examples/mutag_literals.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from rrgcn import RRGCNEmbedder\n",
    "from torch_geometric.datasets import Entities\n",
    "import torch\n",
    "from catboost import CatBoostClassifier\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import classification_report\n",
    "import rdflib"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "## edited entities.py from PyG that also saves node_features\n",
    "import logging\n",
    "import os\n",
    "import os.path as osp\n",
    "from collections import Counter\n",
    "from typing import Callable, List, Optional\n",
    "\n",
    "import numpy as np\n",
    "import torch\n",
    "\n",
    "from torch_geometric.data import (\n",
    "    Data,\n",
    "    InMemoryDataset,\n",
    "    download_url,\n",
    "    extract_tar,\n",
    ")\n",
    "\n",
    "\n",
    "class Entities(InMemoryDataset):\n",
    "    url = 'https://data.dgl.ai/dataset/{}.tgz'\n",
    "\n",
    "    def __init__(self, root: str, name: str,\n",
    "                 transform: Optional[Callable] = None,\n",
    "                 pre_transform: Optional[Callable] = None):\n",
    "        self.name = name.lower()\n",
    "        assert self.name in ['aifb', 'am', 'mutag', 'bgs']\n",
    "        super().__init__(root, transform, pre_transform)\n",
    "        self.data, self.slices = torch.load(self.processed_paths[0])\n",
    "\n",
    "    @property\n",
    "    def raw_dir(self) -> str:\n",
    "        return osp.join(self.root, self.name, 'raw')\n",
    "\n",
    "    @property\n",
    "    def processed_dir(self) -> str:\n",
    "        return osp.join(self.root, self.name, 'processed')\n",
    "\n",
    "    @property\n",
    "    def num_relations(self) -> int:\n",
    "        return self.data.edge_type.max().item() + 1\n",
    "\n",
    "    @property\n",
    "    def num_classes(self) -> int:\n",
    "        return self.data.train_y.max().item() + 1\n",
    "\n",
    "    @property\n",
    "    def raw_file_names(self) -> List[str]:\n",
    "        return [\n",
    "            f'{self.name}_stripped.nt.gz',\n",
    "            'completeDataset.tsv',\n",
    "            'trainingSet.tsv',\n",
    "            'testSet.tsv',\n",
    "        ]\n",
    "\n",
    "    @property\n",
    "    def processed_file_names(self) -> str:\n",
    "        return 'data.pt'\n",
    "\n",
    "    def download(self):\n",
    "        path = download_url(self.url.format(self.name), self.root)\n",
    "        extract_tar(path, self.raw_dir)\n",
    "        os.unlink(path)\n",
    "\n",
    "    def process(self):\n",
    "        import gzip\n",
    "\n",
    "        import pandas as pd\n",
    "        import rdflib as rdf\n",
    "\n",
    "        graph_file, task_file, train_file, test_file = self.raw_paths\n",
    "\n",
    "        with hide_stdout():\n",
    "            g = rdf.Graph()\n",
    "            with gzip.open(graph_file, 'rb') as f:\n",
    "                g.parse(file=f, format='nt')\n",
    "\n",
    "        freq = Counter(g.predicates())\n",
    "\n",
    "        relations = sorted(set(g.predicates()), key=lambda p: -freq.get(p, 0))\n",
    "        subjects = set(g.subjects())\n",
    "        objects = set(g.objects())\n",
    "        nodes = list(subjects.union(objects))\n",
    "\n",
    "        N = len(nodes)\n",
    "        R = 2 * len(relations)\n",
    "\n",
    "        relations_dict = {rel: i for i, rel in enumerate(relations)}\n",
    "        nodes_dict = {node: i for i, node in enumerate(nodes)}\n",
    "\n",
    "        edges = []\n",
    "        node_features = {}\n",
    "        for s, p, o in g.triples((None, None, None)):\n",
    "            src, dst, rel = nodes_dict[s], nodes_dict[o], relations_dict[p]\n",
    "            edges.append([src, dst, 2 * rel])\n",
    "            edges.append([dst, src, 2 * rel + 1])\n",
    "\n",
    "            # SAVE LITERAL VALUES\n",
    "            if isinstance(o, rdf.Literal):\n",
    "                literal_type = p#type(o.value)\n",
    "                if literal_type not in node_features:\n",
    "                    node_features[literal_type] = [[dst], [o.value]]\n",
    "                else:\n",
    "                    node_features[literal_type][0].append(dst)\n",
    "                    node_features[literal_type][1].append(o.value)\n",
    "\n",
    "        edges = torch.tensor(edges, dtype=torch.long).t().contiguous()\n",
    "        perm = (N * R * edges[0] + R * edges[1] + edges[2]).argsort()\n",
    "        edges = edges[:, perm]\n",
    "\n",
    "        edge_index, edge_type = edges[:2], edges[2]\n",
    "\n",
    "        if self.name == 'am':\n",
    "            label_header = 'label_cateogory'\n",
    "            nodes_header = 'proxy'\n",
    "        elif self.name == 'aifb':\n",
    "            label_header = 'label_affiliation'\n",
    "            nodes_header = 'person'\n",
    "        elif self.name == 'mutag':\n",
    "            label_header = 'label_mutagenic'\n",
    "            nodes_header = 'bond'\n",
    "        elif self.name == 'bgs':\n",
    "            label_header = 'label_lithogenesis'\n",
    "            nodes_header = 'rock'\n",
    "\n",
    "        labels_df = pd.read_csv(task_file, sep='\\t')\n",
    "        labels_set = set(labels_df[label_header].values.tolist())\n",
    "        labels_dict = {lab: i for i, lab in enumerate(list(labels_set))}\n",
    "        nodes_dict = {np.unicode(key): val for key, val in nodes_dict.items()}\n",
    "\n",
    "        train_labels_df = pd.read_csv(train_file, sep='\\t')\n",
    "        train_indices, train_labels = [], []\n",
    "        for nod, lab in zip(train_labels_df[nodes_header].values,\n",
    "                            train_labels_df[label_header].values):\n",
    "            train_indices.append(nodes_dict[nod])\n",
    "            train_labels.append(labels_dict[lab])\n",
    "\n",
    "        train_idx = torch.tensor(train_indices, dtype=torch.long)\n",
    "        train_y = torch.tensor(train_labels, dtype=torch.long)\n",
    "\n",
    "        test_labels_df = pd.read_csv(test_file, sep='\\t')\n",
    "        test_indices, test_labels = [], []\n",
    "        for nod, lab in zip(test_labels_df[nodes_header].values,\n",
    "                            test_labels_df[label_header].values):\n",
    "            test_indices.append(nodes_dict[nod])\n",
    "            test_labels.append(labels_dict[lab])\n",
    "\n",
    "        test_idx = torch.tensor(test_indices, dtype=torch.long)\n",
    "        test_y = torch.tensor(test_labels, dtype=torch.long)\n",
    "\n",
    "        data = Data(edge_index=edge_index, edge_type=edge_type,\n",
    "                    train_idx=train_idx, train_y=train_y, test_idx=test_idx,\n",
    "                    test_y=test_y, num_nodes=N, node_features=node_features)\n",
    "\n",
    "        torch.save(self.collate([data]), self.processed_paths[0])\n",
    "\n",
    "    def __repr__(self) -> str:\n",
    "        return f'{self.name.upper()}{self.__class__.__name__}()'\n",
    "\n",
    "\n",
    "class hide_stdout(object):\n",
    "    def __enter__(self):\n",
    "        self.level = logging.getLogger().level\n",
    "        logging.getLogger().setLevel(logging.ERROR)\n",
    "\n",
    "    def __exit__(self, *args):\n",
    "        logging.getLogger().setLevel(self.level)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "device = \"cuda\" if torch.cuda.is_available() else \"cpu\"\n",
    "dataset = Entities('./', 'mutag')\n",
    "data = dataset[0].to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "charge = rdflib.URIRef(\"http://dl-learner.org/carcinogenesis#charge\")\n",
    "filtered_node_features = {}\n",
    "filtered_node_features[0] = [None, None]\n",
    "filtered_node_features[0][0] = torch.tensor(\n",
    "    data.node_features[charge][0], dtype=torch.long\n",
    ")\n",
    "filtered_node_features[0][1] = torch.tensor(\n",
    "    data.node_features[charge][1], dtype=torch.float32\n",
    ").reshape(-1, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_idx, val_idx, y_train, y_val = train_test_split(\n",
    "    data.train_idx,\n",
    "    data.train_y,\n",
    "    stratify=data.train_y.cpu().numpy(),\n",
    "    test_size=0.1,\n",
    "    random_state=42,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "embedder = RRGCNEmbedder(\n",
    "    num_nodes=data.num_nodes,\n",
    "    num_relations=dataset.num_relations,\n",
    "    num_layers=5,\n",
    "    emb_size=2048,\n",
    "    device=device,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1/1 [00:13<00:00, 13.18s/it]\n"
     ]
    }
   ],
   "source": [
    "# for node features to work well, they have to be normalized\n",
    "# you can choose \"standard\" for StandardScaler, \"robust\" for RobustScaler, \"quantile\"\n",
    "# for QuantileTransformer and \"power\" for PowerTransformer\n",
    "#\n",
    "# you could also pass sklearn compatible scalers by passing a dict keyed by\n",
    "# literal type, e.g.: {0: sklearn.preprocessing.StandardScaler()}\n",
    "train_embs = embedder.embeddings(\n",
