Rx1st.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "source": [
    "import tensorflow as tf"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "source": [
    "gpus = tf.config.list_physical_devices(\"GPU\")\n",
    "if gpus:\n",
    "    for gpu in gpus:\n",
    "        tf.config.experimental.set_memory_growth(gpu, True)\n",
    "\n",
    "    logical_gpus = tf.config.experimental.list_logical_devices(\"GPU\")\n",
    "    print(len(gpus), \"Physical GPUs,\", len(logical_gpus), \"Logical GPUs\")"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "1 Physical GPUs, 1 Logical GPUs\n"
     ]
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Fingerprint-based reactivity prediction model\n",
    "\n",
    "This model takes as input the structural fingerprint for each of three reactants and a vector representing the reagents present. To indicate when one reactant is absent (_i.e._ a binary reaction), the model also takes a _presence_ bit vector, which is zero for any component that is missing in the reaction."
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "source": [
    "import itertools\n",
    "import random\n",
    "from math import factorial\n",
    "from os import path\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "from matplotlib import pylab as plt\n",
    "\n",
    "import fast_jtnn\n",
    "import tensorflow as tf\n",
    "import tensorflow.keras as tfk\n",
    "from fast_jtnn.fp_calculator import FingerprintCalculator\n",
    "from rdkit import Chem\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.model_selection import train_test_split\n",
    "from tensorflow.keras.callbacks import EarlyStopping\n",
    "\n",
    "sns.set(\n",
    "    context=\"talk\",\n",
    "    style=\"ticks\",\n",
    "    font=\"Arial\",\n",
    "    font_scale=1.0,\n",
    "    rc={\"svg.fonttype\": \"none\", \"lines.linewidth\": 1.5, \"figure.autolayout\": True},\n",
    ")"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "source": [
    "JTNN_HOME = path.dirname(fast_jtnn.__path__[0])\n",
    "JTNN_MODEL_PATH = path.join(JTNN_HOME, \"data\", \"models\", \"model.iter-20000\")\n",
    "JTNN_MOLS = path.join(JTNN_HOME, \"data\", \"full_train.txt\")\n",
    "JTNN_VOCAB_PATH = path.join(JTNN_HOME, \"data\", \"vocab_full.txt\")"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "source": [
    "chemical_spaces = {\n",
    "    \"simple\": {\n",
    "        \"reactants_dict\": {\n",
    "            \"phenyl_isocyanate\": \"O=C=N/c1ccccc1\",\n",
    "            \"benzylamine\": \"c1ccc(cc1)CN\",\n",
    "            \"cyclopentanone\": \"C1CCC(=O)C1\",\n",
    "            \"cinnamaldehyde\": \"c1ccc(cc1)C=CC=O\",\n",
    "            \"sorbic_alcohol\": \"CC=CC=CCO\",\n",
    "            \"maleic_anhydride\": \"C1=CC(=O)OC1=O\",\n",
    "        },\n",
    "        \"reagents_dict\": {\"base\": \"C1CC2=NCCCN2C1\",},\n",
    "        \"data\": pd.read_csv(\"vector_space_simple.csv\", index_col=None),\n",
    "    },\n",
    "    \"photo\": {\n",
    "        \"reactants_dict\": {\n",
    "            \"bromoacetonitrile\": \"C(C#N)Br\",\n",
    "            \"bromomalonate\": \"CCOC(=O)C(C(=O)OCC)Br\",\n",
    "            \"diethyl_thiobarbituric_acid\": \"CCN1C(=O)CC(=O)N(C1=S)CC\",\n",
    "            \"dimethyl_nitrosoaniline\": \"CN(C)c1ccc(cc1)N=O\",\n",
    "            \"dimethylisoxazole\": \"CC1=CC(=NO1)C\",\n",
    "            \"ethyl_vinyl_ether\": \"CCOC=C\",\n",
    "            \"glycidyl_propargyl_ether\": \"C#CCOCC1CO1\",\n",
    "            \"hexadienal\": \"CC=CC=CC=O\",\n",
    "            \"indole\": \"C1=CC=C2C(=C1)C=CN2\",\n",
    "            \"phenylboronic_acid\": \"B(C1=CC=CC=C1)(O)O\",\n",
    "            \"phenylhydrazine\": \"C1=CC=C(C=C1)NN\",\n",
    "            \"phenylurazole\": \"C1=CC=C(C=C1)N2C(=O)NNC2=O\",\n",
    "            \"thiadiazole\": \"C1=NN=CS1\",\n",
    "            \"tosyl_isocyanide\": \"CC1=CC=C(C=C1)S(=O)(=O)C[N+]#[C-]\",\n",
    "            \"vinyl_pyrrolidinone\": \"C=CN1CCCC1=O\",\n",
    "        },\n",
    "        \"reagents_dict\": {\n",
    "            \"base\": \"CN(C)C1=CC=NC=C1\",\n",
    "            \"acid\": \"C(F)(F)(F)S(=O)(=O)[O-].C(F)(F)(F)S(=O)(=O)[O-].C(F)(F)(F)S(=O)(=O)[O-].[Sc+3]\",\n",
    "            \"pyrylium\": \"F[B-](F)(F)F.c1ccc(cc1)-c2cc([o+]c(c2)-c3ccccc3)-c4ccccc4\",\n",
    "            \"rose_bengal\": \"C1=C2C(=C3C=C(C(=O)C(=C3OC2=C(C(=C1I)O)I)I)I)C4=C(C(=C(C(=C4Cl)Cl)Cl)Cl)C(=O)O\",\n",
    "            \"rubypy\": \"C1=CC2=C(C3=C(C=CC=N3)C=C2)N=C1.C1=CC2=C(C3=C(C=CC=N3)C=C2)N=C1.C1=CC2=C(C3=C(C=CC=N3)C=C2)N=C1.O.O.O.O.O.O.[Cl-].[Cl-].[Ru+2]\",\n",
    "        },\n",
    "        \"data\": pd.read_csv(\"vector_space.csv\", index_col=None),\n",
    "    },\n",
    "}"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "### Choice of chemical space"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "source": [
    "active_space = \"photo\"\n",
    "chemical_space = chemical_spaces[active_space]\n",
    "locals().update(chemical_space)"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "source": [
    "fp_calculator = FingerprintCalculator(\n",
    "    model_path=JTNN_MODEL_PATH, vocab_path=JTNN_VOCAB_PATH\n",
    ")"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stderr",
     "text": [
      "/home/group/mambaforge/envs/Rx1st/lib/python3.8/site-packages/torch/nn/_reduction.py:44: UserWarning: size_average and reduce args will be deprecated, please use reduction='sum' instead.\n",
      "  warnings.warn(warning.format(ret))\n"
     ]
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "source": [
    "reactant_smiles = []\n",
    "reactant_names = []\n",
    "\n",
    "for name, smiles in reactants_dict.items():\n",
    "    reactant_names.append(name)\n",
    "    mol = Chem.MolFromSmiles(smiles)\n",
    "    Chem.SanitizeMol(mol)\n",
    "    reactant_smiles.append(Chem.MolToSmiles(mol))"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "source": [
    "reagent_smiles = []\n",
    "reagent_names = []\n",
    "\n",
    "for name, smiles in reagents_dict.items():\n",
    "    reagent_names.append(name)\n",
    "    mol = Chem.MolFromSmiles(smiles)\n",
    "    Chem.SanitizeMol(mol)\n",
    "    reagent_smiles.append(Chem.MolToSmiles(mol))"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "source": [
    "reactant_fps = fp_calculator(reactant_smiles)\n",
    "reactant_fps.shape"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stderr",
     "text": [
      "100%|██████████| 1/1 [00:00<00:00,  3.66it/s]\n"
     ]
    },
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "(15, 56)"
      ]
     },
     "metadata": {},
     "execution_count": 10
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "source": [
    "data"
   ],
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>bromoacetonitrile</th>\n",
       "      <th>bromomalonate</th>\n",
       "      <th>diethyl_thiobarbituric_acid</th>\n",
       "      <th>dimethyl_nitrosoaniline</th>\n",
       "      <th>dimethylisoxazole</th>\n",
       "      <th>ethyl_vinyl_ether</th>\n",
       "      <th>glycidyl_propargyl_ether</th>\n",
       "      <th>hexadienal</th>\n",
       "      <th>indole</th>\n",
       "      <th>phenylboronic_acid</th>\n",
       "      <th>...</th>\n",
       "      <th>phenylurazole</th>\n",
       "      <th>thiadiazole</th>\n",
       "      <th>tosyl_isocyanide</th>\n",
       "      <th>vinyl_pyrrolidinone</th>\n",
       "      <th>base</th>\n",
       "      <th>acid</th>\n",
       "      <th>pyrylium</th>\n",
       "      <th>rose_bengal</th>\n",
       "      <th>rubypy</th>\n",
