzk/eO3/72t7ObbropGxgYyF566aVswYIF2d69e8/9wGOU935f+9rXsoceeig7depU9vzzz2eXX355dvz48QQTj93g4GD2gx/8IJszZ0522223jfpMue6YLLNHs8wePd/Zo/9kj76xpFH661//OvvMZz5Tev3aa69l73//+7Njx46Vzl5//fXsfe97X3bkyJHSWWtra/bcc8+d01nPRJ77/fnPf85uvPHGYT/uM5/5TPbUU0+dsznPVJ77/cszzzyTXXfdddnHPvaxslqmee44ODiYzZs3Lzt8+HDp7E9/+lP2+uuvn9NZz0TeX8Ply5dnP/zhD7MTJ05kv/3tb7MPfOADZbNM77jjjuzaa6/N7rrrrlGXaTnvmCyzR7PMHj3f2aP/ZI++saSfvt+/f380NjaWXk+dOjVqa2uju7u7dNbd3R0XX3xxTJkypXTW2NgYhULhXI56RvLcb/r06fHQQw+VXvf09EShUIg5c+acy1HPSJ77RUQcOXIk1q1bF3ffffc5nvDs5bnjvn37oq6uLjZv3hwf/vCH4+Mf/3i8/PLLUVdXl2Discn7a7hy5cp4+OGH4/LLL48vf/nLcc899wz7Xu3ns5tuuikefvjheNvb3jbqx8t5x0TYoxH26PnOHv0ne/SNjenbjI63gYGBmDx58rCzmpqaKBaL//WZ6urqYc+cr/Lc79/9/e9/jxtuuCHa2trikksuORcjnpW897v//vvjmmuuiXe84x3ncrxxkeeOR44cib/97W+xd+/e2Lp1a+zZsyeuvfbamDNnTlx66aXneuQxyftreOrUqbjllluira0tnnvuubj11lvj8ssvj3e+853nctwz8kZ/oXM575gIe/Q/2aPnH3v0n+zRN5b0d0prampiaGho2Nnx48ejtrZ22DODg4PDnikWi8OeOV/lud+/9PT0xOc///l473vfG7feeuu5GvGs5Lnfs88+G93d3fH5z3/+XI83LvLcsaqqKk6ePBmrVq2K6urqmD9/fixevDief/75cz3umOW531/+8pd44IEH4gtf+EJUVVVFS0tLNDc3x1NPPXWux50Q5bxjIuzRf2ePnp/sUXs0r6RROmvWrNi3b1/pdX9/fxw7dixmzpxZOmtoaIj+/v44evRo6Wzv3r3n/f9zish3v4iIV155Jdra2qKlpSXuv//+uOCC5H8pQi557vfkk0/Gnj174sorr4wrrrgiXn311bj++uvj8ccfTzDx2OW548yZM2PSpEnx+uuvl85OnDgRWRl8s7Q89+vr64t//OMfw37cm970prjwwqSfaBk35bxjIuzRf7FHz1/2qD2aV9L/1S5atCgOHToUW7ZsiaGhoejs7IyWlpZhf8airq4ulixZEg8++GAMDQ3FY489Fv39/XHFFVcknDyfPPcbHByM66+/Ptra2uJb3/pWwmnHLs/97rnnnti9e3fs3Lkzdu7cGTNmzIiurq5obW1NOHl+ee44derUWLp0aXR2dsbg4GDs2rUrtm/fHldddVXCyfPJc793v/vdcdFFF8WPf/zjOHXqVGzfvj127NgRS5cuTTj5+CnnHRNhj0bYo+c7e9Qeze0svyDrrP3+97/PPv3pT2dNTU3Zl770peyvf/1rdvDgwaypqSk7ePBglmVZdvjw4ezaa6/N5s+fn7W2tmZ/+MMfEk+d3xvd74knnsje8573ZE1NTcP+eeKJJ1KPnkueX79/V25fNZpl+e7Y39+ffeMb38g+8IEPZFdddVXZ/PplWb77/fGPf8xWrFiRzZ8/P1u+fHn2m9/8JvHUY/fggw+Wvmq0knZMltmj9uj5zx61R/OYlGVl8HvjAABUtPL4QzcAAFQ0UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMn9HwU/ZlxKGdtiAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 