    "    data.edge_index,\n",
    "    data.edge_type,\n",
    "    node_features=filtered_node_features,\n",
    "    node_features_scalers=\"robust\",\n",
    "    idx=train_idx,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1/1 [00:13<00:00, 13.09s/it]\n",
      "100%|██████████| 1/1 [00:13<00:00, 13.22s/it]\n"
     ]
    }
   ],
   "source": [
    "# only fit node feature scalers on nodes reachable from train nodes,\n",
    "# for val and test nodes, reuse the fit scalers using embedder.get_last_fit_scalers()\n",
    "val_embs = embedder.embeddings(\n",
    "    data.edge_index,\n",
    "    data.edge_type,\n",
    "    node_features=filtered_node_features,\n",
    "    node_features_scalers=embedder.get_last_fit_scalers(),\n",
    "    idx=val_idx,\n",
    ")\n",
    "test_embs = embedder.embeddings(\n",
    "    data.edge_index,\n",
    "    data.edge_type,\n",
    "    node_features=filtered_node_features,\n",
    "    node_features_scalers=embedder.get_last_fit_scalers(),\n",
    "    idx=data.test_idx,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Warning: less than 75% gpu memory available for training. Free: 6391.4375 Total: 11178.5\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Learning rate set to 0.080278\n",
      "0:\tlearn: 0.6394734\ttest: 0.6480066\tbest: 0.6480066 (0)\ttotal: 143ms\tremaining: 2m 22s\n",
      "1:\tlearn: 0.6043138\ttest: 0.6160162\tbest: 0.6160162 (1)\ttotal: 281ms\tremaining: 2m 20s\n",
      "2:\tlearn: 0.5708737\ttest: 0.5830180\tbest: 0.5830180 (2)\ttotal: 417ms\tremaining: 2m 18s\n",
      "3:\tlearn: 0.5516711\ttest: 0.5606688\tbest: 0.5606688 (3)\ttotal: 549ms\tremaining: 2m 16s\n",
      "4:\tlearn: 0.5237342\ttest: 0.5517336\tbest: 0.5517336 (4)\ttotal: 690ms\tremaining: 2m 17s\n",
      "5:\tlearn: 0.4953295\ttest: 0.5436822\tbest: 0.5436822 (5)\ttotal: 835ms\tremaining: 2m 18s\n",
      "6:\tlearn: 0.4811546\ttest: 0.5335904\tbest: 0.5335904 (6)\ttotal: 923ms\tremaining: 2m 10s\n",
      "7:\tlearn: 0.4684954\ttest: 0.5225592\tbest: 0.5225592 (7)\ttotal: 1.06s\tremaining: 2m 11s\n",
      "8:\tlearn: 0.4389852\ttest: 0.4974404\tbest: 0.4974404 (8)\ttotal: 1.21s\tremaining: 2m 12s\n",
      "9:\tlearn: 0.4252482\ttest: 0.5018680\tbest: 0.4974404 (8)\ttotal: 1.35s\tremaining: 2m 13s\n",
      "10:\tlearn: 0.4090147\ttest: 0.4854347\tbest: 0.4854347 (10)\ttotal: 1.48s\tremaining: 2m 13s\n",
      "11:\tlearn: 0.3973811\ttest: 0.4763975\tbest: 0.4763975 (11)\ttotal: 1.61s\tremaining: 2m 12s\n",
      "12:\tlearn: 0.3905145\ttest: 0.4769150\tbest: 0.4763975 (11)\ttotal: 1.75s\tremaining: 2m 13s\n",
      "13:\tlearn: 0.3768700\ttest: 0.4592112\tbest: 0.4592112 (13)\ttotal: 1.9s\tremaining: 2m 13s\n",
      "14:\tlearn: 0.3652158\ttest: 0.4563049\tbest: 0.4563049 (14)\ttotal: 2.04s\tremaining: 2m 14s\n",
      "15:\tlearn: 0.3613058\ttest: 0.4502937\tbest: 0.4502937 (15)\ttotal: 2.17s\tremaining: 2m 13s\n",
      "16:\tlearn: 0.3505124\ttest: 0.4429693\tbest: 0.4429693 (16)\ttotal: 2.31s\tremaining: 2m 13s\n",
      "17:\tlearn: 0.3422220\ttest: 0.4325638\tbest: 0.4325638 (17)\ttotal: 2.46s\tremaining: 2m 14s\n",
      "18:\tlearn: 0.3307356\ttest: 0.4306741\tbest: 0.4306741 (18)\ttotal: 2.6s\tremaining: 2m 14s\n",
      "19:\tlearn: 0.3191573\ttest: 0.4169085\tbest: 0.4169085 (19)\ttotal: 2.75s\tremaining: 2m 14s\n",
      "20:\tlearn: 0.3086800\ttest: 0.4135951\tbest: 0.4135951 (20)\ttotal: 2.88s\tremaining: 2m 14s\n",
      "21:\tlearn: 0.3065781\ttest: 0.4155235\tbest: 0.4135951 (20)\ttotal: 3.02s\tremaining: 2m 14s\n",
      "22:\tlearn: 0.2989077\ttest: 0.4136530\tbest: 0.4135951 (20)\ttotal: 3.15s\tremaining: 2m 14s\n",
      "23:\tlearn: 0.2916565\ttest: 0.4075632\tbest: 0.4075632 (23)\ttotal: 3.29s\tremaining: 2m 13s\n",
      "24:\tlearn: 0.2855821\ttest: 0.3986323\tbest: 0.3986323 (24)\ttotal: 3.43s\tremaining: 2m 13s\n",
      "25:\tlearn: 0.2793476\ttest: 0.3985705\tbest: 0.3985705 (25)\ttotal: 3.57s\tremaining: 2m 13s\n",
      "26:\tlearn: 0.2707357\ttest: 0.3873832\tbest: 0.3873832 (26)\ttotal: 3.71s\tremaining: 2m 13s\n",
      "27:\tlearn: 0.2607142\ttest: 0.3856811\tbest: 0.3856811 (27)\ttotal: 3.85s\tremaining: 2m 13s\n",
      "28:\tlearn: 0.2552234\ttest: 0.3859423\tbest: 0.3856811 (27)\ttotal: 3.98s\tremaining: 2m 13s\n",
      "29:\tlearn: 0.2451941\ttest: 0.3814318\tbest: 0.3814318 (29)\ttotal: 4.12s\tremaining: 2m 13s\n",
      "30:\tlearn: 0.2384325\ttest: 0.3803245\tbest: 0.3803245 (30)\ttotal: 4.26s\tremaining: 2m 13s\n",
      "31:\tlearn: 0.2341930\ttest: 0.3784893\tbest: 0.3784893 (31)\ttotal: 4.4s\tremaining: 2m 13s\n",
      "32:\tlearn: 0.2291283\ttest: 0.3784789\tbest: 0.3784789 (32)\ttotal: 4.54s\tremaining: 2m 13s\n",
      "33:\tlearn: 0.2249077\ttest: 0.3802365\tbest: 0.3784789 (32)\ttotal: 4.68s\tremaining: 2m 13s\n",
      "34:\tlearn: 0.2201818\ttest: 0.3738123\tbest: 0.3738123 (34)\ttotal: 4.81s\tremaining: 2m 12s\n",
      "35:\tlearn: 0.2139037\ttest: 0.3737398\tbest: 0.3737398 (35)\ttotal: 4.95s\tremaining: 2m 12s\n",
      "36:\tlearn: 0.2054661\ttest: 0.3636225\tbest: 0.3636225 (36)\ttotal: 5.09s\tremaining: 2m 12s\n",
      "37:\tlearn: 0.1992937\ttest: 0.3633832\tbest: 0.3633832 (37)\ttotal: 5.23s\tremaining: 2m 12s\n",
      "38:\tlearn: 0.1948788\ttest: 0.3626177\tbest: 0.3626177 (38)\ttotal: 5.37s\tremaining: 2m 12s\n",
      "39:\tlearn: 0.1907176\ttest: 0.3593376\tbest: 0.3593376 (39)\ttotal: 5.5s\tremaining: 2m 12s\n",
      "40:\tlearn: 0.1871773\ttest: 0.3583662\tbest: 0.3583662 (40)\ttotal: 5.64s\tremaining: 2m 12s\n",
      "41:\tlearn: 0.1834963\ttest: 0.3552711\tbest: 0.3552711 (41)\ttotal: 5.78s\tremaining: 2m 11s\n",
      "42:\tlearn: 0.1815070\ttest: 0.3538181\tbest: 0.3538181 (42)\ttotal: 5.92s\tremaining: 2m 11s\n",
      "43:\tlearn: 0.1765646\ttest: 0.3492678\tbest: 0.3492678 (43)\ttotal: 6.06s\tremaining: 2m 11s\n",
      "44:\tlearn: 0.1731536\ttest: 0.3490360\tbest: 0.3490360 (44)\ttotal: 6.2s\tremaining: 2m 11s\n",
      "45:\tlearn: 0.1697881\ttest: 0.3412557\tbest: 0.3412557 (45)\ttotal: 6.34s\tremaining: 2m 11s\n",
      "46:\tlearn: 0.1678169\ttest: 0.3397595\tbest: 0.3397595 (46)\ttotal: 6.47s\tremaining: 2m 11s\n",
      "47:\tlearn: 0.1643208\ttest: 0.3384613\tbest: 0.3384613 (47)\ttotal: 6.61s\tremaining: 2m 11s\n",
      "48:\tlearn: 0.1618450\ttest: 0.3354773\tbest: 0.3354773 (48)\ttotal: 6.75s\tremaining: 2m 10s\n",
      "49:\tlearn: 0.1574525\ttest: 0.3266881\tbest: 0.3266881 (49)\ttotal: 6.9s\tremaining: 2m 11s\n",
      "50:\tlearn: 0.1544889\ttest: 0.3253375\tbest: 0.3253375 (50)\ttotal: 7.04s\tremaining: 2m 11s\n",
      "51:\tlearn: 0.1532718\ttest: 0.3250737\tbest: 0.3250737 (51)\ttotal: 7.17s\tremaining: 2m 10s\n",
      "52:\tlearn: 0.1511556\ttest: 0.3212326\tbest: 0.3212326 (52)\ttotal: 7.3s\tremaining: 2m 10s\n",
      "53:\tlearn: 0.1509050\ttest: 0.3206753\tbest: 0.3206753 (53)\ttotal: 7.43s\tremaining: 2m 10s\n",
      "54:\tlearn: 0.1476562\ttest: 0.3217598\tbest: 0.3206753 (53)\ttotal: 7.57s\tremaining: 2m 10s\n",
      "55:\tlearn: 0.1449258\ttest: 0.3338838\tbest: 0.3206753 (53)\ttotal: 7.71s\tremaining: 2m 9s\n",
      "56:\tlearn: 0.1414641\ttest: 0.3317339\tbest: 0.3206753 (53)\ttotal: 7.85s\tremaining: 2m 9s\n",
      "57:\tlearn: 0.1388720\ttest: 0.3277775\tbest: 0.3206753 (53)\ttotal: 7.99s\tremaining: 2m 9s\n",
      "58:\tlearn: 0.1357670\ttest: 0.3333109\tbest: 0.3206753 (53)\ttotal: 8.13s\tremaining: 2m 9s\n",
      "59:\tlearn: 0.1327153\ttest: 0.3322467\tbest: 0.3206753 (53)\ttotal: 8.27s\tremaining: 2m 9s\n",
      "60:\tlearn: 0.1304341\ttest: 0.3314288\tbest: 0.3206753 (53)\ttotal: 8.41s\tremaining: 2m 9s\n",
      "61:\tlearn: 0.1260786\ttest: 0.3308448\tbest: 0.3206753 (53)\ttotal: 8.55s\tremaining: 2m 9s\n",
      "62:\tlearn: 0.1245370\ttest: 0.3319546\tbest: 0.3206753 (53)\ttotal: 8.69s\tremaining: 2m 9s\n",
      "63:\tlearn: 0.1230790\ttest: 0.3315655\tbest: 0.3206753 (53)\ttotal: 8.83s\tremaining: 2m 9s\n",
      "64:\tlearn: 0.1213712\ttest: 0.3351130\tbest: 0.3206753 (53)\ttotal: 8.96s\tremaining: 2m 8s\n",
      "65:\tlearn: 0.1196199\ttest: 0.3341032\tbest: 0.3206753 (53)\ttotal: 9.1s\tremaining: 2m 8s\n",
      "66:\tlearn: 0.1171452\ttest: 0.3258258\tbest: 0.3206753 (53)\ttotal: 9.24s\tremaining: 2m 8s\n",
      "67:\tlearn: 0.1126769\ttest: 0.3322236\tbest: 0.3206753 (53)\ttotal: 9.39s\tremaining: 2m 8s\n",
      "68:\tlearn: 0.1101640\ttest: 0.3354506\tbest: 0.3206753 (53)\ttotal: 9.54s\tremaining: 2m 8s\n",
      "69:\tlearn: 0.1072610\ttest: 0.3296324\tbest: 0.3206753 (53)\ttotal: 9.68s\tremaining: 2m 8s\n",
      "70:\tlearn: 0.1051247\ttest: 0.3258160\tbest: 0.3206753 (53)\ttotal: 9.83s\tremaining: 2m 8s\n",
      "71:\tlearn: 0.1036423\ttest: 0.3222940\tbest: 0.3206753 (53)\ttotal: 9.99s\tremaining: 2m 8s\n",
      "72:\tlearn: 0.1023105\ttest: 0.3258978\tbest: 0.3206753 (53)\ttotal: 10.1s\tremaining: 2m 8s\n",
      "73:\tlearn: 0.1012915\ttest: 0.3248717\tbest: 0.3206753 (53)\ttotal: 10.3s\tremaining: 2m 8s\n",
      "74:\tlearn: 0.0999672\ttest: 0.3239194\tbest: 0.3206753 (53)\ttotal: 10.4s\tremaining: 2m 8s\n",
      "75:\tlearn: 0.0982347\ttest: 0.3215438\tbest: 0.3206753 (53)\ttotal: 10.6s\tremaining: 2m 8s\n",