       "      <th>prediction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.691278</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.165638</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.693691</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.013263</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.699631</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1013</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.802568</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1014</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.593064</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1015</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.474527</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1016</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.588960</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1017</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.508840</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1018 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      bromoacetonitrile  bromomalonate  diethyl_thiobarbituric_acid  \\\n",
       "0                     0              0                            1   \n",
       "1                     0              0                            0   \n",
       "2                     0              1                            0   \n",
       "3                     0              0                            1   \n",
       "4                     0              0                            1   \n",
       "...                 ...            ...                          ...   \n",
       "1013                  0              0                            0   \n",
       "1014                  1              0                            1   \n",
       "1015                  1              0                            0   \n",
       "1016                  0              0                            0   \n",
       "1017                  1              0                            1   \n",
       "\n",
       "      dimethyl_nitrosoaniline  dimethylisoxazole  ethyl_vinyl_ether  \\\n",
       "0                           0                  0                  1   \n",
       "1                           0                  1                  0   \n",
       "2                           0                  1                  0   \n",
       "3                           0                  0                  0   \n",
       "4                           0                  1                  0   \n",
       "...                       ...                ...                ...   \n",
       "1013                        0                  0                  1   \n",
       "1014                        0                  0                  0   \n",
       "1015                        1                  0                  1   \n",
       "1016                        0                  0                  1   \n",
       "1017                        0                  0                  0   \n",
       "\n",
       "      glycidyl_propargyl_ether  hexadienal  indole  phenylboronic_acid  ...  \\\n",
       "0                            0           0       0                   0  ...   \n",
       "1                            1           0       0                   0  ...   \n",
       "2                            0           0       0                   0  ...   \n",
       "3                            0           0       0                   1  ...   \n",
       "4                            0           0       0                   0  ...   \n",
       "...                        ...         ...     ...                 ...  ...   \n",
       "1013                         0           0       0                   0  ...   \n",
       "1014                         0           0       0                   0  ...   \n",
       "1015                         0           0       0                   0  ...   \n",
       "1016                         0           0       0                   0  ...   \n",
       "1017                         0           0       0                   0  ...   \n",
       "\n",
       "      phenylurazole  thiadiazole  tosyl_isocyanide  vinyl_pyrrolidinone  base  \\\n",
       "0                 0            0                 0                    0     1   \n",
       "1                 0            0                 0                    1     0   \n",
       "2                 0            0                 1                    0     0   \n",
       "3                 0            0                 0                    0     0   \n",
       "4                 0            0                 0                    0     0   \n",
       "...             ...          ...               ...                  ...   ...   \n",
       "1013              1            0                 0                    0     0   \n",
       "1014              1            0                 0                    0     0   \n",
       "1015              0            0                 0                    0     0   \n",
       "1016              1            0                 0                    0     0   \n",
       "1017              1            0                 0                    0     0   \n",
       "\n",
       "      acid  pyrylium  rose_bengal  rubypy  prediction  \n",
       "0        0         0            1       0    0.691278  \n",
       "1        0         0            0       1    0.165638  \n",
       "2        0         1            0       0    0.693691  \n",
       "3        0         0            0       0    0.013263  \n",
       "4        0         0            0       0    0.699631  \n",
       "...    ...       ...          ...     ...         ...  \n",
       "1013     0         0            0       1    0.802568  \n",
       "1014     0         1            0       0    0.593064  \n",
       "1015     1         0            0       0    0.474527  \n",
       "1016     0         0            0       0    0.588960  \n",
       "1017     0         0            0       0    0.508840  \n",
       "\n",
       "[1018 rows x 21 columns]"
      ]
     },
     "metadata": {},
     "execution_count": 11
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "# Load dataset and associated fingerprints"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "source": [
    "reactant_mat = data[reactant_names].to_numpy(dtype=bool)\n",
    "present_mat = np.sort(data[reactant_names].to_numpy(dtype=\"float32\"))[:, :-4:-1]\n",
    "reagent_mat = data[reagent_names].to_numpy(dtype=\"float32\")\n",
    "predictions = data.prediction.to_numpy(dtype=\"float32\")"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "### Add reactivity class weights to balance the data"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "source": [
    "n_cats = 4\n",
    "counts, edges = np.histogram(predictions, bins=n_cats)\n",
    "weights = 1 / counts * len(predictions) / n_cats\n",
    "bins = np.digitize(predictions, edges) - 1\n",
    "bins = np.minimum(bins, n_cats - 1)\n",
    "\n",
    "sample_weights = weights[bins]"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "source": [
    "reaction_fps = np.zeros(\n",
    "    (reactant_mat.shape[0], 3, reactant_fps.shape[1]), dtype=\"float32\"\n",
    ")\n",
    "\n",
    "for i, slice in enumerate(reactant_mat):\n",
    "    reaction_fps[i, : sum(slice)] = reactant_fps[slice]\n",
    "\n",
    "reaction_fps.shape"
   ],
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "(1018, 3, 56)"
      ]
     },
     "metadata": {},
     "execution_count": 14
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "source": [
    "(\n",
    "    reactant_mat_train,\n",
    "    reactant_mat_test,\n",
    "    reactant_train,\n",
    "    reactant_test,\n",
    "    present_train,\n",
    "    present_test,\n",
    "    reagent_train,\n",
    "    reagent_test,\n",
    "    pred_train,\n",
    "    pred_test,\n",
    "    weights_train,\n",
    "    weights_test,\n",
    ") = train_test_split(\n",
    "    reactant_mat,\n",
    "    reaction_fps,\n",
    "    present_mat,\n",
    "    reagent_mat,\n",
    "    predictions,\n",
    "    sample_weights,\n",
    "    test_size=0.1,\n",
    "    random_state=2018,\n",
    ")"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "# Model"
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Normalization layer\n",
    "\n",
    "Shift/scale molecular fingerprints to have a unit normal distribution around 0."