800x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAqUAAAFkCAYAAAAZhEK+AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/OQEPoAAAACXBIWXMAAA9hAAAPYQGoP6dpAAAbtElEQVR4nO3db2yVZ/nA8Yu5lLYrylSKGkbLMILOhRY2QTDoqsQ/pC7GmGLMJFU359wanJItm9uY2xyGOOuG0mwzvjToG9mYhP3xhduEIQRNDFsWz4AyitLqKoNyWoHn98J4ftYW9xRabs7x80mW7Nx9OFx3mFe+UkonZVmWBQAAJHRB6gEAAECUAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBI7oyi9JFHHonbb7991I/19vZGe3t7NDc3x/Lly2P37t1nNSBApbJLAf7fmKJ0aGgoOjs74/vf//5pn7njjjti7ty58cILL8R1110XN998c5w8efKsBwWoFHYpwEhjitJ777039uzZEytWrBj140ePHo1nn302brjhhqiqqoqrr746pkyZEtu3bx+XYQEqgV0KMNKFY3n4pptuimnTpsVDDz0Uf/7zn0d8vLu7Oy6++OKYMmVK6ayxsTEKhUIsWbJk1Pc8fPhw9Pb2jjg/cuRIFAqFeN/73heTJ08ey5gAb2hwcDBeffXV+NCHPhRvfetbz+nPPd671B4FUhnPXTqmKJ02bdp//fjAwMCIxVddXR3FYvG0P2bjxo2xfv36sYwBMG7WrVsXn/70p8/pzzneu9QeBVIbj106pih9IzU1NTE4ODjsrFgsRm1t7Wl/TFtbW7S0tIw4f/HFF+P222+PdevWxezZs8dzTIAoFAqxevXqmDFjRupRRhjrLrVHgVTGc5eOa5Q2NDREf39/HD16NOrq6iIiYu/evaf9c1MREfX19VFfX3/aj8+ePTsuu+yy8RwToOR8/LT2WHepPQqkNh67dFz/ntK6urpYsmRJPPjggzE0NBSPPfZY9Pf3xxVXXDGePw1ARbNLgf9FZx2lPT090dzcHD09PRHxz68q3bdvX3zwgx+MRx99NH70ox9FVVXVWQ8KUMnsUuB/3Rl9+v6mm24q/fu73vWuYX+p87Rp0+Lhhx8++8kAKpxdCvD/fJtRAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQ3JiidNeuXdHa2hpNTU3R3t4efX19I545ePBgrFy5MhYsWBCf+MQn4plnnhm3YQEqgV0KMFLuKC0Wi9HR0REdHR2xY8eOaGhoiLVr14547r777osrr7wydu7cGXfeeWd84xvfiGKxOK5DA5QruxRgdLmjdNu2bTF9+vRYtmxZVFVVxapVq2Lr1q0xMDAw7Lnu7u44depUnDp1KiZNmhQ1NTXjPjRAubJLAUZ3Yd4H9+/fH42NjaXXU6dOjdra2uju7o65c+eWzleuXBl33313dHV1RUREZ2dnVFdXn/Z9Dx8+HL29vSPOC4VC3tEAysZE7FJ7FKgEuaN0YGAgJk+ePOyspqZmxKeTTp06Fbfccku0tbXFc889F7feemtcfvnl8c53vnPU9924cWOsX7/+DEYHKD8TsUvtUaAS5I7SmpqaGBoaGnZ2/PjxqK2tLb3+y1/+Eg888EBs27YtLrjggmhpaYnm5uZ46qmn4otf/OKo79vW1hYtLS0jzguFQqxevTrveABlYSJ2qT0KVILcUTpr1qzYvHlz6XV/f38cO3YsZs6cWTrr6+uLf/zjH8N+3Jve9Ka48MLT/zT19fVRX18/lpkBytZE7FJ7FKgEub/QadGiRXHo0KHYsmVLDA0NRWdnZ7S0tAz7M07vfve746KLLoof//jHcerUqdi+fXvs2LEjli5dOiHDA5QbuxRgdLmjtLq6OjZs2BBdXV2xcOHCOHDgQKxZsyZ6enqiubk5enp6YvLkydHV1RXPP/98XHnllXHvvffGD37wg5gxY8ZE3gGgbNilAKPL/en7iIh58+bFpk2bRpzv3r279O+XXXZZ/OxnPzv7yQAqlF0KMJJvMwoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJjSlKd+3aFa2trdHU1BTt7e3R19c34plisRh33nlnLFmyJJYuXRq/+MUvxm1YgEpglwKMlDtKi8VidHR0REdHR+zYsSMaGhpi7dq1I5677777or+/P55++ul45JFH4nvf+17s27dvPGcGKFt2KcDockfptm3bYvr06bFs2bKoqqqKVatWxdatW2NgYKD0zNDQUDz++ONxxx13RE1NTcyZMyc2btwYb3/72ydkeIByY5cCjC53lO7fvz8aGxtLr6dOnRq1tbXR3d1dOtu3b1/U1dXF5s2b48Mf/nB8/OMfj5dffjnq6urGdWiAcmWXAozuwrwPDgwMxOTJk4ed1dTURLFYLL0+cuRI