      "76:\tlearn: 0.0959250\ttest: 0.3202095\tbest: 0.3202095 (76)\ttotal: 10.7s\tremaining: 2m 8s\n",
      "77:\tlearn: 0.0944880\ttest: 0.3210167\tbest: 0.3202095 (76)\ttotal: 10.8s\tremaining: 2m 8s\n",
      "78:\tlearn: 0.0925091\ttest: 0.3204880\tbest: 0.3202095 (76)\ttotal: 11s\tremaining: 2m 8s\n",
      "79:\tlearn: 0.0908743\ttest: 0.3195239\tbest: 0.3195239 (79)\ttotal: 11.1s\tremaining: 2m 7s\n",
      "80:\tlearn: 0.0900491\ttest: 0.3182658\tbest: 0.3182658 (80)\ttotal: 11.3s\tremaining: 2m 7s\n",
      "81:\tlearn: 0.0881514\ttest: 0.3192497\tbest: 0.3182658 (80)\ttotal: 11.4s\tremaining: 2m 7s\n",
      "82:\tlearn: 0.0863894\ttest: 0.3182498\tbest: 0.3182498 (82)\ttotal: 11.5s\tremaining: 2m 7s\n",
      "83:\tlearn: 0.0851710\ttest: 0.3161407\tbest: 0.3161407 (83)\ttotal: 11.7s\tremaining: 2m 7s\n",
      "84:\tlearn: 0.0839694\ttest: 0.3138317\tbest: 0.3138317 (84)\ttotal: 11.8s\tremaining: 2m 7s\n",
      "85:\tlearn: 0.0823216\ttest: 0.3094795\tbest: 0.3094795 (85)\ttotal: 12s\tremaining: 2m 7s\n",
      "86:\tlearn: 0.0805074\ttest: 0.3065704\tbest: 0.3065704 (86)\ttotal: 12.1s\tremaining: 2m 7s\n",
      "87:\tlearn: 0.0794605\ttest: 0.3053204\tbest: 0.3053204 (87)\ttotal: 12.3s\tremaining: 2m 7s\n",
      "88:\tlearn: 0.0779377\ttest: 0.3052209\tbest: 0.3052209 (88)\ttotal: 12.4s\tremaining: 2m 7s\n",
      "89:\tlearn: 0.0774295\ttest: 0.3060644\tbest: 0.3052209 (88)\ttotal: 12.5s\tremaining: 2m 6s\n",
      "90:\tlearn: 0.0762737\ttest: 0.3000170\tbest: 0.3000170 (90)\ttotal: 12.7s\tremaining: 2m 6s\n",
      "91:\tlearn: 0.0746868\ttest: 0.2979074\tbest: 0.2979074 (91)\ttotal: 12.8s\tremaining: 2m 6s\n",
      "92:\tlearn: 0.0736565\ttest: 0.2968077\tbest: 0.2968077 (92)\ttotal: 13s\tremaining: 2m 6s\n",
      "93:\tlearn: 0.0724170\ttest: 0.2947924\tbest: 0.2947924 (93)\ttotal: 13.1s\tremaining: 2m 6s\n",
      "94:\tlearn: 0.0722137\ttest: 0.2952092\tbest: 0.2947924 (93)\ttotal: 13.3s\tremaining: 2m 6s\n",
      "95:\tlearn: 0.0714167\ttest: 0.2931717\tbest: 0.2931717 (95)\ttotal: 13.4s\tremaining: 2m 6s\n",
      "96:\tlearn: 0.0692388\ttest: 0.2939510\tbest: 0.2931717 (95)\ttotal: 13.5s\tremaining: 2m 6s\n",
      "97:\tlearn: 0.0675936\ttest: 0.2906260\tbest: 0.2906260 (97)\ttotal: 13.7s\tremaining: 2m 5s\n",
      "98:\tlearn: 0.0661831\ttest: 0.2861740\tbest: 0.2861740 (98)\ttotal: 13.8s\tremaining: 2m 5s\n",
      "99:\tlearn: 0.0654865\ttest: 0.2850338\tbest: 0.2850338 (99)\ttotal: 14s\tremaining: 2m 5s\n",
      "100:\tlearn: 0.0639433\ttest: 0.2888764\tbest: 0.2850338 (99)\ttotal: 14.1s\tremaining: 2m 5s\n",
      "101:\tlearn: 0.0630505\ttest: 0.2882900\tbest: 0.2850338 (99)\ttotal: 14.2s\tremaining: 2m 5s\n",
      "102:\tlearn: 0.0622483\ttest: 0.2874611\tbest: 0.2850338 (99)\ttotal: 14.4s\tremaining: 2m 5s\n",
      "103:\tlearn: 0.0612813\ttest: 0.2869061\tbest: 0.2850338 (99)\ttotal: 14.5s\tremaining: 2m 5s\n",
      "104:\tlearn: 0.0601793\ttest: 0.2867024\tbest: 0.2850338 (99)\ttotal: 14.7s\tremaining: 2m 5s\n",
      "105:\tlearn: 0.0591212\ttest: 0.2881703\tbest: 0.2850338 (99)\ttotal: 14.8s\tremaining: 2m 5s\n",
      "106:\tlearn: 0.0581428\ttest: 0.2872324\tbest: 0.2850338 (99)\ttotal: 15s\tremaining: 2m 4s\n",
      "107:\tlearn: 0.0569301\ttest: 0.2852462\tbest: 0.2850338 (99)\ttotal: 15.1s\tremaining: 2m 4s\n",
      "108:\tlearn: 0.0557871\ttest: 0.2885857\tbest: 0.2850338 (99)\ttotal: 15.3s\tremaining: 2m 4s\n",
      "109:\tlearn: 0.0547642\ttest: 0.2895187\tbest: 0.2850338 (99)\ttotal: 15.4s\tremaining: 2m 4s\n",
      "110:\tlearn: 0.0540875\ttest: 0.2899529\tbest: 0.2850338 (99)\ttotal: 15.6s\tremaining: 2m 4s\n",
      "111:\tlearn: 0.0538629\ttest: 0.2898326\tbest: 0.2850338 (99)\ttotal: 15.7s\tremaining: 2m 4s\n",
      "112:\tlearn: 0.0532058\ttest: 0.2903000\tbest: 0.2850338 (99)\ttotal: 15.8s\tremaining: 2m 4s\n",
      "113:\tlearn: 0.0522369\ttest: 0.2951868\tbest: 0.2850338 (99)\ttotal: 16s\tremaining: 2m 4s\n",
      "114:\tlearn: 0.0518011\ttest: 0.2970304\tbest: 0.2850338 (99)\ttotal: 16.1s\tremaining: 2m 3s\n",
      "115:\tlearn: 0.0508782\ttest: 0.2932861\tbest: 0.2850338 (99)\ttotal: 16.3s\tremaining: 2m 3s\n",
      "116:\tlearn: 0.0504505\ttest: 0.2925049\tbest: 0.2850338 (99)\ttotal: 16.4s\tremaining: 2m 3s\n",
      "117:\tlearn: 0.0497144\ttest: 0.2913651\tbest: 0.2850338 (99)\ttotal: 16.5s\tremaining: 2m 3s\n",
      "118:\tlearn: 0.0492032\ttest: 0.2923355\tbest: 0.2850338 (99)\ttotal: 16.7s\tremaining: 2m 3s\n",
      "119:\tlearn: 0.0486542\ttest: 0.2916584\tbest: 0.2850338 (99)\ttotal: 16.8s\tremaining: 2m 3s\n",
      "120:\tlearn: 0.0480729\ttest: 0.2914616\tbest: 0.2850338 (99)\ttotal: 17s\tremaining: 2m 3s\n",
      "121:\tlearn: 0.0471557\ttest: 0.2948131\tbest: 0.2850338 (99)\ttotal: 17.1s\tremaining: 2m 3s\n",
      "122:\tlearn: 0.0468941\ttest: 0.2942767\tbest: 0.2850338 (99)\ttotal: 17.3s\tremaining: 2m 3s\n",
      "123:\tlearn: 0.0460941\ttest: 0.2932198\tbest: 0.2850338 (99)\ttotal: 17.4s\tremaining: 2m 2s\n",
      "124:\tlearn: 0.0447811\ttest: 0.2948380\tbest: 0.2850338 (99)\ttotal: 17.5s\tremaining: 2m 2s\n",
      "125:\tlearn: 0.0442445\ttest: 0.2939233\tbest: 0.2850338 (99)\ttotal: 17.7s\tremaining: 2m 2s\n",
      "126:\tlearn: 0.0438456\ttest: 0.2951953\tbest: 0.2850338 (99)\ttotal: 17.8s\tremaining: 2m 2s\n",
      "127:\tlearn: 0.0435927\ttest: 0.2950152\tbest: 0.2850338 (99)\ttotal: 18s\tremaining: 2m 2s\n",
      "128:\tlearn: 0.0430205\ttest: 0.2931064\tbest: 0.2850338 (99)\ttotal: 18.1s\tremaining: 2m 2s\n",
      "129:\tlearn: 0.0424757\ttest: 0.2891753\tbest: 0.2850338 (99)\ttotal: 18.2s\tremaining: 2m 2s\n",
      "130:\tlearn: 0.0416623\ttest: 0.2882146\tbest: 0.2850338 (99)\ttotal: 18.4s\tremaining: 2m 2s\n",
      "131:\tlearn: 0.0410081\ttest: 0.2878628\tbest: 0.2850338 (99)\ttotal: 18.5s\tremaining: 2m 1s\n",
      "132:\tlearn: 0.0402487\ttest: 0.2873745\tbest: 0.2850338 (99)\ttotal: 18.7s\tremaining: 2m 1s\n",
      "133:\tlearn: 0.0399032\ttest: 0.2862932\tbest: 0.2850338 (99)\ttotal: 18.8s\tremaining: 2m 1s\n",
      "134:\tlearn: 0.0391681\ttest: 0.2857948\tbest: 0.2850338 (99)\ttotal: 19s\tremaining: 2m 1s\n",
      "135:\tlearn: 0.0388594\ttest: 0.2881171\tbest: 0.2850338 (99)\ttotal: 19.1s\tremaining: 2m 1s\n",
      "136:\tlearn: 0.0386129\ttest: 0.2886100\tbest: 0.2850338 (99)\ttotal: 19.2s\tremaining: 2m 1s\n",
      "137:\tlearn: 0.0384269\ttest: 0.2888902\tbest: 0.2850338 (99)\ttotal: 19.4s\tremaining: 2m 1s\n",
      "138:\tlearn: 0.0381998\ttest: 0.2885165\tbest: 0.2850338 (99)\ttotal: 19.5s\tremaining: 2m\n",
      "139:\tlearn: 0.0379844\ttest: 0.2883801\tbest: 0.2850338 (99)\ttotal: 19.7s\tremaining: 2m\n",
      "140:\tlearn: 0.0375099\ttest: 0.2893149\tbest: 0.2850338 (99)\ttotal: 19.8s\tremaining: 2m\n",
      "141:\tlearn: 0.0371117\ttest: 0.2901357\tbest: 0.2850338 (99)\ttotal: 19.9s\tremaining: 2m\n",
      "142:\tlearn: 0.0369790\ttest: 0.2879561\tbest: 0.2850338 (99)\ttotal: 20.1s\tremaining: 2m\n",
      "143:\tlearn: 0.0363731\ttest: 0.2858002\tbest: 0.2850338 (99)\ttotal: 20.2s\tremaining: 2m\n",
      "144:\tlearn: 0.0357672\ttest: 0.2864962\tbest: 0.2850338 (99)\ttotal: 20.4s\tremaining: 2m\n",
      "145:\tlearn: 0.0354699\ttest: 0.2863945\tbest: 0.2850338 (99)\ttotal: 20.5s\tremaining: 2m\n",
      "146:\tlearn: 0.0344779\ttest: 0.2808405\tbest: 0.2808405 (146)\ttotal: 20.7s\tremaining: 1m 59s\n",
      "147:\tlearn: 0.0337543\ttest: 0.2759626\tbest: 0.2759626 (147)\ttotal: 20.8s\tremaining: 1m 59s\n",
      "148:\tlearn: 0.0333159\ttest: 0.2769837\tbest: 0.2759626 (147)\ttotal: 20.9s\tremaining: 1m 59s\n",
      "149:\tlearn: 0.0330401\ttest: 0.2765098\tbest: 0.2759626 (147)\ttotal: 21.1s\tremaining: 1m 59s\n",
      "150:\tlearn: 0.0326211\ttest: 0.2799751\tbest: 0.2759626 (147)\ttotal: 21.2s\tremaining: 1m 59s\n",
      "151:\tlearn: 0.0321623\ttest: 0.2793529\tbest: 0.2759626 (147)\ttotal: 21.4s\tremaining: 1m 59s\n",
      "152:\tlearn: 0.0316942\ttest: 0.2756948\tbest: 0.2756948 (152)\ttotal: 21.5s\tremaining: 1m 59s\n",
      "153:\tlearn: 0.0312007\ttest: 0.2723317\tbest: 0.2723317 (153)\ttotal: 21.7s\tremaining: 1m 59s\n",
      "154:\tlearn: 0.0311201\ttest: 0.2727474\tbest: 0.2723317 (153)\ttotal: 21.8s\tremaining: 1m 58s\n",
      "155:\tlearn: 0.0309582\ttest: 0.2735951\tbest: 0.2723317 (153)\ttotal: 21.9s\tremaining: 1m 58s\n",
      "156:\tlearn: 0.0307566\ttest: 0.2746155\tbest: 0.2723317 (153)\ttotal: 22.1s\tremaining: 1m 58s\n",
      "157:\tlearn: 0.0305786\ttest: 0.2750122\tbest: 0.2723317 (153)\ttotal: 22.2s\tremaining: 1m 58s\n",
      "158:\tlearn: 0.0303490\ttest: 0.2727463\tbest: 0.2723317 (153)\ttotal: 22.4s\tremaining: 1m 58s\n",
      "159:\tlearn: 0.0302020\ttest: 0.2729467\tbest: 0.2723317 (153)\ttotal: 22.5s\tremaining: 1m 58s\n",
      "160:\tlearn: 0.0301256\ttest: 0.2727873\tbest: 0.2723317 (153)\ttotal: 22.6s\tremaining: 1m 58s\n",
      "161:\tlearn: 0.0297507\ttest: 0.2759361\tbest: 0.2723317 (153)\ttotal: 22.8s\tremaining: 1m 57s\n",
      "162:\tlearn: 0.0293072\ttest: 0.2750962\tbest: 0.2723317 (153)\ttotal: 22.9s\tremaining: 1m 57s\n",
      "163:\tlearn: 0.0288688\ttest: 0.2765252\tbest: 0.2723317 (153)\ttotal: 23.1s\tremaining: 1m 57s\n",