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "source": [
    "with open(JTNN_MOLS, \"r\") as f:\n",
    "    smiles = [l.strip() for l in f]\n",
    "\n",
    "fps = fp_calculator(smiles[:5000])"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stderr",
     "text": [
      "  4%|▍         | 1/25 [00:02<00:51,  2.16s/it]RDKit WARNING: [11:20:48] WARNING: not removing hydrogen atom without neighbors\n",
      "RDKit WARNING: [11:20:48] WARNING: not removing hydrogen atom without neighbors\n",
      " 80%|████████  | 20/25 [00:41<00:09,  1.98s/it]RDKit WARNING: [11:21:27] WARNING: not removing hydrogen atom without neighbors\n",
      "RDKit WARNING: [11:21:27] WARNING: not removing hydrogen atom without neighbors\n",
      "RDKit WARNING: [11:21:27] WARNING: not removing hydrogen atom without neighbors\n",
      "RDKit WARNING: [11:21:27] WARNING: not removing hydrogen atom without neighbors\n",
      "100%|██████████| 25/25 [00:50<00:00,  2.01s/it]\n"
     ]
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "source": [
    "normalizer = tfk.layers.experimental.preprocessing.Normalization()\n",
    "normalizer.adapt(fps)"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "source": [
    "plt.hist(normalizer(fps).numpy()[:, 1:5]);"
   ],
   "outputs": [
    {
     "output_type": "display_data",
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {}
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Model definition"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "source": [
    "def nn_model(\n",
    "    dropout0=0.3,\n",
    "    dropout1=0.2,\n",
    "    normalizer=normalizer,\n",
    "    fp_length=56,\n",
    "    n_reagents=reagent_mat.shape[-1],\n",
    "):\n",
    "    fp_inputs = [tfk.Input(shape=(fp_length)) for _ in range(3)]\n",
    "    present_inputs = [tfk.Input(shape=(1,)) for _ in range(3)]\n",
    "    rx_input = tfk.Input(shape=(n_reagents,))\n",
    "\n",
    "    fp_processing = tfk.Sequential(\n",
    "        [\n",
    "            normalizer,\n",
    "            tfk.layers.GaussianNoise(stddev=0.02),\n",
    "            tfk.layers.Dropout(dropout0),\n",
    "            tfk.layers.Dense(32),\n",
    "            tfk.layers.LeakyReLU(),\n",
    "            tfk.layers.Dropout(dropout0),\n",
    "            tfk.layers.Dense(16),\n",
    "            tfk.layers.LeakyReLU(),\n",
    "            tfk.layers.Dropout(dropout1),\n",
    "        ]\n",
    "    )\n",
    "    rx = tfk.layers.Dense(3, activation=\"relu\")(rx_input)\n",
    "\n",
    "    result = tfk.layers.Concatenate()(\n",
    "        [*[fp_processing(fp_inputs[i]) for i in range(3)], *present_inputs, rx,]\n",
    "    )\n",
    "\n",
    "    rest = tfk.Sequential(\n",
    "        [\n",
    "            tfk.layers.Dense(24),\n",
    "            tfk.layers.BatchNormalization(),\n",
    "            tfk.layers.LeakyReLU(),\n",
    "            tfk.layers.Dense(16),\n",
    "            tfk.layers.BatchNormalization(),\n",
    "            tfk.layers.LeakyReLU(),\n",
    "            tfk.layers.Dense(8),\n",
    "            tfk.layers.BatchNormalization(),\n",
    "            tfk.layers.LeakyReLU(),\n",
    "            tfk.layers.Dense(4),\n",
    "            tfk.layers.LeakyReLU(),\n",
    "            tfk.layers.Dense(1, activation=\"sigmoid\"),\n",
    "        ]\n",
    "    )\n",
    "\n",
    "    result = rest(result)\n",
    "\n",
    "    model = tfk.Model(inputs=[*fp_inputs, *present_inputs, rx_input,], outputs=result)\n",
    "\n",
    "    return tfk.Model(\n",
    "        inputs=[*fp_inputs, *present_inputs, rx_input],\n",
    "        outputs=[\n",
    "            model([*perm_fps, *perm_present, rx_input])\n",
    "            for perm_fps, perm_present in zip(\n",
    "                itertools.permutations(fp_inputs, 3),\n",
    "                itertools.permutations(present_inputs, 3),\n",
    "            )\n",
    "        ],\n",
    "    )"
   ],
   "outputs": [],
   "metadata": {
    "tags": []
   }
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "source": [
    "mdl = nn_model()\n",
    "\n",
    "mdl.compile(optimizer=\"adam\", loss=\"mse\", metrics=[\"mae\"])\n",
    "\n",
    "history = mdl.fit(\n",
    "    x=[\n",
    "        *reactant_train.transpose(1, 0, 2),\n",
    "        *present_train.transpose(1, 0),\n",
    "        reagent_train,\n",
    "    ],\n",
    "    y=[pred_train] * factorial(3),\n",
    "    sample_weight=weights_train,\n",
    "    batch_size=8,\n",
    "    validation_data=(\n",
    "        [\n",
    "            *reactant_test.transpose(1, 0, 2),\n",
    "            *present_test.transpose(1, 0),\n",
    "            reagent_test,\n",
    "        ],\n",
    "        [pred_test] * factorial(3),\n",
    "        weights_test,\n",
    "    ),\n",
    "    callbacks=[EarlyStopping(patience=100, verbose=1, restore_best_weights=True)],\n",
    "    epochs=4000,\n",
    "    verbose=0,\n",
    ")"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "Restoring model weights from the end of the best epoch.\n",
      "Epoch 00528: early stopping\n"
     ]
    }
   ],
   "metadata": {
    "tags": []
   }
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "source": [
    "plt.plot(history.history[\"loss\"], label=\"train\")\n",
    "plt.plot(history.history[\"val_loss\"], label=\"test\")\n",
    "plt.legend()\n",
    "plt.xlabel(\"Epoch #\")\n",
    "plt.ylabel(\"Loss (MSE)\")"
   ],
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Loss (MSE)')"
      ]
     },
     "metadata": {},
     "execution_count": 21
    },
    {
     "output_type": "display_data",
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {}
    }
   ],
   "metadata": {
    "tags": []
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "### Evaluating model accuracy"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "source": [
    "pred = mdl(\n",
    "    [*reactant_test.transpose(1, 0, 2), *present_test.transpose(1, 0), reagent_test]\n",
    ")\n",
    "\n",
    "pred = np.stack(pred).mean(axis=(0, 2))\n",
    "delta = pred_test - pred\n",
    "\n",
    "{\n",
    "    \"Mean squared error (baseline)\": np.mean(delta * delta),\n",
    "    \"Mean absolute error (baseline)\": np.mean(np.abs(delta)),\n",
    "    \"Mean squared error (baseline, weighted)\": np.mean(weights_test * delta * delta),\n",
    "    \"Mean absolute error (baseline, weighted)\": np.mean(weights_test * np.abs(delta)),\n",
    "}"
   ],
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "{'Mean squared error (baseline)': 0.04184484,\n",
       " 'Mean absolute error (baseline)': 0.16380449,\n",
       " 'Mean squared error (baseline, weighted)': 0.04097987653828008,\n",
       " 'Mean absolute error (baseline, weighted)': 0.16728816836219026}"
      ]
     },
     "metadata": {},
     "execution_count": 22
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "#### Comparison to baseline accuracy\n",
    "As a baseline model we propose one with a constant prediction equal to the average reactivity in the training set"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "source": [
    "baseline_pred = pred_train.mean()\n",
    "\n",
    "delta = pred_test - baseline_pred\n",
    "\n",
    "{\n",
    "    \"Mean squared error (baseline)\": np.mean(delta * delta),\n",
    "    \"Mean absolute error (baseline)\": np.mean(np.abs(delta)),\n",
    "    \"Mean squared error (baseline, weighted)\": np.mean(weights_test * delta * delta),\n",
    "    \"Mean absolute error (baseline, weighted)\": np.mean(weights_test * np.abs(delta)),\n",
    "}"
   ],
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "{'Mean squared error (baseline)': 0.07958939,\n",
       " 'Mean absolute error (baseline)': 0.24156313,\n",
       " 'Mean squared error (baseline, weighted)': 0.11189074322693884,\n",
       " 'Mean absolute error (baseline, weighted)': 0.2889966966332564}"
      ]
     },
     "metadata": {},
     "execution_count": 23
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "#### Comparing to dummy model\n",
    "The dummy model predicts the most common class (0: unreactive) for all observations"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "source": [
    "dummy_pred = np.zeros_like(pred_test)\n",
    "\n",
    "delta = pred_test - dummy_pred\n",
    "\n",
    "{\n",
    "    \"Mean squared error (baseline)\": np.mean(delta * delta),\n",
    "    \"Mean absolute error (baseline)\": np.mean(np.abs(delta)),\n",
    "    \"Mean squared error (baseline, weighted)\": np.mean(weights_test * delta * delta),\n",
    "    \"Mean absolute error (baseline, weighted)\": np.mean(weights_test * np.abs(delta)),\n",
    "}"
   ],
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "{'Mean squared error (baseline)': 0.22068562,\n",