/O1vf4u9e/fG1q1bY8+ePXHttdfGnDlz4tJLLx31fQ8fPhy9vb0jzguFQt7RAMrGROxSexSoBLmjtKamJoaGhoadHT9+PGpra0uvq6qq4uTJk7Fq1aqorq6O+fPnx+LFi+P5558/bZRu3Lgx1q9ff4bjA5SXidil9ihQCXJH6axZs2Lz5s2l1/39/XHs2LGYOXNm6WzmzJkxadKkeP311+Otb31rREScOHEisiw77fu2tbVFS0vLiPNCoRCrV6/OOx5AWZiIXWqPApUg958pXbRoURw6dCi2bNkSQ0ND0dnZGS0tLVFdXV16ZurUqbF06dLo7OyMwcHB2LVrV2zfvj2uuuqq075vfX19XHbZZSP+mT179tndDOA8NBG71B4FKkHuKK2uro4NGzZEV1dXLFy4MA4cOBBr1qyJnp6eaG5ujp6enoiIWLduXUyaNCmWLl0aq1evjvvuuy8uueSSCbsAQDmxSwFGl/vT9xER8+bNi02bNo043717d+nf3/KWt8QDDzxw9pMBVCi7FGAk32YUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkN6Yo3bVrV7S2tkZTU1O0t7dHX1/faZ997bXXYvHixfHCCy+c9ZAAlcQuBRgpd5QWi8Xo6OiIjo6O2LFjRzQ0NMTatWtP+/y9994br7322rgMCVAp7FKA0eWO0m3btsX06dNj2bJlUVVVFatWrYqtW7fGwMDAiGd//etfx9GjR2PGjBnjOixAubNLAUZ3Yd4H9+/fH42NjaXXU6dOjdra2uju7o65c+eWzo8cORLr1q2Ln/70p3HNNde84fsePnw4ent7R5wXCoW8owGUjYnYpfYoUAlyR+nAwEBMnjx52FlNTU0Ui8VhZ/fff39cc8018Y53vCPX+27cuDHWr1+fdwyAsjYRu9QeBSpB7iitqamJoaGhYWfHjx+P2tra0utnn302uru747vf/W7uAdra2qKlpWXEeaFQiNWrV+d+H4ByMBG71B4FKkHuKJ01a1Zs3ry59Lq/vz+OHTsWM2fOLJ09+eSTsWfPnrjyyisjIuLYsWNx/fXXx3e+851obW0d9X3r6+ujvr7+TOcHKCsTsUvtUaAS5P5Cp0WLFsWhQ4diy5YtMTQ0FJ2dndHS0hLV1dWlZ+65557YvXt37Ny5M3bu3BkzZsyIrq6u0wYpwP8auxRgdLmjtLq6OjZs2BBdXV2xcOHCOHDgQKxZsyZ6enqiubk5enp6JnJOgIpglwKMLven7yMi5s2bF5s2bRpxvnv37lGff+qpp85sKoAKZpcCjOTbjAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByY4rSXbt2RWtrazQ1NUV7e3v09fWNeGbPnj2xYsWKWLBgQXzyk5+Mp59+etyGBagEdinASLmjtFgsRkdHR3R0dMSOHTuioaEh1q5dO+yZkydPxte//vX47Gc/G7/73e/irrvuiltuuSUOHjw47oMDlCO7FGB0uaN027ZtMX369Fi2bFlUVVXFqlWrYuvWrTEwMFB6pq+vL97//vfH5z73ubjgggti0aJF0dDQEC+++OKEDA9QbuxSgNHljtL9+/dHY2Nj6fXUqVOjtrY2uru7S2fTp0+Phx56qPS6p6cnCoVCzJkzZ3ymBShzdinA6C7M++DAwEBMnjx52FlNTU0Ui8VRn//73/8eN9xwQ7S1tcUll1xy2vc9fPhw9Pb2jjgvFAp5RwMoGxOxS+1RoBLkjtKampoYGhoadnb8+PGora0d8WxPT0985StfiXnz5sWtt976X99348aNsX79+rxjAJS1idil9ihQCXJH6axZs2Lz5s2l1/39/XHs2LGYOXPmsOdeeeWVWLlyZVx99dXxrW996w3ft62tLVpaWkacFwqFWL16dd7xAMrCROxSexSoBLmjdNGiRXHbbbfFli1b4qMf/Wh0dnZGS0tLVFdXl54ZHByM66+/Ptra2uLGG2/M9b719fVRX18/9skBytBE7FJ7FKgEub/Qqbq6OjZs2BBdXV2xcOHCOHDgQKxZsyZ6enqiubk5enp64plnnon9+/fHT37yk2hubi7986tf/Woi7wBQNuxSgNHl/p3SiIh58+bFpk2bRpzv3r07IiLe9a53xac+9anxmQygQtmlACP5NqMAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACA5UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMmJUgAAkhOlAAAkJ0oBAEhOlAIAkJwoBQAgOVEKAEByohQAgOREKQAAyYlSAACSE6UAACQnSgEASE6UAgCQnCgFACC5MUXprl27orW1NZqamqK9vT36+vpGPNPb2xvt7e3R3Nwcy5cvj927d4/bsACVwC4FGCl3lBaLxejo6IiOjo7YsWNHNDQ0xNq1a0c8d8cdd8TcuXPjhRdeiOuuuy5uvvnmOHny5LgODVCu7FKA0eWO0m3btsX06dNj2bJlUVVVFatWrYqtW7fGwMBA6ZmjR4/Gs88+GzfccENUVVXF1VdfHVOmTInt27dPyPAA5cYuBRjdhXkf3L9/fzQ2NpZeT506NWpra6O7uzvmzp0bERHd3d1x8cUXx5QpU0rPNTY2RqFQiCVLloz6vocPH47e3t4R5y+++GJERBQKhbwjAuT2r90yODh4Tn/eidil9iiQynju0txROjAwEJMnTx52VlNTE8Vi8b8+U11dPeyZ/7Rx48ZYv379aT++evXqvCMCjNmePXti/vz55+znm4hdao8CqY3HLs0dpTU1NTE0NDTs7Pjx41FbWzvsmf8s5WKxOOyZ/9TW1hYtLS0jzv/whz/E3XffHd/97ndLv3tQSQqFQqxevTrWrVsXs2fPTj3OuKv0+0VU/h0r/X4vvfRS3Hbbbef8bhOxS+3RyvxvtNLvF1H5d6z0+0WM7y7NHaWzZs2KzZs3l1739/fHsWPHYubMmaWzhoaG6O/vj6NHj0ZdXV1EROzduzdWrFhx2vetr6+P+vr603587ty5cdlll+Uds+zMnj3b/cpcpd+x0u/35je/+Zz+fBOxS+3Ryv5vtNLvF1H5d6z0+0WMzy7N/YVOixYtikOHDsWWLVtiaGgoOjs7o6WlJaqrq0vP1NXVxZIlS+LBBx+MoaGheOyxx6K/vz+uuOKKsx4UoBLYpQCjyx2l1dXVsWHDhujq6oqFCxfGgQMHYs2aNdHT0xPNzc3R09MTERH33ntv7Nu3Lz74wQ/Go48+Gj/60Y+iqqpqwi4AUE7sUoDR5f70fUTEvHnzYtOmTSPO//0vdZ42bVo8/PDDZz8ZQIWySwFG8m1GAQBI7ryN0mnTpsWNN94Y06ZNSz3KhHC/8lfpd3S/8lfpd3S/8lfpd6z0+0WM7x0nZVmWjcNMAABwxs7b3ykFAOB/hygFACA5UQoAQHLJo3TXrl3R2toaTU1N0d7eHn19fSOe6e3tjfb29mhubo7ly5cP+2tTznd57rdnz55YsWJFLFiwID75yU/G008/nWDSM5Pnfv/y2muvxeLFi+OFF144hxOevTx3LBaLceedd8aSJUti6dKl8Ytf/CLBpGcmz/0OHjwYK1eujAULFsQnPvGJeOaZZxJMenYeeeSRuP3220f9WDnvmAh7NMIePd/Zo/ZoLllCx48fzxYvXpw9+eST2eDgYHbXXXdl3/zmN0c899WvfjVbu3ZtNjg4mP3yl7/MPvKRj2QnTpxIMPHY5LnfiRMnso985CPZz3/+8+zkyZPZtm3bsvnz52evvvpqoqnzy/vr9y8333xzNnfu3Gz79u3ncMqzk/eO3/72t7ObbropGxgYyF566aVswYIF2d69e8/9wGOU935f+9rXsoceeig7depU9vzzz2eXX355dvz48QQTj93g4GD2gx/8IJszZ0522223jfpMue6YLLNHs8wePd/Zo/9kj76xpFH661//OvvMZz5Tev3aa69l73//+7Njx46Vzl5//fXsfe97X3bkyJHSWWtra/bcc8+d01nPRJ77/fnPf85uvPHGYT/uM5/5TPbUU0+dsznPVJ77/cszzzyTXXfdddnHPvaxslqmee44ODiYzZs3Lzt8+HDp7E9/+lP2+uuvn9NZz0TeX8Ply5dnP/zhD7MTJ05kv/3tb7MPfOADZbNM77jjjuzaa6/N7rrrrlGXaTnvmCyzR7PMHj3f2aP/ZI++saSfvt+/f380NjaWXk+dOjVqa2uju7u7dNbd3R0XX3xxTJkypXTW2NgYhULhXI56RvLcb/r06fHQQw+VXvf09EShUIg5c+acy1HPSJ77RUQcOXIk1q1bF3ffffc5nvDs5bnjvn37oq6uLjZv3hwf/vCH4+Mf/3i8/PLLUVdXl2Discn7a7hy5cp4+OGH4/LLL48vf/nLcc899