      "164:\tlearn: 0.0286660\ttest: 0.2751428\tbest: 0.2723317 (153)\ttotal: 23.2s\tremaining: 1m 57s\n",
      "165:\tlearn: 0.0285545\ttest: 0.2745524\tbest: 0.2723317 (153)\ttotal: 23.4s\tremaining: 1m 57s\n",
      "166:\tlearn: 0.0281744\ttest: 0.2757988\tbest: 0.2723317 (153)\ttotal: 23.5s\tremaining: 1m 57s\n",
      "167:\tlearn: 0.0277873\ttest: 0.2731711\tbest: 0.2723317 (153)\ttotal: 23.6s\tremaining: 1m 57s\n",
      "168:\tlearn: 0.0275073\ttest: 0.2739425\tbest: 0.2723317 (153)\ttotal: 23.8s\tremaining: 1m 56s\n",
      "169:\tlearn: 0.0271659\ttest: 0.2757181\tbest: 0.2723317 (153)\ttotal: 23.9s\tremaining: 1m 56s\n",
      "170:\tlearn: 0.0266242\ttest: 0.2760726\tbest: 0.2723317 (153)\ttotal: 24.1s\tremaining: 1m 56s\n",
      "171:\tlearn: 0.0263716\ttest: 0.2773178\tbest: 0.2723317 (153)\ttotal: 24.2s\tremaining: 1m 56s\n",
      "172:\tlearn: 0.0259225\ttest: 0.2761062\tbest: 0.2723317 (153)\ttotal: 24.4s\tremaining: 1m 56s\n",
      "173:\tlearn: 0.0257382\ttest: 0.2760375\tbest: 0.2723317 (153)\ttotal: 24.5s\tremaining: 1m 56s\n",
      "174:\tlearn: 0.0256745\ttest: 0.2752305\tbest: 0.2723317 (153)\ttotal: 24.6s\tremaining: 1m 56s\n",
      "175:\tlearn: 0.0255465\ttest: 0.2745530\tbest: 0.2723317 (153)\ttotal: 24.8s\tremaining: 1m 56s\n",
      "176:\tlearn: 0.0254174\ttest: 0.2723296\tbest: 0.2723296 (176)\ttotal: 24.9s\tremaining: 1m 55s\n",
      "177:\tlearn: 0.0252618\ttest: 0.2719670\tbest: 0.2719670 (177)\ttotal: 25.1s\tremaining: 1m 55s\n",
      "178:\tlearn: 0.0251403\ttest: 0.2724206\tbest: 0.2719670 (177)\ttotal: 25.2s\tremaining: 1m 55s\n",
      "179:\tlearn: 0.0246870\ttest: 0.2706709\tbest: 0.2706709 (179)\ttotal: 25.4s\tremaining: 1m 55s\n",
      "180:\tlearn: 0.0246103\ttest: 0.2705156\tbest: 0.2705156 (180)\ttotal: 25.5s\tremaining: 1m 55s\n",
      "181:\tlearn: 0.0245409\ttest: 0.2707372\tbest: 0.2705156 (180)\ttotal: 25.6s\tremaining: 1m 55s\n",
      "182:\tlearn: 0.0241114\ttest: 0.2712215\tbest: 0.2705156 (180)\ttotal: 25.8s\tremaining: 1m 55s\n",
      "183:\tlearn: 0.0240144\ttest: 0.2716659\tbest: 0.2705156 (180)\ttotal: 25.9s\tremaining: 1m 54s\n",
      "184:\tlearn: 0.0237717\ttest: 0.2722746\tbest: 0.2705156 (180)\ttotal: 26.1s\tremaining: 1m 54s\n",
      "185:\tlearn: 0.0235573\ttest: 0.2719672\tbest: 0.2705156 (180)\ttotal: 26.2s\tremaining: 1m 54s\n",
      "186:\tlearn: 0.0234469\ttest: 0.2704860\tbest: 0.2704860 (186)\ttotal: 26.4s\tremaining: 1m 54s\n",
      "187:\tlearn: 0.0232978\ttest: 0.2709987\tbest: 0.2704860 (186)\ttotal: 26.5s\tremaining: 1m 54s\n",
      "188:\tlearn: 0.0230661\ttest: 0.2709798\tbest: 0.2704860 (186)\ttotal: 26.6s\tremaining: 1m 54s\n",
      "189:\tlearn: 0.0229619\ttest: 0.2724014\tbest: 0.2704860 (186)\ttotal: 26.8s\tremaining: 1m 54s\n",
      "190:\tlearn: 0.0227059\ttest: 0.2740151\tbest: 0.2704860 (186)\ttotal: 26.9s\tremaining: 1m 54s\n",
      "191:\tlearn: 0.0225802\ttest: 0.2741058\tbest: 0.2704860 (186)\ttotal: 27.1s\tremaining: 1m 53s\n",
      "192:\tlearn: 0.0224862\ttest: 0.2744837\tbest: 0.2704860 (186)\ttotal: 27.2s\tremaining: 1m 53s\n",
      "193:\tlearn: 0.0224410\ttest: 0.2748062\tbest: 0.2704860 (186)\ttotal: 27.3s\tremaining: 1m 53s\n",
      "194:\tlearn: 0.0223527\ttest: 0.2760362\tbest: 0.2704860 (186)\ttotal: 27.5s\tremaining: 1m 53s\n",
      "195:\tlearn: 0.0222674\ttest: 0.2759379\tbest: 0.2704860 (186)\ttotal: 27.6s\tremaining: 1m 53s\n",
      "196:\tlearn: 0.0222052\ttest: 0.2761238\tbest: 0.2704860 (186)\ttotal: 27.8s\tremaining: 1m 53s\n",
      "197:\tlearn: 0.0218980\ttest: 0.2769714\tbest: 0.2704860 (186)\ttotal: 27.9s\tremaining: 1m 53s\n",
      "198:\tlearn: 0.0217185\ttest: 0.2776643\tbest: 0.2704860 (186)\ttotal: 28.1s\tremaining: 1m 52s\n",
      "199:\tlearn: 0.0216402\ttest: 0.2774234\tbest: 0.2704860 (186)\ttotal: 28.2s\tremaining: 1m 52s\n",
      "200:\tlearn: 0.0214964\ttest: 0.2779772\tbest: 0.2704860 (186)\ttotal: 28.3s\tremaining: 1m 52s\n",
      "201:\tlearn: 0.0214672\ttest: 0.2772250\tbest: 0.2704860 (186)\ttotal: 28.5s\tremaining: 1m 52s\n",
      "202:\tlearn: 0.0213738\ttest: 0.2772630\tbest: 0.2704860 (186)\ttotal: 28.6s\tremaining: 1m 52s\n",
      "203:\tlearn: 0.0212471\ttest: 0.2777414\tbest: 0.2704860 (186)\ttotal: 28.8s\tremaining: 1m 52s\n",
      "204:\tlearn: 0.0210903\ttest: 0.2760704\tbest: 0.2704860 (186)\ttotal: 28.9s\tremaining: 1m 52s\n",
      "205:\tlearn: 0.0210607\ttest: 0.2757442\tbest: 0.2704860 (186)\ttotal: 29s\tremaining: 1m 51s\n",
      "206:\tlearn: 0.0209799\ttest: 0.2751494\tbest: 0.2704860 (186)\ttotal: 29.2s\tremaining: 1m 51s\n",
      "207:\tlearn: 0.0209515\ttest: 0.2746509\tbest: 0.2704860 (186)\ttotal: 29.3s\tremaining: 1m 51s\n",
      "208:\tlearn: 0.0207918\ttest: 0.2750441\tbest: 0.2704860 (186)\ttotal: 29.5s\tremaining: 1m 51s\n",
      "209:\tlearn: 0.0207050\ttest: 0.2743553\tbest: 0.2704860 (186)\ttotal: 29.6s\tremaining: 1m 51s\n",
      "210:\tlearn: 0.0205038\ttest: 0.2732220\tbest: 0.2704860 (186)\ttotal: 29.7s\tremaining: 1m 51s\n",
      "211:\tlearn: 0.0204309\ttest: 0.2726186\tbest: 0.2704860 (186)\ttotal: 29.9s\tremaining: 1m 51s\n",
      "212:\tlearn: 0.0200462\ttest: 0.2726959\tbest: 0.2704860 (186)\ttotal: 30s\tremaining: 1m 50s\n",
      "213:\tlearn: 0.0197874\ttest: 0.2714662\tbest: 0.2704860 (186)\ttotal: 30.2s\tremaining: 1m 50s\n",
      "214:\tlearn: 0.0197705\ttest: 0.2711338\tbest: 0.2704860 (186)\ttotal: 30.3s\tremaining: 1m 50s\n",
      "215:\tlearn: 0.0193884\ttest: 0.2703297\tbest: 0.2703297 (215)\ttotal: 30.5s\tremaining: 1m 50s\n",
      "216:\tlearn: 0.0191110\ttest: 0.2698702\tbest: 0.2698702 (216)\ttotal: 30.6s\tremaining: 1m 50s\n",
      "217:\tlearn: 0.0190258\ttest: 0.2697780\tbest: 0.2697780 (217)\ttotal: 30.7s\tremaining: 1m 50s\n",
      "218:\tlearn: 0.0189444\ttest: 0.2687319\tbest: 0.2687319 (218)\ttotal: 30.9s\tremaining: 1m 50s\n",
      "219:\tlearn: 0.0188911\ttest: 0.2690709\tbest: 0.2687319 (218)\ttotal: 31s\tremaining: 1m 50s\n",
      "220:\tlearn: 0.0185719\ttest: 0.2674192\tbest: 0.2674192 (220)\ttotal: 31.2s\tremaining: 1m 49s\n",
      "221:\tlearn: 0.0183840\ttest: 0.2668677\tbest: 0.2668677 (221)\ttotal: 31.3s\tremaining: 1m 49s\n",
      "222:\tlearn: 0.0182017\ttest: 0.2639457\tbest: 0.2639457 (222)\ttotal: 31.5s\tremaining: 1m 49s\n",
      "223:\tlearn: 0.0180557\ttest: 0.2644021\tbest: 0.2639457 (222)\ttotal: 31.6s\tremaining: 1m 49s\n",
      "224:\tlearn: 0.0179687\ttest: 0.2653796\tbest: 0.2639457 (222)\ttotal: 31.7s\tremaining: 1m 49s\n",
      "225:\tlearn: 0.0179240\ttest: 0.2650868\tbest: 0.2639457 (222)\ttotal: 31.9s\tremaining: 1m 49s\n",
      "226:\tlearn: 0.0179018\ttest: 0.2650166\tbest: 0.2639457 (222)\ttotal: 32s\tremaining: 1m 49s\n",
      "227:\tlearn: 0.0178229\ttest: 0.2653917\tbest: 0.2639457 (222)\ttotal: 32.2s\tremaining: 1m 48s\n",
      "228:\tlearn: 0.0176096\ttest: 0.2659295\tbest: 0.2639457 (222)\ttotal: 32.3s\tremaining: 1m 48s\n",
      "229:\tlearn: 0.0175477\ttest: 0.2652680\tbest: 0.2639457 (222)\ttotal: 32.5s\tremaining: 1m 48s\n",
      "230:\tlearn: 0.0173482\ttest: 0.2656377\tbest: 0.2639457 (222)\ttotal: 32.6s\tremaining: 1m 48s\n",
      "231:\tlearn: 0.0172052\ttest: 0.2672841\tbest: 0.2639457 (222)\ttotal: 32.8s\tremaining: 1m 48s\n",
      "232:\tlearn: 0.0171378\ttest: 0.2671430\tbest: 0.2639457 (222)\ttotal: 32.9s\tremaining: 1m 48s\n",
      "233:\tlearn: 0.0170383\ttest: 0.2666133\tbest: 0.2639457 (222)\ttotal: 33s\tremaining: 1m 48s\n",
      "234:\tlearn: 0.0170005\ttest: 0.2665384\tbest: 0.2639457 (222)\ttotal: 33.2s\tremaining: 1m 48s\n",
      "235:\tlearn: 0.0168800\ttest: 0.2666475\tbest: 0.2639457 (222)\ttotal: 33.3s\tremaining: 1m 47s\n",
      "236:\tlearn: 0.0168590\ttest: 0.2664937\tbest: 0.2639457 (222)\ttotal: 33.5s\tremaining: 1m 47s\n",
      "237:\tlearn: 0.0168043\ttest: 0.2660172\tbest: 0.2639457 (222)\ttotal: 33.6s\tremaining: 1m 47s\n",
      "238:\tlearn: 0.0167663\ttest: 0.2654602\tbest: 0.2639457 (222)\ttotal: 33.7s\tremaining: 1m 47s\n",
      "239:\tlearn: 0.0167370\ttest: 0.2656325\tbest: 0.2639457 (222)\ttotal: 33.9s\tremaining: 1m 47s\n",
      "240:\tlearn: 0.0166227\ttest: 0.2652555\tbest: 0.2639457 (222)\ttotal: 34s\tremaining: 1m 47s\n",
      "241:\tlearn: 0.0165747\ttest: 0.2653073\tbest: 0.2639457 (222)\ttotal: 34.2s\tremaining: 1m 47s\n",
      "242:\tlearn: 0.0164159\ttest: 0.2649318\tbest: 0.2639457 (222)\ttotal: 34.3s\tremaining: 1m 46s\n",
      "243:\tlearn: 0.0163244\ttest: 0.2652159\tbest: 0.2639457 (222)\ttotal: 34.5s\tremaining: 1m 46s\n",
      "244:\tlearn: 0.0162213\ttest: 0.2664560\tbest: 0.2639457 (222)\ttotal: 34.6s\tremaining: 1m 46s\n",
      "245:\tlearn: 0.0161364\ttest: 0.2652724\tbest: 0.2639457 (222)\ttotal: 34.8s\tremaining: 1m 46s\n",
      "246:\tlearn: 0.0160317\ttest: 0.2650518\tbest: 0.2639457 (222)\ttotal: 34.9s\tremaining: 1m 46s\n",
      "247:\tlearn: 0.0159637\ttest: 0.2649028\tbest: 0.2639457 (222)\ttotal: 35s\tremaining: 1m 46s\n",
      "248:\tlearn: 0.0158972\ttest: 0.2647785\tbest: 0.2639457 (222)\ttotal: 35.2s\tremaining: 1m 46s\n",
      "249:\tlearn: 0.0158293\ttest: 0.2639996\tbest: 0.2639457 (222)\ttotal: 35.3s\tremaining: 1m 46s\n",
      "250:\tlearn: 0.0157903\ttest: 0.2633502\tbest: 0.2633502 (250)\ttotal: 35.5s\tremaining: 1m 45s\n",
      "251:\tlearn: 0.0156566\ttest: 0.2632003\tbest: 0.2632003 (251)\ttotal: 35.6s\tremaining: 1m 45s\n",
      "252:\tlearn: 0.0155580\ttest: 0.2615125\tbest: 0.2615125 (252)\ttotal: 35.9s\tremaining: 1m 45s\n",
      "253:\tlearn: 0.0155292\ttest: 0.2609414\tbest: 0.2609414 (253)\ttotal: 36s\tremaining: 1m 45s\n",
      "254:\tlearn: 0.0153252\ttest: 0.2615544\tbest: 0.2609414 (253)\ttotal: 36.1s\tremaining: 1m 45s\n",
      "255:\tlearn: 0.0152254\ttest: 0.2607900\tbest: 0.2607900 (255)\ttotal: 36.3s\tremaining: 1m 45s\n",
      "256:\tlearn: 0.0152114\ttest: 0.2600148\tbest: 0.2600148 (256)\ttotal: 36.4s\tremaining: 1m 45s\n",
      "257:\tlearn: 0.0151662\ttest: 0.2607993\tbest: 0.2600148 (256)\ttotal: 36.6s\tremaining: 1m 45s\n",
      "258:\tlearn: 0.0150374\ttest: 0.2623879\tbest: 0.2600148 (256)\ttotal: 36.7s\tremaining: 1m 45s\n",
      "259:\tlearn: 0.0149696\ttest: 0.2615912\tbest: 0.2600148 (256)\ttotal: 36.9s\tremaining: 1m 44s\n",
      "260:\tlearn: 0.0148259\ttest: 0.2609797\tbest: 0.2600148 (256)\ttotal: 37s\tremaining: 1m 44s\n",
      "261:\tlearn: 0.0146882\ttest: 0.2595877\tbest: 0.2595877 (261)\ttotal: 37.2s\tremaining: 1m 44s\n",
      "262:\tlearn: 0.0146707\ttest: 0.2594465\tbest: 0.2594465 (262)\ttotal: 37.4s\tremaining: 1m 44s\n",
      "263:\tlearn: 0.0144895\ttest: 0.2571859\tbest: 0.2571859 (263)\ttotal: 37.5s\tremaining: 1m 44s\n",
      "264:\tlearn: 0.0144115\ttest: 0.2559697\tbest: 0.2559697 (264)\ttotal: 37.7s\tremaining: 1m 44s\n",
      "265:\tlearn: 0.0142509\ttest: 0.2567723\tbest: 0.2559697 (264)\ttotal: 37.9s\tremaining: 1m 44s\n",
      "266:\tlearn: 0.0142304\ttest: 0.2562601\tbest: 0.2559697 (264)\ttotal: 38s\tremaining: 1m 44s\n",
      "267:\tlearn: 0.0141820\ttest: 0.2562055\tbest: 0.2559697 (264)\ttotal: 38.2s\tremaining: 1m 44s\n",
      "268:\tlearn: 0.0141513\ttest: 0.2562476\tbest: 0.2559697 (264)\ttotal: 38.3s\tremaining: 1m 44s\n",
      "269:\tlearn: 0.0141069\ttest: 0.2545152\tbest: 0.2545152 (269)\ttotal: 38.5s\tremaining: 1m 43s\n",
      "270:\tlearn: 0.0140630\ttest: 0.2546901\tbest: 0.2545152 (269)\ttotal: 38.6s\tremaining: 1m 43s\n",
      "271:\tlearn: 0.0139999\ttest: 0.2547826\tbest: 0.2545152 (269)\ttotal: 38.8s\tremaining: 1m 43s\n",
      "272:\tlearn: 0.0139640\ttest: 0.2554810\tbest: 0.2545152 (269)\ttotal: 38.9s\tremaining: 1m 43s\n",
      "273:\tlearn: 0.0139480\ttest: 0.2556390\tbest: 0.2545152 (269)\ttotal: 39s\tremaining: 1m 43s\n",
      "274:\tlearn: 0.0138814\ttest: 0.2554406\tbest: 0.2545152 (269)\ttotal: 39.2s\tremaining: 1m 43s\n",
      "275:\tlearn: 0.0137544\ttest: 0.2558657\tbest: 0.2545152 (269)\ttotal: 39.3s\tremaining: 1m 43s\n",
      "276:\tlearn: 0.0137131\ttest: 0.2556103\tbest: 0.2545152 (269)\ttotal: 39.5s\tremaining: 1m 43s\n",
      "277:\tlearn: 0.0136938\ttest: 0.2552960\tbest: 0.2545152 (269)\ttotal: 39.6s\tremaining: 1m 42s\n",
      "278:\tlearn: 0.0136299\ttest: 0.2542246\tbest: 0.2542246 (278)\ttotal: 39.8s\tremaining: 1m 42s\n",
      "279:\tlearn: 0.0135608\ttest: 0.2539622\tbest: 0.2539622 (279)\ttotal: 39.9s\tremaining: 1m 42s\n",
      "280:\tlearn: 0.0135119\ttest: 0.2536103\tbest: 0.2536103 (280)\ttotal: 40.1s\tremaining: 1m 42s\n",
      "281:\tlearn: 0.0133911\ttest: 0.2534819\tbest: 0.2534819 (281)\ttotal: 40.2s\tremaining: 1m 42s\n",
      "282:\tlearn: 0.0133606\ttest: 0.2535869\tbest: 0.2534819 (281)\ttotal: 40.3s\tremaining: 1m 42s\n",
      "283:\tlearn: 0.0133445\ttest: 0.2538760\tbest: 0.2534819 (281)\ttotal: 40.5s\tremaining: 1m 42s\n",
      "284:\tlearn: 0.0132253\ttest: 0.2533146\tbest: 0.2533146 (284)\ttotal: 40.6s\tremaining: 1m 41s\n",
      "285:\tlearn: 0.0132079\ttest: 0.2531767\tbest: 0.2531767 (285)\ttotal: 40.8s\tremaining: 1m 41s\n",
      "286:\tlearn: 0.0131427\ttest: 0.2539343\tbest: 0.2531767 (285)\ttotal: 40.9s\tremaining: 1m 41s\n",
      "287:\tlearn: 0.0131118\ttest: 0.2540228\tbest: 0.2531767 (285)\ttotal: 41.1s\tremaining: 1m 41s\n",
      "288:\tlearn: 0.0130754\ttest: 0.2539035\tbest: 0.2531767 (285)\ttotal: 41.2s\tremaining: 1m 41s\n",
      "289:\tlearn: 0.0130196\ttest: 0.2545514\tbest: 0.2531767 (285)\ttotal: 41.3s\tremaining: 1m 41s\n",
      "290:\tlearn: 0.0128557\ttest: 0.2552015\tbest: 0.2531767 (285)\ttotal: 41.5s\tremaining: 1m 41s\n",
      "291:\tlearn: 0.0128471\ttest: 0.2554828\tbest: 0.2531767 (285)\ttotal: 41.6s\tremaining: 1m 40s\n",
      "292:\tlearn: 0.0128068\ttest: 0.2563395\tbest: 0.2531767 (285)\ttotal: 41.8s\tremaining: 1m 40s\n",
      "293:\tlearn: 0.0127239\ttest: 0.2555302\tbest: 0.2531767 (285)\ttotal: 41.9s\tremaining: 1m 40s\n",
      "294:\tlearn: 0.0125470\ttest: 0.2539733\tbest: 0.2531767 (285)\ttotal: 42s\tremaining: 1m 40s\n",
      "295:\tlearn: 0.0125371\ttest: 0.2535943\tbest: 0.2531767 (285)\ttotal: 42.2s\tremaining: 1m 40s\n",
      "296:\tlearn: 0.0124121\ttest: 0.2568347\tbest: 0.2531767 (285)\ttotal: 42.3s\tremaining: 1m 40s\n",
      "297:\tlearn: 0.0123820\ttest: 0.2567483\tbest: 0.2531767 (285)\ttotal: 42.5s\tremaining: 1m 40s\n",
      "298:\tlearn: 0.0123389\ttest: 0.2561871\tbest: 0.2531767 (285)\ttotal: 42.6s\tremaining: 1m 39s\n",
      "299:\tlearn: 0.0123219\ttest: 0.2563703\tbest: 0.2531767 (285)\ttotal: 42.7s\tremaining: 1m 39s\n",
      "300:\tlearn: 0.0122998\ttest: 0.2556855\tbest: 0.2531767 (285)\ttotal: 42.9s\tremaining: 1m 39s\n",
      "301:\tlearn: 0.0122060\ttest: 0.2560710\tbest: 0.2531767 (285)\ttotal: 43s\tremaining: 1m 39s\n",
      "302:\tlearn: 0.0121518\ttest: 0.2552128\tbest: 0.2531767 (285)\ttotal: 43.2s\tremaining: 1m 39s\n",
      "303:\tlearn: 0.0121301\ttest: 0.2553612\tbest: 0.2531767 (285)\ttotal: 43.3s\tremaining: 1m 39s\n",
      "304:\tlearn: 0.0121024\ttest: 0.2552814\tbest: 0.2531767 (285)\ttotal: 43.4s\tremaining: 1m 38s\n",
      "305:\tlearn: 0.0120044\ttest: 0.2552018\tbest: 0.2531767 (285)\ttotal: 43.6s\tremaining: 1m 38s\n",
      "306:\tlearn: 0.0119620\ttest: 0.2547649\tbest: 0.2531767 (285)\ttotal: 43.7s\tremaining: 1m 38s\n",
      "307:\tlearn: 0.0118565\ttest: 0.2563643\tbest: 0.2531767 (285)\ttotal: 43.9s\tremaining: 1m 38s\n",
      "308:\tlearn: 0.0118282\ttest: 0.2570013\tbest: 0.2531767 (285)\ttotal: 44s\tremaining: 1m 38s\n",
      "309:\tlearn: 0.0117807\ttest: 0.2555201\tbest: 0.2531767 (285)\ttotal: 44.2s\tremaining: 1m 38s\n",
      "310:\tlearn: 0.0117667\ttest: 0.2549907\tbest: 0.2531767 (285)\ttotal: 44.3s\tremaining: 1m 38s\n",
      "311:\tlearn: 0.0116529\ttest: 0.2533799\tbest: 0.2531767 (285)\ttotal: 44.4s\tremaining: 1m 38s\n",
      "312:\tlearn: 0.0116159\ttest: 0.2534296\tbest: 0.2531767 (285)\ttotal: 44.6s\tremaining: 1m 37s\n",
      "313:\tlearn: 0.0116073\ttest: 0.2535719\tbest: 0.2531767 (285)\ttotal: 44.7s\tremaining: 1m 37s\n",
      "314:\tlearn: 0.0115954\ttest: 0.2536929\tbest: 0.2531767 (285)\ttotal: 44.9s\tremaining: 1m 37s\n",
      "315:\tlearn: 0.0114899\ttest: 0.2536323\tbest: 0.2531767 (285)\ttotal: 45s\tremaining: 1m 37s\n",
      "316:\tlearn: 0.0113088\ttest: 0.2523333\tbest: 0.2523333 (316)\ttotal: 45.1s\tremaining: 1m 37s\n",
      "317:\tlearn: 0.0112698\ttest: 0.2509166\tbest: 0.2509166 (317)\ttotal: 45.3s\tremaining: 1m 37s\n",
      "318:\tlearn: 0.0112284\ttest: 0.2513417\tbest: 0.2509166 (317)\ttotal: 45.4s\tremaining: 1m 37s\n",
      "319:\tlearn: 0.0111533\ttest: 0.2527073\tbest: 0.2509166 (317)\ttotal: 45.6s\tremaining: 1m 36s\n",
      "320:\tlearn: 0.0110977\ttest: 0.2519800\tbest: 0.2509166 (317)\ttotal: 45.8s\tremaining: 1m 36s\n",
      "321:\tlearn: 0.0110855\ttest: 0.2525028\tbest: 0.2509166 (317)\ttotal: 45.9s\tremaining: 1m 36s\n",
      "322:\tlearn: 0.0110663\ttest: 0.2517260\tbest: 0.2509166 (317)\ttotal: 46s\tremaining: 1m 36s\n",
      "323:\tlearn: 0.0109968\ttest: 0.2523825\tbest: 0.2509166 (317)\ttotal: 46.2s\tremaining: 1m 36s\n",
      "324:\tlearn: 0.0109281\ttest: 0.2505121\tbest: 0.2505121 (324)\ttotal: 46.3s\tremaining: 1m 36s\n",
      "325:\tlearn: 0.0109015\ttest: 0.2513806\tbest: 0.2505121 (324)\ttotal: 46.5s\tremaining: 1m 36s\n",
      "326:\tlearn: 0.0108569\ttest: 0.2512665\tbest: 0.2505121 (324)\ttotal: 46.6s\tremaining: 1m 35s\n",
      "327:\tlearn: 0.0108410\ttest: 0.2512305\tbest: 0.2505121 (324)\ttotal: 46.8s\tremaining: 1m 35s\n",
      "328:\tlearn: 0.0107817\ttest: 0.2511900\tbest: 0.2505121 (324)\ttotal: 46.9s\tremaining: 1m 35s\n",
      "329:\tlearn: 0.0107646\ttest: 0.2511318\tbest: 0.2505121 (324)\ttotal: 47s\tremaining: 1m 35s\n",
      "330:\tlearn: 0.0107433\ttest: 0.2516040\tbest: 0.2505121 (324)\ttotal: 47.2s\tremaining: 1m 35s\n",
      "331:\tlearn: 0.0107103\ttest: 0.2519463\tbest: 0.2505121 (324)\ttotal: 47.3s\tremaining: 1m 35s\n",
      "332:\tlearn: 0.0106556\ttest: 0.2519442\tbest: 0.2505121 (324)\ttotal: 47.5s\tremaining: 1m 35s\n",
      "333:\tlearn: 0.0106158\ttest: 0.2502394\tbest: 0.2502394 (333)\ttotal: 47.6s\tremaining: 1m 34s\n",
      "334:\tlearn: 0.0104899\ttest: 0.2515431\tbest: 0.2502394 (333)\ttotal: 47.7s\tremaining: 1m 34s\n",
      "335:\tlearn: 0.0104501\ttest: 0.2517295\tbest: 0.2502394 (333)\ttotal: 47.9s\tremaining: 1m 34s\n",
      "336:\tlearn: 0.0104138\ttest: 0.2510858\tbest: 0.2502394 (333)\ttotal: 48s\tremaining: 1m 34s\n",