       " 'Mean absolute error (baseline)': 0.37645924,\n",
       " 'Mean squared error (baseline, weighted)': 0.3591672072836551,\n",
       " 'Mean absolute error (baseline, weighted)': 0.5402900532565874}"
      ]
     },
     "metadata": {},
     "execution_count": 24
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "### Predicted–expectation correlation"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "source": [
    "dfs = []\n",
    "outputs = [\n",
    "    mdl(\n",
    "        [\n",
    "            *reactant_test.transpose(1, 0, 2),\n",
    "            *present_test.transpose(1, 0),\n",
    "            reagent_test,\n",
    "        ],\n",
    "        training=True,\n",
    "    )\n",
    "    for _ in range(10)\n",
    "]\n",
    "for j, output in enumerate(outputs):\n",
    "    for i, o in enumerate(output):\n",
    "        dfs.append(\n",
    "            pd.DataFrame(\n",
    "                {\n",
    "                    \"point\": np.arange(len(o)),\n",
    "                    \"permutation\": i,\n",
    "                    \"sample\": j,\n",
    "                    \"expected\": pred_test,\n",
    "                    \"predicted\": o.numpy().flatten(),\n",
    "                }\n",
    "            )\n",
    "        )\n",
    "results = pd.concat(dfs, ignore_index=True)"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "source": [
    "f = sns.lmplot(\n",
    "    data=results,\n",
    "    x=\"expected\",\n",
    "    y=\"predicted\",\n",
    "    col=\"permutation\",\n",
    "    hue=\"permutation\",\n",
    "    col_wrap=3,\n",
    "    x_jitter=0.02,\n",
    "    scatter_kws={\"s\": 20},\n",
    "    x_estimator=np.mean,\n",
    "    sharex=False,\n",
    "    sharey=False,\n",
    ")\n",
    "\n",
    "f.savefig(f\"Model stats - {active_space}.svg\")"
   ],
   "outputs": [
    {
     "output_type": "display_data",
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1080x720 with 6 Axes>"
      ]
     },
     "metadata": {}
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "source": [
    "palette = sns.mpl_palette(\"tab10\", 10)\n",
    "perm0 = results[results.permutation == 0]\n",
    "grp = perm0.groupby(\"point\")\n",
    "f, a = plt.subplots(figsize=(4.9, 4.6))\n",
    "\n",
    "a.errorbar(\n",
    "    grp.mean().expected,\n",
    "    grp.mean().predicted,\n",
    "    yerr=grp.std().predicted,\n",
    "    fmt=\".\",\n",
    "    ecolor=\"k\",\n",
    "    elinewidth=1.5,\n",
    "    capsize=3,\n",
    ")\n",
    "\n",
    "sns.regplot(\n",
    "    data=perm0,\n",
    "    x=\"expected\",\n",
    "    y=\"predicted\",\n",
    "    x_jitter=0.02,\n",
    "    ax=a,\n",
    "    scatter=False,\n",
    "    line_kws={\"color\": palette[1]},\n",
    ")\n",
    "\n",
    "a.set_xlim(0, 1)\n",
    "a.set_ylim(0, 1)\n",
    "a.set_xlabel(\"Observed (via Reactify)\")\n",
    "a.set_ylabel(\"Predicted\")\n",
    "\n",
    "f.savefig(f\"Model stats 2 - {active_space}.svg\", transparent=True, bbox_inches=\"tight\")"
   ],
   "outputs": [
    {
     "output_type": "display_data",
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 352.8x331.2 with 1 Axes>"
      ]
     },
     "metadata": {}
    }
   ],
   "metadata": {
    "tags": []
   }
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "source": [
    "results[results.permutation == 0]"
   ],
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>point</th>\n",
       "      <th>permutation</th>\n",
       "      <th>sample</th>\n",
       "      <th>expected</th>\n",
       "      <th>predicted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.179568</td>\n",
       "      <td>0.132643</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.613760</td>\n",
       "      <td>0.458202</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.005880</td>\n",
       "      <td>0.287438</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.326243</td>\n",
       "      <td>0.274094</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.747098</td>\n",
       "      <td>0.763883</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5605</th>\n",
       "      <td>97</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>0.058275</td>\n",
       "      <td>0.329135</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5606</th>\n",
       "      <td>98</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>0.572912</td>\n",
       "      <td>0.375411</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5607</th>\n",
       "      <td>99</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>0.742973</td>\n",
       "      <td>0.515639</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5608</th>\n",
       "      <td>100</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>0.004380</td>\n",
       "      <td>0.102145</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5609</th>\n",
       "      <td>101</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>0.806563</td>\n",
       "      <td>0.557737</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1020 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      point  permutation  sample  expected  predicted\n",
       "0         0            0       0  0.179568   0.132643\n",
       "1         1            0       0  0.613760   0.458202\n",
       "2         2            0       0  0.005880   0.287438\n",
       "3         3            0       0  0.326243   0.274094\n",
       "4         4            0       0  0.747098   0.763883\n",
       "...     ...          ...     ...       ...        ...\n",
       "5605     97            0       9  0.058275   0.329135\n",
       "5606     98            0       9  0.572912   0.375411\n",
       "5607     99            0       9  0.742973   0.515639\n",
       "5608    100            0       9  0.004380   0.102145\n",
       "5609    101            0       9  0.806563   0.557737\n",
       "\n",
       "[1020 rows x 5 columns]"
      ]
     },
     "metadata": {},
     "execution_count": 28
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "### Confusion matrix"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "source": [
    "f, a = plt.subplots(figsize=(4, 4))\n",
    "expected_classes = np.round(pred_test * 4)\n",
    "out_classes = np.round(\n",
    "    mdl(\n",
    "        [*reactant_test.transpose(1, 0, 2), *present_test.transpose(1, 0), reagent_test]\n",
    "    )[0]\n",
    "    * 4\n",
    ")\n",
    "cm = confusion_matrix(expected_classes, out_classes, labels=range(4))\n",
    "ax = sns.heatmap(cm, square=True, annot=True, ax=a, cbar=False)\n",
    "ax.set_xlabel(\"Predicted class\")\n",
    "ax.set_ylabel(\"True class\");"
   ],
   "outputs": [
    {
     "output_type": "display_data",
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {}
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Error-uncertainty correlation"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "source": [
    "results[\"diff\"] = (results[\"predicted\"] - results[\"expected\"]).abs()\n",
    "results[\"diff2\"] = results[\"diff\"] ** 2\n",
    "grp = results.groupby([\"point\", \"permutation\"])"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "source": [
    "correlation_df = pd.DataFrame(\n",
    "    {\n",
    "        \"Error (mse)\": grp[\"diff\"].mean(),\n",
    "        \"Error (mae)\": grp[\"diff2\"].mean(),\n",
    "        \"Uncertainty\": grp[\"predicted\"].std(),\n",
    "    },\n",
    ")"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "source": [
    "# c = correlation_df.loc[(slice(None), 5), :]\n",
    "c = correlation_df\n",
    "sns.lmplot(\n",
    "    \"Uncertainty\", \"Error (mse)\", data=c, scatter_kws={\"s\": 10, \"color\": \"grey\"},\n",
    ")\n",
    "\n",
    "plt.arrow(0, -0.02, 0.22, 0.13, width=0.002, head_width=0.0, fc=\"grey\", ls=\":\")\n",
    "\n",
    "plt.savefig(\n",
    "    \"Error-uncertainty correlation.png\",\n",
    "    transparent=True,\n",
    "    dpi=200,\n",
    "    pad_inches=0.0,\n",
    "    bbox_inches=\"tight\",\n",
    ")"
   ],
   "outputs": [
    {
     "output_type": "display_data",
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {}
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "source": [
    "error_perm = np.argsort(c[\"Error (mse)\"])\n",
    "uncertainty_perm = np.argsort(c[\"Uncertainty\"])\n",
    "cov = np.cov(np.stack([error_perm, uncertainty_perm]))"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "### Spearman correlation coefficient"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "source": [
    "cov[0, 1] / cov[0, 0]"
   ],
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "0.05761196141095975"
      ]
     },
     "metadata": {},
     "execution_count": 55
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Benchmarking on an imaginary chemical space"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "source": [