wz7Xu3ns5tuuikefvjheNvb3jbqx8t5x0TYoxH26PnOHv0ne/SNjenbjI63gYGBmDx58rCzmpqaKBaL//WZ6urqYc+cr/Lc79/9/e9/jxtuuCHa2trikksuORcjnpW897v//vvjmmuuiXe84x3ncrxxkeeOR44cib/97W+xd+/e2Lp1a+zZsyeuvfbamDNnTlx66aXneuQxyftreOrUqbjllluira0tnnvuubj11lvj8ssvj3e+853nctwz8kZ/oXM575gIe/Q/2aPnH3v0n+zRN5b0d0prampiaGho2Nnx48ejtrZ22DODg4PDnikWi8OeOV/lud+/9PT0xOc///l473vfG7feeuu5GvGs5Lnfs88+G93d3fH5z3/+XI83LvLcsaqqKk6ePBmrVq2K6urqmD9/fixevDief/75cz3umOW531/+8pd44IEH4gtf+EJUVVVFS0tLNDc3x1NPPXWux50Q5bxjIuzRf2ePnp/sUXs0r6RROmvWrNi3b1/pdX9/fxw7dixmzpxZOmtoaIj+/v44evRo6Wzv3r3n/f9zish3v4iIV155Jdra2qKlpSXuv//+uOCC5H8pQi557vfkk0/Gnj174sorr4wrrrgiXn311bj++uvj8ccfTzDx2OW548yZM2PSpEnx+uuvl85OnDgRWRl8s7Q89+vr64t//OMfw37cm970prjwwqSfaBk35bxjIuzRf7FHz1/2qD2aV9L/1S5atCgOHToUW7ZsiaGhoejs7IyWlpZhf8airq4ulixZEg8++GAMDQ3FY489Fv39/XHFFVcknDyfPPcbHByM66+/Ptra2uJb3/pWwmnHLs/97rnnnti9e3fs3Lkzdu7cGTNmzIiurq5obW1NOHl+ee44derUWLp0aXR2dsbg4GDs2rUrtm/fHldddVXCyfPJc793v/vdcdFFF8WPf/zjOHXqVGzfvj127NgRS5cuTTj5+CnnHRNhj0bYo+c7e9Qeze0svyDrrP3+97/PPv3pT2dNTU3Zl770peyvf/1rdvDgwaypqSk7ePBglmVZdvjw4ezaa6/N5s+fn7W2tmZ/+MMfEk+d3xvd74knnsje8573ZE1NTcP+eeKJJ1KPnkueX79/V25fNZpl+e7Y39+ffeMb38g+8IEPZFdddVXZ/PplWb77/fGPf8xWrFiRzZ8/P1u+fHn2m9/8JvHUY/fggw+Wvmq0knZMltmj9uj5zx61R/OYlGVl8HvjAABUtPL4QzcAAFQ0UQoAQHKiFACA5EQpAADJiVIAAJITpQAAJCdKAQBITpQCAJCcKAUAIDlRCgBAcqIUAIDkRCkAAMn9HwU/ZlxKGdtiAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 800x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for ratio in myrange:\n",
    "    config.update(dict(use_train_set_ratio=ratio))\n",
    "    tnn.model_pipeline(config)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Run model test only"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "config = dict(\n",
    "    use_model_from='json',\n",
    "    test_mode='val',\n",
    ")\n",
    "val_result_df = tnn.test(\n",
    "    config,\n",
    "    json_file='./models/dna-nnn-swift-lake-118-adj.json',\n",
    "    debug=False,\n",
    "    log_wandb=False,\n",
    "    save_val_result_df=True\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "val_result_df.to_csv('./models/swift-lake-118-ddX_val_result_df.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.18"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}