      "337:\tlearn: 0.0102705\ttest: 0.2494073\tbest: 0.2494073 (337)\ttotal: 48.2s\tremaining: 1m 34s\n",
      "338:\tlearn: 0.0102217\ttest: 0.2481397\tbest: 0.2481397 (338)\ttotal: 48.3s\tremaining: 1m 34s\n",
      "339:\tlearn: 0.0101924\ttest: 0.2472293\tbest: 0.2472293 (339)\ttotal: 48.5s\tremaining: 1m 34s\n",
      "340:\tlearn: 0.0101456\ttest: 0.2470874\tbest: 0.2470874 (340)\ttotal: 48.6s\tremaining: 1m 33s\n",
      "341:\tlearn: 0.0101315\ttest: 0.2467096\tbest: 0.2467096 (341)\ttotal: 48.7s\tremaining: 1m 33s\n",
      "342:\tlearn: 0.0100709\ttest: 0.2466306\tbest: 0.2466306 (342)\ttotal: 48.9s\tremaining: 1m 33s\n",
      "343:\tlearn: 0.0100643\ttest: 0.2467306\tbest: 0.2466306 (342)\ttotal: 49s\tremaining: 1m 33s\n",
      "344:\tlearn: 0.0099901\ttest: 0.2470926\tbest: 0.2466306 (342)\ttotal: 49.2s\tremaining: 1m 33s\n",
      "345:\tlearn: 0.0099786\ttest: 0.2468768\tbest: 0.2466306 (342)\ttotal: 49.3s\tremaining: 1m 33s\n",
      "346:\tlearn: 0.0099726\ttest: 0.2467836\tbest: 0.2466306 (342)\ttotal: 49.5s\tremaining: 1m 33s\n",
      "347:\tlearn: 0.0099213\ttest: 0.2463868\tbest: 0.2463868 (347)\ttotal: 49.6s\tremaining: 1m 32s\n",
      "348:\tlearn: 0.0098908\ttest: 0.2466551\tbest: 0.2463868 (347)\ttotal: 49.8s\tremaining: 1m 32s\n",
      "349:\tlearn: 0.0098173\ttest: 0.2469137\tbest: 0.2463868 (347)\ttotal: 49.9s\tremaining: 1m 32s\n",
      "350:\tlearn: 0.0097515\ttest: 0.2473559\tbest: 0.2463868 (347)\ttotal: 50s\tremaining: 1m 32s\n",
      "351:\tlearn: 0.0097038\ttest: 0.2474691\tbest: 0.2463868 (347)\ttotal: 50.2s\tremaining: 1m 32s\n",
      "352:\tlearn: 0.0096803\ttest: 0.2475018\tbest: 0.2463868 (347)\ttotal: 50.3s\tremaining: 1m 32s\n",
      "353:\tlearn: 0.0096637\ttest: 0.2475032\tbest: 0.2463868 (347)\ttotal: 50.5s\tremaining: 1m 32s\n",
      "354:\tlearn: 0.0096520\ttest: 0.2475587\tbest: 0.2463868 (347)\ttotal: 50.6s\tremaining: 1m 32s\n",
      "355:\tlearn: 0.0096368\ttest: 0.2474469\tbest: 0.2463868 (347)\ttotal: 50.8s\tremaining: 1m 31s\n",
      "356:\tlearn: 0.0095904\ttest: 0.2476394\tbest: 0.2463868 (347)\ttotal: 50.9s\tremaining: 1m 31s\n",
      "357:\tlearn: 0.0095830\ttest: 0.2480728\tbest: 0.2463868 (347)\ttotal: 51.1s\tremaining: 1m 31s\n",
      "358:\tlearn: 0.0095693\ttest: 0.2482120\tbest: 0.2463868 (347)\ttotal: 51.2s\tremaining: 1m 31s\n",
      "359:\tlearn: 0.0094719\ttest: 0.2499844\tbest: 0.2463868 (347)\ttotal: 51.3s\tremaining: 1m 31s\n",
      "360:\tlearn: 0.0094103\ttest: 0.2495413\tbest: 0.2463868 (347)\ttotal: 51.5s\tremaining: 1m 31s\n",
      "361:\tlearn: 0.0093786\ttest: 0.2491476\tbest: 0.2463868 (347)\ttotal: 51.6s\tremaining: 1m 30s\n",
      "362:\tlearn: 0.0092781\ttest: 0.2490000\tbest: 0.2463868 (347)\ttotal: 51.8s\tremaining: 1m 30s\n",
      "363:\tlearn: 0.0092751\ttest: 0.2492042\tbest: 0.2463868 (347)\ttotal: 51.9s\tremaining: 1m 30s\n",
      "364:\tlearn: 0.0092462\ttest: 0.2494924\tbest: 0.2463868 (347)\ttotal: 52s\tremaining: 1m 30s\n",
      "365:\tlearn: 0.0091665\ttest: 0.2488917\tbest: 0.2463868 (347)\ttotal: 52.2s\tremaining: 1m 30s\n",
      "366:\tlearn: 0.0091568\ttest: 0.2484489\tbest: 0.2463868 (347)\ttotal: 52.3s\tremaining: 1m 30s\n",
      "367:\tlearn: 0.0091126\ttest: 0.2481630\tbest: 0.2463868 (347)\ttotal: 52.5s\tremaining: 1m 30s\n",
      "368:\tlearn: 0.0090187\ttest: 0.2482610\tbest: 0.2463868 (347)\ttotal: 52.6s\tremaining: 1m 29s\n",
      "369:\tlearn: 0.0089864\ttest: 0.2478930\tbest: 0.2463868 (347)\ttotal: 52.8s\tremaining: 1m 29s\n",
      "370:\tlearn: 0.0089753\ttest: 0.2480367\tbest: 0.2463868 (347)\ttotal: 52.9s\tremaining: 1m 29s\n",
      "371:\tlearn: 0.0089496\ttest: 0.2476122\tbest: 0.2463868 (347)\ttotal: 53s\tremaining: 1m 29s\n",
      "372:\tlearn: 0.0089255\ttest: 0.2481745\tbest: 0.2463868 (347)\ttotal: 53.2s\tremaining: 1m 29s\n",
      "373:\tlearn: 0.0089157\ttest: 0.2479834\tbest: 0.2463868 (347)\ttotal: 53.3s\tremaining: 1m 29s\n",
      "374:\tlearn: 0.0088910\ttest: 0.2481216\tbest: 0.2463868 (347)\ttotal: 53.5s\tremaining: 1m 29s\n",
      "375:\tlearn: 0.0088653\ttest: 0.2478217\tbest: 0.2463868 (347)\ttotal: 53.6s\tremaining: 1m 29s\n",
      "376:\tlearn: 0.0088585\ttest: 0.2479921\tbest: 0.2463868 (347)\ttotal: 53.8s\tremaining: 1m 28s\n",
      "377:\tlearn: 0.0088466\ttest: 0.2476383\tbest: 0.2463868 (347)\ttotal: 53.9s\tremaining: 1m 28s\n",
      "378:\tlearn: 0.0087784\ttest: 0.2478080\tbest: 0.2463868 (347)\ttotal: 54.1s\tremaining: 1m 28s\n",
      "379:\tlearn: 0.0087762\ttest: 0.2477795\tbest: 0.2463868 (347)\ttotal: 54.2s\tremaining: 1m 28s\n",
      "380:\tlearn: 0.0087228\ttest: 0.2477628\tbest: 0.2463868 (347)\ttotal: 54.4s\tremaining: 1m 28s\n",
      "381:\tlearn: 0.0086337\ttest: 0.2491863\tbest: 0.2463868 (347)\ttotal: 54.5s\tremaining: 1m 28s\n",
      "382:\tlearn: 0.0086207\ttest: 0.2493294\tbest: 0.2463868 (347)\ttotal: 54.7s\tremaining: 1m 28s\n",
      "383:\tlearn: 0.0085872\ttest: 0.2484787\tbest: 0.2463868 (347)\ttotal: 54.8s\tremaining: 1m 27s\n",
      "384:\tlearn: 0.0085788\ttest: 0.2483600\tbest: 0.2463868 (347)\ttotal: 54.9s\tremaining: 1m 27s\n",
      "385:\tlearn: 0.0085507\ttest: 0.2483615\tbest: 0.2463868 (347)\ttotal: 55.1s\tremaining: 1m 27s\n",
      "386:\tlearn: 0.0084902\ttest: 0.2496316\tbest: 0.2463868 (347)\ttotal: 55.2s\tremaining: 1m 27s\n",
      "387:\tlearn: 0.0084337\ttest: 0.2498414\tbest: 0.2463868 (347)\ttotal: 55.4s\tremaining: 1m 27s\n",
      "388:\tlearn: 0.0084109\ttest: 0.2503054\tbest: 0.2463868 (347)\ttotal: 55.5s\tremaining: 1m 27s\n",
      "389:\tlearn: 0.0083566\ttest: 0.2501047\tbest: 0.2463868 (347)\ttotal: 55.7s\tremaining: 1m 27s\n",
      "390:\tlearn: 0.0083229\ttest: 0.2501184\tbest: 0.2463868 (347)\ttotal: 55.8s\tremaining: 1m 26s\n",
      "391:\tlearn: 0.0082598\ttest: 0.2502575\tbest: 0.2463868 (347)\ttotal: 56s\tremaining: 1m 26s\n",
      "392:\tlearn: 0.0082185\ttest: 0.2506794\tbest: 0.2463868 (347)\ttotal: 56.1s\tremaining: 1m 26s\n",
      "393:\tlearn: 0.0081883\ttest: 0.2503708\tbest: 0.2463868 (347)\ttotal: 56.2s\tremaining: 1m 26s\n",
      "394:\tlearn: 0.0081527\ttest: 0.2494595\tbest: 0.2463868 (347)\ttotal: 56.4s\tremaining: 1m 26s\n",
      "395:\tlearn: 0.0081122\ttest: 0.2519916\tbest: 0.2463868 (347)\ttotal: 56.5s\tremaining: 1m 26s\n",
      "396:\tlearn: 0.0080750\ttest: 0.2522601\tbest: 0.2463868 (347)\ttotal: 56.7s\tremaining: 1m 26s\n",
      "397:\tlearn: 0.0080521\ttest: 0.2522998\tbest: 0.2463868 (347)\ttotal: 56.9s\tremaining: 1m 25s\n",
      "398:\tlearn: 0.0080387\ttest: 0.2525732\tbest: 0.2463868 (347)\ttotal: 57s\tremaining: 1m 25s\n",
      "399:\tlearn: 0.0080211\ttest: 0.2532785\tbest: 0.2463868 (347)\ttotal: 57.2s\tremaining: 1m 25s\n",
      "400:\tlearn: 0.0080130\ttest: 0.2528564\tbest: 0.2463868 (347)\ttotal: 57.3s\tremaining: 1m 25s\n",
      "401:\tlearn: 0.0079797\ttest: 0.2539778\tbest: 0.2463868 (347)\ttotal: 57.5s\tremaining: 1m 25s\n",
      "402:\tlearn: 0.0079524\ttest: 0.2540898\tbest: 0.2463868 (347)\ttotal: 57.6s\tremaining: 1m 25s\n",
      "403:\tlearn: 0.0079074\ttest: 0.2523683\tbest: 0.2463868 (347)\ttotal: 57.7s\tremaining: 1m 25s\n",
      "404:\tlearn: 0.0078788\ttest: 0.2514958\tbest: 0.2463868 (347)\ttotal: 57.9s\tremaining: 1m 25s\n",
      "405:\tlearn: 0.0078743\ttest: 0.2515648\tbest: 0.2463868 (347)\ttotal: 58.1s\tremaining: 1m 24s\n",
      "406:\tlearn: 0.0078615\ttest: 0.2516751\tbest: 0.2463868 (347)\ttotal: 58.2s\tremaining: 1m 24s\n",
      "407:\tlearn: 0.0078128\ttest: 0.2536154\tbest: 0.2463868 (347)\ttotal: 58.4s\tremaining: 1m 24s\n",
      "408:\tlearn: 0.0077836\ttest: 0.2537220\tbest: 0.2463868 (347)\ttotal: 58.5s\tremaining: 1m 24s\n",
      "409:\tlearn: 0.0077005\ttest: 0.2536105\tbest: 0.2463868 (347)\ttotal: 58.6s\tremaining: 1m 24s\n",
      "410:\tlearn: 0.0076536\ttest: 0.2529505\tbest: 0.2463868 (347)\ttotal: 58.8s\tremaining: 1m 24s\n",
      "411:\tlearn: 0.0076451\ttest: 0.2520279\tbest: 0.2463868 (347)\ttotal: 58.9s\tremaining: 1m 24s\n",
      "412:\tlearn: 0.0076336\ttest: 0.2522109\tbest: 0.2463868 (347)\ttotal: 59.1s\tremaining: 1m 23s\n",
      "413:\tlearn: 0.0075985\ttest: 0.2505802\tbest: 0.2463868 (347)\ttotal: 59.2s\tremaining: 1m 23s\n",
      "414:\tlearn: 0.0075813\ttest: 0.2505460\tbest: 0.2463868 (347)\ttotal: 59.3s\tremaining: 1m 23s\n",
      "415:\tlearn: 0.0075710\ttest: 0.2503797\tbest: 0.2463868 (347)\ttotal: 59.5s\tremaining: 1m 23s\n",
      "416:\tlearn: 0.0075036\ttest: 0.2493091\tbest: 0.2463868 (347)\ttotal: 59.6s\tremaining: 1m 23s\n",
      "417:\tlearn: 0.0074832\ttest: 0.2493938\tbest: 0.2463868 (347)\ttotal: 59.8s\tremaining: 1m 23s\n",
      "418:\tlearn: 0.0074708\ttest: 0.2494540\tbest: 0.2463868 (347)\ttotal: 59.9s\tremaining: 1m 23s\n",
      "419:\tlearn: 0.0074621\ttest: 0.2491306\tbest: 0.2463868 (347)\ttotal: 1m\tremaining: 1m 22s\n",
      "420:\tlearn: 0.0074446\ttest: 0.2495807\tbest: 0.2463868 (347)\ttotal: 1m\tremaining: 1m 22s\n",
      "421:\tlearn: 0.0074344\ttest: 0.2500577\tbest: 0.2463868 (347)\ttotal: 1m\tremaining: 1m 22s\n",
      "422:\tlearn: 0.0074269\ttest: 0.2503194\tbest: 0.2463868 (347)\ttotal: 1m\tremaining: 1m 22s\n",