    "synthetic_reactants = {\n",
    "    \"cyclopenten-1-ylboronic acid\": \"OB(O)C1=CCCC1\",\n",
    "    \"Indole-3-carboxaldehyde\": \"O=Cc1c[nH]c2ccccc12\",\n",
    "    \"4-nitroaniline\": \"Nc1ccc(cc1)[N+]([O-])=O\",\n",
    "    \"3,4-Dimethylbenzaldehyde\": \"Cc1ccc(C=O)cc1C\",\n",
    "    \"5-chloro-3-(bromomethyl)-1,2,4-thiadiazole\": \"BrCc1nsc(Cl)n1\",\n",
    "    \"vinyl acetate\": \"C=COC(C)=O\",\n",
    "}\n",
    "\n",
    "synthetic_fps = fp_calculator(list(synthetic_reactants.values()))"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stderr",
     "text": [
      "100%|██████████| 1/1 [00:00<00:00,  8.93it/s]\n"
     ]
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "source": [
    "# binary and three-component combinations\n",
    "combinations = [itertools.combinations(range(len(synthetic_fps)), n) for n in [2, 3]]\n",
    "all_combinations = []\n",
    "for c in combinations:\n",
    "    all_combinations += list(c)"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "source": [
    "synthetic_reactant_fps = np.zeros(\n",
    "    (len(all_combinations), 3, synthetic_fps.shape[1]), dtype=\"float32\"\n",
    ")\n",
    "synthetic_reagents = np.zeros((len(all_combinations), 5), dtype=\"float32\")\n",
    "synthetic_present = np.zeros((len(all_combinations), 3), dtype=\"float32\")\n",
    "\n",
    "for i, comb in enumerate(all_combinations):\n",
    "    comb = list(comb)\n",
    "    synthetic_reactant_fps[i, : len(comb)] = synthetic_fps[comb]\n",
    "    synthetic_present[i, : len(comb)] = 1.0"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "source": [
    "predictions_mat = np.concatenate(\n",
    "    [\n",
    "        mdl(\n",
    "            [\n",
    "                *synthetic_reactant_fps.transpose(1, 0, 2),\n",
    "                *synthetic_present.transpose(1, 0),\n",
    "                synthetic_reagents,\n",
    "            ],\n",
    "            training=True\n",
    "        )\n",
    "        for _ in range(100)\n",
    "    ],\n",
    "    axis=0,\n",
    ")"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "source": [
    "predictions_df = pd.DataFrame(predictions_mat.squeeze())"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "source": [
    "predictions_df.rename(\n",
    "    columns={n: \"+\".join(str(i) for i in comb) for n, comb in enumerate(all_combinations)},\n",
    "    inplace=True,\n",
    ")"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "source": [
    "pd.DataFrame(\n",
    "    {\n",
    "        \"name\": list(synthetic_reactants.keys()),\n",
    "        \"smiles\": list(synthetic_reactants.values()),\n",
    "    },\n",
    ")"
   ],
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>name</th>\n",
       "      <th>smiles</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>cyclopenten-1-ylboronic acid</td>\n",
       "      <td>OB(O)C1=CCCC1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Indole-3-carboxaldehyde</td>\n",
       "      <td>O=Cc1c[nH]c2ccccc12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4-nitroaniline</td>\n",
       "      <td>Nc1ccc(cc1)[N+]([O-])=O</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3,4-Dimethylbenzaldehyde</td>\n",
       "      <td>Cc1ccc(C=O)cc1C</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5-chloro-3-(bromomethyl)-1,2,4-thiadiazole</td>\n",
       "      <td>BrCc1nsc(Cl)n1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>vinyl acetate</td>\n",
       "      <td>C=COC(C)=O</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                         name                   smiles\n",
       "0                cyclopenten-1-ylboronic acid            OB(O)C1=CCCC1\n",
       "1                     Indole-3-carboxaldehyde      O=Cc1c[nH]c2ccccc12\n",
       "2                              4-nitroaniline  Nc1ccc(cc1)[N+]([O-])=O\n",
       "3                    3,4-Dimethylbenzaldehyde          Cc1ccc(C=O)cc1C\n",
       "4  5-chloro-3-(bromomethyl)-1,2,4-thiadiazole           BrCc1nsc(Cl)n1\n",
       "5                               vinyl acetate               C=COC(C)=O"
      ]
     },
     "metadata": {},
     "execution_count": 65
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "source": [
    "f, a = plt.subplots(figsize=(24, 4))\n",
    "sns.boxplot(data=predictions_df, ax=a);"
   ],
   "outputs": [
    {
     "output_type": "display_data",
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1728x288 with 1 Axes>"
      ]
     },
     "metadata": {}
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "# Chemical space exploration"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "source": [
    "class Explorer:\n",
    "    def __init__(self):\n",
    "        pass\n",
    "\n",
    "    def train(\n",
    "        self,\n",
    "        reactant_train,\n",
    "        present_train,\n",
    "        reagent_train,\n",
    "        pred_train,\n",
    "        callbacks=None,\n",
    "        weights_train=None,\n",
    "        fit_kwargs=None,\n",
    "    ):\n",
    "        self.model = nn_model(normalizer=normalizer)\n",
    "        self.model.compile(optimizer=\"adam\", loss=\"mse\", metrics=[\"mae\"])\n",
    "        if callbacks is None:\n",
    "            callbacks = [\n",
    "                EarlyStopping(patience=20, verbose=0, restore_best_weights=True)\n",
    "            ]\n",
    "        self.model.fit(\n",
    "            x=[\n",
    "                *reactant_train.transpose(1, 0, 2),\n",
    "                *present_train.transpose(1, 0),\n",
    "                reagent_train,\n",
    "            ],\n",
    "            y=[pred_train] * factorial(3),\n",
    "            sample_weight=weights_train,\n",
    "            batch_size=8,\n",
    "            validation_data=(\n",
    "                [\n",
    "                    *reactant_test.transpose(1, 0, 2),\n",
    "                    *present_test.transpose(1, 0),\n",
    "                    reagent_test,\n",
    "                ],\n",
    "                [pred_test] * factorial(3),\n",
    "                weights_test,\n",
    "            ),\n",
    "            callbacks=callbacks,\n",
    "            epochs=4000,\n",
    "            verbose=0,\n",
    "            **(fit_kwargs or {})\n",
    "        )\n",
    "\n",
    "    def test(self, reactant_test, present_test, reagent_test):\n",
    "        return np.mean(\n",
    "            np.concatenate(\n",
    "                self.model(\n",
    "                    [\n",
    "                        *reactant_test.transpose(1, 0, 2),\n",
    "                        *present_test.transpose(1, 0),\n",
    "                        reagent_test,\n",
    "                    ]\n",
    "                ),\n",
    "                axis=1,\n",
    "            ),\n",
    "            axis=1,\n",
    "        )"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "source": [
    "class ChemicalSpace:\n",
    "    \"\"\"\n",
    "    Class to manage the simulated chemical space. It keeps track of experiments that\n",
    "    have been seen or unseen by the algorithm.\n",
    "    methods:\n",
    "    random_guess: assign a random percentage to the seen reactions\n",
    "    update_explored_indexes: assign the selected indexes to the seen reactions\n",
    "    update_df: based on seen/unseen indexes it splits the complete dataset\n",
    "    accordingly\n",
    "    \"\"\"\n",
    "\n",
    "    def __init__(\n",
    "        self,\n",
    "        reactant_fps,\n",
    "        present_mat,\n",
    "        reagent_mat,\n",
    "        predictions,\n",
    "        sample_weights,\n",
    "        seed=2018,\n",
    "    ):\n",
    "        random.seed(seed)\n",
    "\n",
    "        self.all_reactant_fps = reactant_fps\n",
    "        self.all_present_mat = present_mat\n",
    "        self.all_reagent_mat = reagent_mat\n",
    "        self.all_predictions = predictions\n",
    "        self.all_sample_weights = sample_weights\n",
    "\n",
    "        self.unseen = list(range(len(self.all_reactant_fps)))\n",
    "        self.seen = []\n",
    "        self.space_len = len(self.unseen)\n",
    "\n",
    "        random.shuffle(self.unseen)\n",
    "\n",
    "    def random_guess(self, reveal_fraction: float):\n",
    "        num_of_rxns = int(self.space_len * reveal_fraction)\n",
    "        random_idxs = []\n",
    "\n",
    "        for idx in self.unseen:\n",
    "            self.seen.append(idx)\n",
    "            random_idxs.append(idx)\n",
    "            if len(random_idxs) == num_of_rxns:\n",
    "                break\n",
    "        for idx in random_idxs:\n",
    "            self.unseen.remove(idx)\n",
    "\n",
    "        self.update_df()\n",
    "\n",
    "    def update_explored_indexes(self, indexes):\n",
    "        for idx in indexes:\n",
    "            self.seen.append(idx)\n",
    "            self.unseen.remove(idx)\n",
    "        self.update_df()\n",
    "\n",
    "    def update_df(self):\n",
    "        self.reactant_fps = self.all_reactant_fps[self.seen]\n",
    "        self.present_mat = self.all_present_mat[self.seen]\n",
    "        self.reagent_mat = self.all_reagent_mat[self.seen]\n",
    "        self.predictions = self.all_predictions[self.seen]\n",
    "        self.sample_weights = self.all_sample_weights[self.seen]\n",
    "\n",
    "        self.unseen_reactant_fps = self.all_reactant_fps[self.unseen]\n",
    "        self.unseen_present_mat = self.all_present_mat[self.unseen]\n",
    "        self.unseen_reagent_mat = self.all_reagent_mat[self.unseen]\n",
    "        self.unseen_predictions = self.all_predictions[self.unseen]\n",
    "        self.unseen_sample_weights = self.all_sample_weights[self.unseen]\n",
    "\n",
    "\n",
    "class Simulation:\n",
    "    \"\"\"\n",
    "    class used for the simulated exploration of the chemical space. It trains the\n",
    "    neural netowork on pregressively bigger datasets and tries to explore the space\n",
    "    performing the reactive combinations first.\n",
    "    \"\"\"\n",
    "\n",
    "    def __init__(\n",
    "        self,\n",
    "        chemical_space: ChemicalSpace,\n",
    "        reveal_fraction=0.1,\n",
    "        screen_size=50,\n",
    "        monitor_reactions=None,\n",
    "    ):\n",
    "        self.chemical_space = chemical_space\n",
    "        self.reveal_fraction = reveal_fraction\n",
    "        self.screen_size = screen_size\n",
    "        self.model = Explorer()\n",
    "        self.results = []\n",
    "        if monitor_reactions:\n",
    "            self.monitor_reactions = {i: -1 for i in monitor_reactions}\n",
    "        else:\n",
    "            self.monitor_reactions = None\n",
    "\n",
    "    def explore_space(self):\n",
    "        # defines the batch size for exploration of chemical space\n",
    "        screen_size = 50\n",
    "        # randomly guess select 10 % of chemical space\n",
    "        self.chemical_space.random_guess(self.reveal_fraction)\n",
    "\n",
    "        # keep statistics for plotting\n",
    "        rxn = [0.5 * len(self.chemical_space.seen)]\n",
    "        width = [0.95 * len(self.chemical_space.seen)]\n",
    "        self.results.append([i for i in self.chemical_space.predictions])\n",
    "\n",
    "        # Evaluate average yield of selected reactions\n",
    "        avg_real_yield = [np.mean(self.chemical_space.predictions)]\n",
    "        # Evaluate standard deviation\n",
    "        stds = [np.std(self.chemical_space.predictions)]\n",
    "\n",
    "        iteration = 0\n",
    "\n",
    "        while not len(self.chemical_space.unseen) == 0:\n",
    "            # Add data to plot\n",
    "            if iteration == 0:\n",
    "                rxn.append(2 * rxn[-1] + 0.5 * screen_size)\n",
    "            else:\n",
    "                if len(self.chemical_space.unseen) < screen_size:\n",
    "                    screen_size = len(self.chemical_space.unseen)\n",
    "                rxn.append(rxn[-1] + screen_size)\n",
    "\n",
    "            width.append(0.95 * screen_size)\n",
    "\n",
    "            # train neural network with seen data\n",
    "            print(f\"Training on {len(self.chemical_space.seen)}\")\n",
    "            self.model.train(\n",
    "                self.chemical_space.reactant_fps,\n",
    "                self.chemical_space.present_mat,\n",
    "                self.chemical_space.reagent_mat,\n",
    "                self.chemical_space.predictions,\n",
    "                weights_train=self.chemical_space.sample_weights,\n",
    "            )\n",
    "\n",
    "            # assign reactivity to unseen data\n",
    "            self.yp = self.model.test(\n",
    "                self.chemical_space.unseen_reactant_fps,\n",
    "                self.chemical_space.unseen_present_mat,\n",
    "                self.chemical_space.unseen_reagent_mat,\n",
    "            )\n",
    "\n",
    "            # Sort the dataframe by predicted yield\n",
    "            unseen_df = pd.DataFrame(\n",
    "                {\n",
    "                    \"prediction\": self.yp,\n",
    "                    \"indexes\": self.chemical_space.unseen,\n",
    "                    \"real_react\": self.chemical_space.unseen_predictions,\n",
    "                }\n",
    "            ).sort_values(by=[\"prediction\"], ascending=False)\n",
    "            # Get a dataframe with best candidates from prediction\n",
    "            self.best_results = unseen_df.head(self.screen_size)\n",
    "            # Get idxs of best candidates\n",
    "            best_results_idxs = list(self.best_results[\"indexes\"])\n",
    "            # Evaluate real reactivitity of selected batch of reactions\n",
    "            avg_real_yield.append(np.mean(self.best_results[\"real_react\"]))\n",
    "            # Calcuate and print the standard deviation for current batch of reactions\n",
    "            stds.append(np.std(self.best_results[\"real_react\"]))\n",
    "            self.results.append(list(self.best_results[\"real_react\"]))\n",
    "            # Add the current batch of reactions to explored reactions so in\n",
    "            # the next iteration neural network can be trained on updated data\n",
    "            self.chemical_space.update_explored_indexes(best_results_idxs)\n",
    "\n",
    "            if self.monitor_reactions:\n",
    "                for rxn_no in self.monitor_reactions:\n",
    "                    if (\n",
    "                        rxn_no in self.chemical_space.seen\n",
    "                        and self.monitor_reactions[rxn_no] < 0\n",
    "                    ):\n",
    "                        self.monitor_reactions[rxn_no] = iteration\n",
    "\n",
    "                # terminate search if all reactions of interest found\n",
    "                for val in self.monitor_reactions.values():\n",
    "                    if val < 0:\n",
    "                        # not found yet; continue exploring\n",
    "                        break\n",
    "                else:\n",
    "                    # all reactions of interest found; terminate search\n",
    "                    return rxn, avg_real_yield, stds, width, self.results\n",
    "            iteration += 1\n",
    "\n",
    "        return rxn, avg_real_yield, stds, width, self.results"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "source": [
    "cs = ChemicalSpace(\n",
    "    reactant_fps=reaction_fps,\n",
    "    present_mat=present_mat,\n",
    "    reagent_mat=reagent_mat,\n",
    "    predictions=predictions,\n",
    "    sample_weights=sample_weights,\n",
    ")"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "source": [
    "simulation = Simulation(cs)"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "source": [
    "rxn, avg_real_yield, stds, width, results = simulation.explore_space()"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "Training on 551\n",
      "Training on 601\n",
      "Training on 651\n",
      "Training on 701\n",
      "Training on 751\n",
      "Training on 801\n",
      "Training on 851\n",
      "Training on 901\n",
      "Training on 951\n",
      "Training on 1001\n"
     ]
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "source": [
    "plt.figure(figsize=(4.9, 4.6))\n",
    "colors = [\"steelblue\" for i in range(len(rxn))]\n",
    "colors[0] = \"coral\"\n",
    "plt.bar(\n",
    "    rxn[:-1], avg_real_yield[:-1], yerr=stds[:-1], width=width[:-1], color=colors[:-1],\n",
    ")\n",
    "# plt.title(\"Average reactivity per reactions batch\")\n",
    "plt.xlabel(\"Reaction Number\")\n",
    "plt.ylabel(\"Reactivity\")\n",
    "plt.ylim(0, 1)\n",
    "plt.savefig(\n",
    "    f\"Simulated exploration {active_space}.svg\", transparent=True, bbox_inches=\"tight\"\n",
    ")"
   ],
   "outputs": [
    {
     "output_type": "display_data",
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 352.8x331.2 with 1 Axes>"
      ]
     },
     "metadata": {}
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "### Simulating the discovery of reactions of interest"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "source": [
    "reactions_of_interest = [\n",
    "    50,  # Formation of trimer #29\n",
    "    187,  # Photochemical reaction forming compound #28\n",
    "]"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "source": [
    "data.loc[reactions_of_interest]"
   ],
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>bromoacetonitrile</th>\n",