      "423:\tlearn: 0.0074054\ttest: 0.2503679\tbest: 0.2463868 (347)\ttotal: 1m\tremaining: 1m 22s\n",
      "424:\tlearn: 0.0073901\ttest: 0.2501710\tbest: 0.2463868 (347)\ttotal: 1m\tremaining: 1m 22s\n",
      "425:\tlearn: 0.0073516\ttest: 0.2505964\tbest: 0.2463868 (347)\ttotal: 1m\tremaining: 1m 22s\n",
      "426:\tlearn: 0.0072960\ttest: 0.2506396\tbest: 0.2463868 (347)\ttotal: 1m 1s\tremaining: 1m 21s\n",
      "427:\tlearn: 0.0072781\ttest: 0.2507422\tbest: 0.2463868 (347)\ttotal: 1m 1s\tremaining: 1m 21s\n",
      "428:\tlearn: 0.0072214\ttest: 0.2527098\tbest: 0.2463868 (347)\ttotal: 1m 1s\tremaining: 1m 21s\n",
      "429:\tlearn: 0.0072056\ttest: 0.2520212\tbest: 0.2463868 (347)\ttotal: 1m 1s\tremaining: 1m 21s\n",
      "430:\tlearn: 0.0071920\ttest: 0.2524643\tbest: 0.2463868 (347)\ttotal: 1m 1s\tremaining: 1m 21s\n",
      "431:\tlearn: 0.0071802\ttest: 0.2520206\tbest: 0.2463868 (347)\ttotal: 1m 1s\tremaining: 1m 21s\n",
      "432:\tlearn: 0.0071728\ttest: 0.2516589\tbest: 0.2463868 (347)\ttotal: 1m 1s\tremaining: 1m 21s\n",
      "433:\tlearn: 0.0071695\ttest: 0.2517712\tbest: 0.2463868 (347)\ttotal: 1m 2s\tremaining: 1m 20s\n",
      "434:\tlearn: 0.0071466\ttest: 0.2524255\tbest: 0.2463868 (347)\ttotal: 1m 2s\tremaining: 1m 20s\n",
      "435:\tlearn: 0.0070930\ttest: 0.2533018\tbest: 0.2463868 (347)\ttotal: 1m 2s\tremaining: 1m 20s\n",
      "436:\tlearn: 0.0070775\ttest: 0.2529748\tbest: 0.2463868 (347)\ttotal: 1m 2s\tremaining: 1m 20s\n",
      "437:\tlearn: 0.0070653\ttest: 0.2534750\tbest: 0.2463868 (347)\ttotal: 1m 2s\tremaining: 1m 20s\n",
      "438:\tlearn: 0.0070598\ttest: 0.2534687\tbest: 0.2463868 (347)\ttotal: 1m 2s\tremaining: 1m 20s\n",
      "439:\tlearn: 0.0070277\ttest: 0.2533135\tbest: 0.2463868 (347)\ttotal: 1m 2s\tremaining: 1m 20s\n",
      "440:\tlearn: 0.0069980\ttest: 0.2534350\tbest: 0.2463868 (347)\ttotal: 1m 3s\tremaining: 1m 19s\n",
      "441:\tlearn: 0.0069681\ttest: 0.2542328\tbest: 0.2463868 (347)\ttotal: 1m 3s\tremaining: 1m 19s\n",
      "442:\tlearn: 0.0068889\ttest: 0.2547277\tbest: 0.2463868 (347)\ttotal: 1m 3s\tremaining: 1m 19s\n",
      "443:\tlearn: 0.0068844\ttest: 0.2548999\tbest: 0.2463868 (347)\ttotal: 1m 3s\tremaining: 1m 19s\n",
      "444:\tlearn: 0.0068404\ttest: 0.2552262\tbest: 0.2463868 (347)\ttotal: 1m 3s\tremaining: 1m 19s\n",
      "445:\tlearn: 0.0068051\ttest: 0.2547610\tbest: 0.2463868 (347)\ttotal: 1m 3s\tremaining: 1m 19s\n",
      "446:\tlearn: 0.0067915\ttest: 0.2549354\tbest: 0.2463868 (347)\ttotal: 1m 3s\tremaining: 1m 19s\n",
      "447:\tlearn: 0.0067859\ttest: 0.2548513\tbest: 0.2463868 (347)\ttotal: 1m 4s\tremaining: 1m 18s\n",
      "bestTest = 0.2463867848\n",
      "bestIteration = 347\n",
      "Shrink model to first 348 iterations.\n"
     ]
    }
   ],
   "source": [
    "task_type = \"GPU\" if torch.cuda.is_available() else \"CPU\"\n",
    "clf = CatBoostClassifier(\n",
    "    iterations=1000,\n",
    "    early_stopping_rounds=100,\n",
    "    task_type=task_type,\n",
    "    random_seed=42,\n",
    "    use_best_model=True,\n",
    "    auto_class_weights=\"Balanced\",\n",
    ")\n",
    "clf = clf.fit(\n",
    "    train_embs.cpu().numpy(),\n",
    "    y_train.cpu().numpy(),\n",
    "    eval_set=(val_embs.cpu().numpy(), y_val.cpu().numpy()),\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.96      0.88      0.91        49\n",
      "           1       0.74      0.89      0.81        19\n",
      "\n",
      "    accuracy                           0.88        68\n",
      "   macro avg       0.85      0.89      0.86        68\n",
      "weighted avg       0.90      0.88      0.89        68\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(clf.predict(test_embs.cpu().numpy()), data.test_y.cpu().numpy()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1/1 [00:13<00:00, 13.30s/it]\n",
      "100%|██████████| 1/1 [00:13<00:00, 13.24s/it]\n",
      "100%|██████████| 1/1 [00:13<00:00, 13.30s/it]\n",
      "Warning: less than 75% gpu memory available for training. Free: 6389.4375 Total: 11178.5\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Learning rate set to 0.080278\n",
      "0:\tlearn: 0.6538378\ttest: 0.6605008\tbest: 0.6605008 (0)\ttotal: 142ms\tremaining: 2m 22s\n",
      "1:\tlearn: 0.6171080\ttest: 0.6224138\tbest: 0.6224138 (1)\ttotal: 283ms\tremaining: 2m 21s\n",
      "2:\tlearn: 0.5819143\ttest: 0.6069669\tbest: 0.6069669 (2)\ttotal: 418ms\tremaining: 2m 18s\n",
      "3:\tlearn: 0.5595998\ttest: 0.5866527\tbest: 0.5866527 (3)\ttotal: 555ms\tremaining: 2m 18s\n",
      "4:\tlearn: 0.5399098\ttest: 0.5812235\tbest: 0.5812235 (4)\ttotal: 699ms\tremaining: 2m 19s\n",
      "5:\tlearn: 0.5284173\ttest: 0.5648007\tbest: 0.5648007 (5)\ttotal: 839ms\tremaining: 2m 19s\n",
      "6:\tlearn: 0.5058470\ttest: 0.5629640\tbest: 0.5629640 (6)\ttotal: 979ms\tremaining: 2m 18s\n",
      "7:\tlearn: 0.4844777\ttest: 0.5601503\tbest: 0.5601503 (7)\ttotal: 1.12s\tremaining: 2m 18s\n",
      "8:\tlearn: 0.4593012\ttest: 0.5369273\tbest: 0.5369273 (8)\ttotal: 1.27s\tremaining: 2m 19s\n",
      "9:\tlearn: 0.4459104\ttest: 0.5212256\tbest: 0.5212256 (9)\ttotal: 1.41s\tremaining: 2m 19s\n",
      "10:\tlearn: 0.4331912\ttest: 0.5184409\tbest: 0.5184409 (10)\ttotal: 1.56s\tremaining: 2m 19s\n",
      "11:\tlearn: 0.4240206\ttest: 0.5131389\tbest: 0.5131389 (11)\ttotal: 1.7s\tremaining: 2m 19s\n",
      "12:\tlearn: 0.4116658\ttest: 0.5120075\tbest: 0.5120075 (12)\ttotal: 1.83s\tremaining: 2m 19s\n",
      "13:\tlearn: 0.3969961\ttest: 0.4919836\tbest: 0.4919836 (13)\ttotal: 1.98s\tremaining: 2m 19s\n",
      "14:\tlearn: 0.3901306\ttest: 0.4937906\tbest: 0.4919836 (13)\ttotal: 2.12s\tremaining: 2m 19s\n",
      "15:\tlearn: 0.3806338\ttest: 0.4842854\tbest: 0.4842854 (15)\ttotal: 2.26s\tremaining: 2m 18s\n",
      "16:\tlearn: 0.3665703\ttest: 0.4843272\tbest: 0.4842854 (15)\ttotal: 2.39s\tremaining: 2m 18s\n",
      "17:\tlearn: 0.3554519\ttest: 0.4703519\tbest: 0.4703519 (17)\ttotal: 2.53s\tremaining: 2m 18s\n",
      "18:\tlearn: 0.3500488\ttest: 0.4671425\tbest: 0.4671425 (18)\ttotal: 2.67s\tremaining: 2m 17s\n",
      "19:\tlearn: 0.3358262\ttest: 0.4544250\tbest: 0.4544250 (19)\ttotal: 2.81s\tremaining: 2m 17s\n",
      "20:\tlearn: 0.3305343\ttest: 0.4547336\tbest: 0.4544250 (19)\ttotal: 2.94s\tremaining: 2m 17s\n",
      "21:\tlearn: 0.3250376\ttest: 0.4507987\tbest: 0.4507987 (21)\ttotal: 3.07s\tremaining: 2m 16s\n",
      "22:\tlearn: 0.3195861\ttest: 0.4496883\tbest: 0.4496883 (22)\ttotal: 3.21s\tremaining: 2m 16s\n",
      "23:\tlearn: 0.3124007\ttest: 0.4494263\tbest: 0.4494263 (23)\ttotal: 3.35s\tremaining: 2m 16s\n",
      "24:\tlearn: 0.3039500\ttest: 0.4446065\tbest: 0.4446065 (24)\ttotal: 3.49s\tremaining: 2m 16s\n",
      "25:\tlearn: 0.2962702\ttest: 0.4372247\tbest: 0.4372247 (25)\ttotal: 3.63s\tremaining: 2m 16s\n",
      "26:\tlearn: 0.2910181\ttest: 0.4398968\tbest: 0.4372247 (25)\ttotal: 3.78s\tremaining: 2m 16s\n",
      "27:\tlearn: 0.2829489\ttest: 0.4294043\tbest: 0.4294043 (27)\ttotal: 3.92s\tremaining: 2m 15s\n",
      "28:\tlearn: 0.2787090\ttest: 0.4257016\tbest: 0.4257016 (28)\ttotal: 4.05s\tremaining: 2m 15s\n",
      "29:\tlearn: 0.2718640\ttest: 0.4255345\tbest: 0.4255345 (29)\ttotal: 4.19s\tremaining: 2m 15s\n",
      "30:\tlearn: 0.2645727\ttest: 0.4296659\tbest: 0.4255345 (29)\ttotal: 4.33s\tremaining: 2m 15s\n",
      "31:\tlearn: 0.2548197\ttest: 0.4338701\tbest: 0.4255345 (29)\ttotal: 4.47s\tremaining: 2m 15s\n",
      "32:\tlearn: 0.2485479\ttest: 0.4312503\tbest: 0.4255345 (29)\ttotal: 4.61s\tremaining: 2m 15s\n",
      "33:\tlearn: 0.2428647\ttest: 0.4277478\tbest: 0.4255345 (29)\ttotal: 4.75s\tremaining: 2m 15s\n",
      "34:\tlearn: 0.2370475\ttest: 0.4247740\tbest: 0.4247740 (34)\ttotal: 4.89s\tremaining: 2m 14s\n",
      "35:\tlearn: 0.2321884\ttest: 0.4218834\tbest: 0.4218834 (35)\ttotal: 5.03s\tremaining: 2m 14s\n",
      "36:\tlearn: 0.2284110\ttest: 0.4184027\tbest: 0.4184027 (36)\ttotal: 5.17s\tremaining: 2m 14s\n",
      "37:\tlearn: 0.2222448\ttest: 0.4124408\tbest: 0.4124408 (37)\ttotal: 5.32s\tremaining: 2m 14s\n",
      "38:\tlearn: 0.2203592\ttest: 0.4103889\tbest: 0.4103889 (38)\ttotal: 5.45s\tremaining: 2m 14s\n",
      "39:\tlearn: 0.2176135\ttest: 0.4110621\tbest: 0.4103889 (38)\ttotal: 5.58s\tremaining: 2m 14s\n",
      "40:\tlearn: 0.2138759\ttest: 0.4082875\tbest: 0.4082875 (40)\ttotal: 5.72s\tremaining: 2m 13s\n",
      "41:\tlearn: 0.2082849\ttest: 0.4072212\tbest: 0.4072212 (41)\ttotal: 5.85s\tremaining: 2m 13s\n",
      "42:\tlearn: 0.2063867\ttest: 0.4051114\tbest: 0.4051114 (42)\ttotal: 5.99s\tremaining: 2m 13s\n",
      "43:\tlearn: 0.2037525\ttest: 0.4061551\tbest: 0.4051114 (42)\ttotal: 6.12s\tremaining: 2m 13s\n",
      "44:\tlearn: 0.2010851\ttest: 0.4029345\tbest: 0.4029345 (44)\ttotal: 6.26s\tremaining: 2m 12s\n",
      "45:\tlearn: 0.1978036\ttest: 0.3986981\tbest: 0.3986981 (45)\ttotal: 6.4s\tremaining: 2m 12s\n",
      "46:\tlearn: 0.1958619\ttest: 0.3981579\tbest: 0.3981579 (46)\ttotal: 6.54s\tremaining: 2m 12s\n",
      "47:\tlearn: 0.1892164\ttest: 0.4004138\tbest: 0.3981579 (46)\ttotal: 6.68s\tremaining: 2m 12s\n",