       "      <th>bromomalonate</th>\n",
       "      <th>diethyl_thiobarbituric_acid</th>\n",
       "      <th>dimethyl_nitrosoaniline</th>\n",
       "      <th>dimethylisoxazole</th>\n",
       "      <th>ethyl_vinyl_ether</th>\n",
       "      <th>glycidyl_propargyl_ether</th>\n",
       "      <th>hexadienal</th>\n",
       "      <th>indole</th>\n",
       "      <th>phenylboronic_acid</th>\n",
       "      <th>...</th>\n",
       "      <th>phenylurazole</th>\n",
       "      <th>thiadiazole</th>\n",
       "      <th>tosyl_isocyanide</th>\n",
       "      <th>vinyl_pyrrolidinone</th>\n",
       "      <th>base</th>\n",
       "      <th>acid</th>\n",
       "      <th>pyrylium</th>\n",
       "      <th>rose_bengal</th>\n",
       "      <th>rubypy</th>\n",
       "      <th>prediction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>50</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.819559</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>187</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.877370</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     bromoacetonitrile  bromomalonate  diethyl_thiobarbituric_acid  \\\n",
       "50                   0              1                            0   \n",
       "187                  1              0                            0   \n",
       "\n",
       "     dimethyl_nitrosoaniline  dimethylisoxazole  ethyl_vinyl_ether  \\\n",
       "50                         0                  0                  0   \n",
       "187                        0                  0                  0   \n",
       "\n",
       "     glycidyl_propargyl_ether  hexadienal  indole  phenylboronic_acid  ...  \\\n",
       "50                          0           0       0                   0  ...   \n",
       "187                         0           0       0                   0  ...   \n",
       "\n",
       "     phenylurazole  thiadiazole  tosyl_isocyanide  vinyl_pyrrolidinone  base  \\\n",
       "50               0            0                 1                    0     0   \n",
       "187              0            0                 0                    0     0   \n",
       "\n",
       "     acid  pyrylium  rose_bengal  rubypy  prediction  \n",
       "50      0         1            0       0    0.819559  \n",
       "187     0         0            0       1    0.877370  \n",
       "\n",
       "[2 rows x 21 columns]"
      ]
     },
     "metadata": {},
     "execution_count": 79
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "source": [
    "found = {i: [] for i in reactions_of_interest}\n",
    "\n",
    "for i in range(100):\n",
    "    print(f\"---------------ITERATION {i}-------------------\")\n",
    "    print(found)\n",
    "    cs = ChemicalSpace(\n",
    "        reactant_fps=reaction_fps,\n",
    "        present_mat=present_mat,\n",
    "        reagent_mat=reagent_mat,\n",
    "        predictions=predictions,\n",
    "        sample_weights=sample_weights,\n",
    "        seed=i,\n",
    "    )\n",
    "    sim = Simulation(cs, monitor_reactions=reactions_of_interest)\n",
    "    rxn, avg_real_yield, stds, width, results = sim.explore_space()\n",
    "    for rxn in found:\n",
    "        found[rxn].append(sim.monitor_reactions[rxn])"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "---------------ITERATION 0-------------------\n",
      "{50: [], 187: []}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "---------------ITERATION 1-------------------\n",
      "{50: [4], 187: [5]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "---------------ITERATION 2-------------------\n",
      "{50: [4, 5], 187: [5, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "---------------ITERATION 3-------------------\n",
      "{50: [4, 5, 7], 187: [5, 2, 4]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "Training on 551\n",
      "---------------ITERATION 4-------------------\n",
      "{50: [4, 5, 7, 9], 187: [5, 2, 4, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "---------------ITERATION 5-------------------\n",
      "{50: [4, 5, 7, 9, 6], 187: [5, 2, 4, 1, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "---------------ITERATION 6-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4], 187: [5, 2, 4, 1, 2, 6]}\n",
      "Training on 101\n",
      "---------------ITERATION 7-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0], 187: [5, 2, 4, 1, 2, 6, 0]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "---------------ITERATION 8-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0], 187: [5, 2, 4, 1, 2, 6, 0, 3]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "---------------ITERATION 9-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "---------------ITERATION 10-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "---------------ITERATION 11-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "---------------ITERATION 12-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "---------------ITERATION 13-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "---------------ITERATION 14-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "---------------ITERATION 15-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "---------------ITERATION 16-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "---------------ITERATION 17-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "---------------ITERATION 18-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "---------------ITERATION 19-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "---------------ITERATION 20-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "---------------ITERATION 21-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "---------------ITERATION 22-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "---------------ITERATION 23-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "---------------ITERATION 24-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "---------------ITERATION 25-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "Training on 551\n",
      "Training on 601\n",
      "---------------ITERATION 26-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "Training on 551\n",
      "---------------ITERATION 27-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "Training on 551\n",
      "---------------ITERATION 28-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "---------------ITERATION 29-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "---------------ITERATION 30-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "---------------ITERATION 31-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "---------------ITERATION 32-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "---------------ITERATION 33-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "---------------ITERATION 34-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "---------------ITERATION 35-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "Training on 551\n",
      "---------------ITERATION 36-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "---------------ITERATION 37-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "---------------ITERATION 38-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "---------------ITERATION 39-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "---------------ITERATION 40-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "---------------ITERATION 41-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "---------------ITERATION 42-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "---------------ITERATION 43-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "---------------ITERATION 44-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "---------------ITERATION 45-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "---------------ITERATION 46-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "Training on 551\n",
      "Training on 601\n",
      "---------------ITERATION 47-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "---------------ITERATION 48-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "Training on 551\n",
      "Training on 601\n",
      "---------------ITERATION 49-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "Training on 551\n",
      "Training on 601\n",
      "---------------ITERATION 50-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "---------------ITERATION 51-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "---------------ITERATION 52-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "---------------ITERATION 53-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "---------------ITERATION 54-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "---------------ITERATION 55-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "Training on 551\n",
      "Training on 601\n",
      "---------------ITERATION 56-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "---------------ITERATION 57-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "---------------ITERATION 58-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "---------------ITERATION 59-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "---------------ITERATION 60-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "---------------ITERATION 61-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "---------------ITERATION 62-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "---------------ITERATION 63-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "Training on 551\n",
      "---------------ITERATION 64-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "---------------ITERATION 65-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "---------------ITERATION 66-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "---------------ITERATION 67-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "---------------ITERATION 68-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "---------------ITERATION 69-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "---------------ITERATION 70-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "---------------ITERATION 71-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "---------------ITERATION 72-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "---------------ITERATION 73-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "---------------ITERATION 74-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "---------------ITERATION 75-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "---------------ITERATION 76-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "---------------ITERATION 77-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "---------------ITERATION 78-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "---------------ITERATION 79-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6, 4], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "---------------ITERATION 80-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6, 4, 6], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1, 2, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "Training on 551\n",
      "---------------ITERATION 81-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6, 4, 6, 9], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1, 2, 2, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "---------------ITERATION 82-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6, 4, 6, 9, 8], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1, 2, 2, 1, 0]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "---------------ITERATION 83-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6, 4, 6, 9, 8, 5], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1, 2, 2, 1, 0, 4]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "Training on 551\n",
      "---------------ITERATION 84-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6, 4, 6, 9, 8, 5, 9], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1, 2, 2, 1, 0, 4, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "---------------ITERATION 85-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6, 4, 6, 9, 8, 5, 9, 5], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1, 2, 2, 1, 0, 4, 2, 3]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "---------------ITERATION 86-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6, 4, 6, 9, 8, 5, 9, 5, 0], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1, 2, 2, 1, 0, 4, 2, 3, 3]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "Training on 551\n",
      "---------------ITERATION 87-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6, 4, 6, 9, 8, 5, 9, 5, 0, 9], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1, 2, 2, 1, 0, 4, 2, 3, 3, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "---------------ITERATION 88-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6, 4, 6, 9, 8, 5, 9, 5, 0, 9, 7], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1, 2, 2, 1, 0, 4, 2, 3, 3, 2, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "---------------ITERATION 89-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6, 4, 6, 9, 8, 5, 9, 5, 0, 9, 7, 6], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1, 2, 2, 1, 0, 4, 2, 3, 3, 2, 1, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "---------------ITERATION 90-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6, 4, 6, 9, 8, 5, 9, 5, 0, 9, 7, 6, 6], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1, 2, 2, 1, 0, 4, 2, 3, 3, 2, 1, 2, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "---------------ITERATION 91-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6, 4, 6, 9, 8, 5, 9, 5, 0, 9, 7, 6, 6, 8], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1, 2, 2, 1, 0, 4, 2, 3, 3, 2, 1, 2, 1, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "---------------ITERATION 92-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6, 4, 6, 9, 8, 5, 9, 5, 0, 9, 7, 6, 6, 8, 7], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1, 2, 2, 1, 0, 4, 2, 3, 3, 2, 1, 2, 1, 2, 0]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "---------------ITERATION 93-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6, 4, 6, 9, 8, 5, 9, 5, 0, 9, 7, 6, 6, 8, 7, 7], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1, 2, 2, 1, 0, 4, 2, 3, 3, 2, 1, 2, 1, 2, 0, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "---------------ITERATION 94-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6, 4, 6, 9, 8, 5, 9, 5, 0, 9, 7, 6, 6, 8, 7, 7, 6], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1, 2, 2, 1, 0, 4, 2, 3, 3, 2, 1, 2, 1, 2, 0, 2, 5]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "---------------ITERATION 95-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6, 4, 6, 9, 8, 5, 9, 5, 0, 9, 7, 6, 6, 8, 7, 7, 6, 3], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1, 2, 2, 1, 0, 4, 2, 3, 3, 2, 1, 2, 1, 2, 0, 2, 5, 1]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "---------------ITERATION 96-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6, 4, 6, 9, 8, 5, 9, 5, 0, 9, 7, 6, 6, 8, 7, 7, 6, 3, 7], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1, 2, 2, 1, 0, 4, 2, 3, 3, 2, 1, 2, 1, 2, 0, 2, 5, 1, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "---------------ITERATION 97-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6, 4, 6, 9, 8, 5, 9, 5, 0, 9, 7, 6, 6, 8, 7, 7, 6, 3, 7, 4], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1, 2, 2, 1, 0, 4, 2, 3, 3, 2, 1, 2, 1, 2, 0, 2, 5, 1, 2, 0]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n",
      "Training on 451\n",
      "Training on 501\n",
      "---------------ITERATION 98-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6, 4, 6, 9, 8, 5, 9, 5, 0, 9, 7, 6, 6, 8, 7, 7, 6, 3, 7, 4, 8], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1, 2, 2, 1, 0, 4, 2, 3, 3, 2, 1, 2, 1, 2, 0, 2, 5, 1, 2, 0, 2]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "---------------ITERATION 99-------------------\n",
      "{50: [4, 5, 7, 9, 6, 4, 0, 0, 7, 8, 8, 3, 8, 6, 8, 8, 4, 7, 7, 5, 0, 6, 6, 8, 3, 10, 9, 9, 8, 6, 5, 6, 8, 8, 5, 9, 0, 7, 5, 3, 5, 7, 6, 7, 6, 4, 10, 7, 10, 10, 7, 5, 6, 7, 7, 10, 2, 0, 4, 3, 7, 8, 8, 9, 5, 0, 4, 4, 5, 5, 0, 6, 3, 7, 2, 6, 4, 6, 4, 6, 9, 8, 5, 9, 5, 0, 9, 7, 6, 6, 8, 7, 7, 6, 3, 7, 4, 8, 5], 187: [5, 2, 4, 1, 2, 6, 0, 3, 0, 0, 2, 3, 0, 0, 4, 3, 1, 2, 1, 2, 6, 3, 4, 2, 1, 1, 3, 2, 1, 2, 3, 1, 4, 1, 2, 2, 5, 2, 0, 2, 4, 1, 5, 2, 2, 4, 0, 6, 1, 1, 1, 2, 1, 0, 2, 0, 1, 2, 3, 1, 5, 0, 2, 3, 4, 1, 2, 1, 2, 3, 1, 0, 0, 0, 3, 2, 5, 1, 2, 2, 1, 0, 4, 2, 3, 3, 2, 1, 2, 1, 2, 0, 2, 5, 1, 2, 0, 2, 0]}\n",
      "Training on 101\n",
      "Training on 151\n",
      "Training on 201\n",
      "Training on 251\n",
      "Training on 301\n",
      "Training on 351\n",
      "Training on 401\n"
     ]
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "Average number of iterations before reaction discovery (mean number of iterations using random search = 10)"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "source": [
    "{rxn: np.mean(iters) for rxn, iters in found.items()}"
   ],
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "{50: 5.82, 187: 2.0}"
      ]
     },
     "metadata": {},
     "execution_count": 81
    }
   ],
   "metadata": {}
  }
 ],
 "metadata": {
  "kernelspec": {
   "name": "python3",
   "display_name": "Python 3.8.0 64-bit ('Rx1st': conda)"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.0"
  },
  "interpreter": {
   "hash": "018d584ab618c879dcf95ac471267a0117a7b6da20a163fd0dfde0cd8f900583"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}