      "48:\tlearn: 0.1859059\ttest: 0.3954145\tbest: 0.3954145 (48)\ttotal: 6.82s\tremaining: 2m 12s\n",
      "49:\tlearn: 0.1839587\ttest: 0.3945769\tbest: 0.3945769 (49)\ttotal: 6.96s\tremaining: 2m 12s\n",
      "50:\tlearn: 0.1793802\ttest: 0.3927058\tbest: 0.3927058 (50)\ttotal: 7.1s\tremaining: 2m 12s\n",
      "51:\tlearn: 0.1755745\ttest: 0.3875274\tbest: 0.3875274 (51)\ttotal: 7.25s\tremaining: 2m 12s\n",
      "52:\tlearn: 0.1721384\ttest: 0.3859242\tbest: 0.3859242 (52)\ttotal: 7.39s\tremaining: 2m 12s\n",
      "53:\tlearn: 0.1693080\ttest: 0.3823808\tbest: 0.3823808 (53)\ttotal: 7.52s\tremaining: 2m 11s\n",
      "54:\tlearn: 0.1675544\ttest: 0.3808015\tbest: 0.3808015 (54)\ttotal: 7.67s\tremaining: 2m 11s\n",
      "55:\tlearn: 0.1662455\ttest: 0.3790084\tbest: 0.3790084 (55)\ttotal: 7.81s\tremaining: 2m 11s\n",
      "56:\tlearn: 0.1654383\ttest: 0.3756252\tbest: 0.3756252 (56)\ttotal: 7.95s\tremaining: 2m 11s\n",
      "57:\tlearn: 0.1609758\ttest: 0.3795744\tbest: 0.3756252 (56)\ttotal: 8.09s\tremaining: 2m 11s\n",
      "58:\tlearn: 0.1588093\ttest: 0.3791328\tbest: 0.3756252 (56)\ttotal: 8.23s\tremaining: 2m 11s\n",
      "59:\tlearn: 0.1567737\ttest: 0.3718969\tbest: 0.3718969 (59)\ttotal: 8.36s\tremaining: 2m 11s\n",
      "60:\tlearn: 0.1546771\ttest: 0.3743691\tbest: 0.3718969 (59)\ttotal: 8.51s\tremaining: 2m 10s\n",
      "61:\tlearn: 0.1526292\ttest: 0.3724335\tbest: 0.3718969 (59)\ttotal: 8.65s\tremaining: 2m 10s\n",
      "62:\tlearn: 0.1494248\ttest: 0.3687534\tbest: 0.3687534 (62)\ttotal: 8.79s\tremaining: 2m 10s\n",
      "63:\tlearn: 0.1466038\ttest: 0.3691694\tbest: 0.3687534 (62)\ttotal: 8.93s\tremaining: 2m 10s\n",
      "64:\tlearn: 0.1448571\ttest: 0.3693909\tbest: 0.3687534 (62)\ttotal: 9.07s\tremaining: 2m 10s\n",
      "65:\tlearn: 0.1435678\ttest: 0.3683980\tbest: 0.3683980 (65)\ttotal: 9.27s\tremaining: 2m 11s\n",
      "66:\tlearn: 0.1411530\ttest: 0.3664138\tbest: 0.3664138 (66)\ttotal: 9.42s\tremaining: 2m 11s\n",
      "67:\tlearn: 0.1372547\ttest: 0.3636630\tbest: 0.3636630 (67)\ttotal: 9.57s\tremaining: 2m 11s\n",
      "68:\tlearn: 0.1338780\ttest: 0.3632165\tbest: 0.3632165 (68)\ttotal: 9.72s\tremaining: 2m 11s\n",
      "69:\tlearn: 0.1323509\ttest: 0.3611028\tbest: 0.3611028 (69)\ttotal: 9.88s\tremaining: 2m 11s\n",
      "70:\tlearn: 0.1302592\ttest: 0.3618101\tbest: 0.3611028 (69)\ttotal: 10s\tremaining: 2m 11s\n",
      "71:\tlearn: 0.1289592\ttest: 0.3595735\tbest: 0.3595735 (71)\ttotal: 10.2s\tremaining: 2m 11s\n",
      "72:\tlearn: 0.1280238\ttest: 0.3612779\tbest: 0.3595735 (71)\ttotal: 10.3s\tremaining: 2m 10s\n",
      "73:\tlearn: 0.1257951\ttest: 0.3590102\tbest: 0.3590102 (73)\ttotal: 10.5s\tremaining: 2m 11s\n",
      "74:\tlearn: 0.1242672\ttest: 0.3581445\tbest: 0.3581445 (74)\ttotal: 10.7s\tremaining: 2m 11s\n",
      "75:\tlearn: 0.1222948\ttest: 0.3586298\tbest: 0.3581445 (74)\ttotal: 10.8s\tremaining: 2m 11s\n",
      "76:\tlearn: 0.1202608\ttest: 0.3587142\tbest: 0.3581445 (74)\ttotal: 10.9s\tremaining: 2m 11s\n",
      "77:\tlearn: 0.1162712\ttest: 0.3644345\tbest: 0.3581445 (74)\ttotal: 11.1s\tremaining: 2m 11s\n",
      "78:\tlearn: 0.1137296\ttest: 0.3669014\tbest: 0.3581445 (74)\ttotal: 11.2s\tremaining: 2m 10s\n",
      "79:\tlearn: 0.1108229\ttest: 0.3647821\tbest: 0.3581445 (74)\ttotal: 11.4s\tremaining: 2m 10s\n",
      "80:\tlearn: 0.1093501\ttest: 0.3636967\tbest: 0.3581445 (74)\ttotal: 11.5s\tremaining: 2m 10s\n",
      "81:\tlearn: 0.1076778\ttest: 0.3608170\tbest: 0.3581445 (74)\ttotal: 11.7s\tremaining: 2m 10s\n",
      "82:\tlearn: 0.1068973\ttest: 0.3597146\tbest: 0.3581445 (74)\ttotal: 11.8s\tremaining: 2m 10s\n",
      "83:\tlearn: 0.1045565\ttest: 0.3635658\tbest: 0.3581445 (74)\ttotal: 12s\tremaining: 2m 10s\n",
      "84:\tlearn: 0.1024326\ttest: 0.3605812\tbest: 0.3581445 (74)\ttotal: 12.1s\tremaining: 2m 10s\n",
      "85:\tlearn: 0.0996299\ttest: 0.3611368\tbest: 0.3581445 (74)\ttotal: 12.3s\tremaining: 2m 10s\n",
      "86:\tlearn: 0.0984389\ttest: 0.3624161\tbest: 0.3581445 (74)\ttotal: 12.4s\tremaining: 2m 10s\n",
      "87:\tlearn: 0.0972632\ttest: 0.3613778\tbest: 0.3581445 (74)\ttotal: 12.5s\tremaining: 2m 9s\n",
      "88:\tlearn: 0.0967492\ttest: 0.3623160\tbest: 0.3581445 (74)\ttotal: 12.7s\tremaining: 2m 9s\n",
      "89:\tlearn: 0.0963417\ttest: 0.3626883\tbest: 0.3581445 (74)\ttotal: 12.8s\tremaining: 2m 9s\n",
      "90:\tlearn: 0.0941493\ttest: 0.3578558\tbest: 0.3578558 (90)\ttotal: 13s\tremaining: 2m 9s\n",
      "91:\tlearn: 0.0917821\ttest: 0.3582127\tbest: 0.3578558 (90)\ttotal: 13.1s\tremaining: 2m 9s\n",
      "92:\tlearn: 0.0886817\ttest: 0.3555941\tbest: 0.3555941 (92)\ttotal: 13.3s\tremaining: 2m 9s\n",
      "93:\tlearn: 0.0881765\ttest: 0.3547043\tbest: 0.3547043 (93)\ttotal: 13.4s\tremaining: 2m 9s\n",
      "94:\tlearn: 0.0870542\ttest: 0.3570531\tbest: 0.3547043 (93)\ttotal: 13.5s\tremaining: 2m 9s\n",
      "95:\tlearn: 0.0857023\ttest: 0.3567908\tbest: 0.3547043 (93)\ttotal: 13.7s\tremaining: 2m 8s\n",
      "96:\tlearn: 0.0844521\ttest: 0.3553384\tbest: 0.3547043 (93)\ttotal: 13.8s\tremaining: 2m 8s\n",
      "97:\tlearn: 0.0828094\ttest: 0.3518886\tbest: 0.3518886 (97)\ttotal: 14s\tremaining: 2m 8s\n",
      "98:\tlearn: 0.0814604\ttest: 0.3532217\tbest: 0.3518886 (97)\ttotal: 14.1s\tremaining: 2m 8s\n",
      "99:\tlearn: 0.0793539\ttest: 0.3508416\tbest: 0.3508416 (99)\ttotal: 14.3s\tremaining: 2m 8s\n",
      "100:\tlearn: 0.0779341\ttest: 0.3497940\tbest: 0.3497940 (100)\ttotal: 14.4s\tremaining: 2m 8s\n",
      "101:\tlearn: 0.0755171\ttest: 0.3496130\tbest: 0.3496130 (101)\ttotal: 14.6s\tremaining: 2m 8s\n",
      "102:\tlearn: 0.0737236\ttest: 0.3515202\tbest: 0.3496130 (101)\ttotal: 14.7s\tremaining: 2m 8s\n",
      "103:\tlearn: 0.0723779\ttest: 0.3534327\tbest: 0.3496130 (101)\ttotal: 14.9s\tremaining: 2m 7s\n",
      "104:\tlearn: 0.0720375\ttest: 0.3536474\tbest: 0.3496130 (101)\ttotal: 15s\tremaining: 2m 7s\n",
      "105:\tlearn: 0.0705402\ttest: 0.3582266\tbest: 0.3496130 (101)\ttotal: 15.1s\tremaining: 2m 7s\n",
      "106:\tlearn: 0.0687804\ttest: 0.3534804\tbest: 0.3496130 (101)\ttotal: 15.3s\tremaining: 2m 7s\n",
      "107:\tlearn: 0.0667705\ttest: 0.3528885\tbest: 0.3496130 (101)\ttotal: 15.4s\tremaining: 2m 7s\n",
      "108:\tlearn: 0.0655098\ttest: 0.3512380\tbest: 0.3496130 (101)\ttotal: 15.6s\tremaining: 2m 7s\n",
      "109:\tlearn: 0.0643664\ttest: 0.3524885\tbest: 0.3496130 (101)\ttotal: 15.7s\tremaining: 2m 7s\n",
      "110:\tlearn: 0.0628404\ttest: 0.3542997\tbest: 0.3496130 (101)\ttotal: 15.9s\tremaining: 2m 7s\n",
      "111:\tlearn: 0.0611051\ttest: 0.3553097\tbest: 0.3496130 (101)\ttotal: 16s\tremaining: 2m 6s\n",
      "112:\tlearn: 0.0609040\ttest: 0.3540795\tbest: 0.3496130 (101)\ttotal: 16.1s\tremaining: 2m 6s\n",
      "113:\tlearn: 0.0587266\ttest: 0.3533935\tbest: 0.3496130 (101)\ttotal: 16.3s\tremaining: 2m 6s\n",
      "114:\tlearn: 0.0579257\ttest: 0.3547609\tbest: 0.3496130 (101)\ttotal: 16.4s\tremaining: 2m 6s\n",
      "115:\tlearn: 0.0565346\ttest: 0.3562196\tbest: 0.3496130 (101)\ttotal: 16.6s\tremaining: 2m 6s\n",
      "116:\tlearn: 0.0557313\ttest: 0.3546250\tbest: 0.3496130 (101)\ttotal: 16.7s\tremaining: 2m 6s\n",
      "117:\tlearn: 0.0546437\ttest: 0.3575876\tbest: 0.3496130 (101)\ttotal: 16.9s\tremaining: 2m 6s\n",
      "118:\tlearn: 0.0534114\ttest: 0.3549399\tbest: 0.3496130 (101)\ttotal: 17s\tremaining: 2m 5s\n",
      "119:\tlearn: 0.0531509\ttest: 0.3569741\tbest: 0.3496130 (101)\ttotal: 17.2s\tremaining: 2m 5s\n",
      "120:\tlearn: 0.0523275\ttest: 0.3569946\tbest: 0.3496130 (101)\ttotal: 17.3s\tremaining: 2m 5s\n",
      "121:\tlearn: 0.0508160\ttest: 0.3490908\tbest: 0.3490908 (121)\ttotal: 17.4s\tremaining: 2m 5s\n",
      "122:\tlearn: 0.0499428\ttest: 0.3513512\tbest: 0.3490908 (121)\ttotal: 17.6s\tremaining: 2m 5s\n",
      "123:\tlearn: 0.0486890\ttest: 0.3490993\tbest: 0.3490908 (121)\ttotal: 17.7s\tremaining: 2m 5s\n",
      "124:\tlearn: 0.0481393\ttest: 0.3498925\tbest: 0.3490908 (121)\ttotal: 17.9s\tremaining: 2m 5s\n",
      "125:\tlearn: 0.0471562\ttest: 0.3483495\tbest: 0.3483495 (125)\ttotal: 18s\tremaining: 2m 4s\n",
      "126:\tlearn: 0.0465038\ttest: 0.3448544\tbest: 0.3448544 (126)\ttotal: 18.2s\tremaining: 2m 4s\n",
      "127:\tlearn: 0.0460258\ttest: 0.3471332\tbest: 0.3448544 (126)\ttotal: 18.3s\tremaining: 2m 4s\n",
      "128:\tlearn: 0.0456782\ttest: 0.3476852\tbest: 0.3448544 (126)\ttotal: 18.4s\tremaining: 2m 4s\n",
      "129:\tlearn: 0.0450248\ttest: 0.3476020\tbest: 0.3448544 (126)\ttotal: 18.6s\tremaining: 2m 4s\n",
      "130:\tlearn: 0.0441695\ttest: 0.3525059\tbest: 0.3448544 (126)\ttotal: 18.7s\tremaining: 2m 4s\n",
      "131:\tlearn: 0.0437505\ttest: 0.3531608\tbest: 0.3448544 (126)\ttotal: 18.9s\tremaining: 2m 4s\n",
      "132:\tlearn: 0.0426379\ttest: 0.3515638\tbest: 0.3448544 (126)\ttotal: 19s\tremaining: 2m 4s\n",
      "133:\tlearn: 0.0423010\ttest: 0.3503911\tbest: 0.3448544 (126)\ttotal: 19.2s\tremaining: 2m 3s\n",
      "134:\tlearn: 0.0414037\ttest: 0.3498752\tbest: 0.3448544 (126)\ttotal: 19.3s\tremaining: 2m 3s\n",
      "135:\tlearn: 0.0409608\ttest: 0.3506309\tbest: 0.3448544 (126)\ttotal: 19.5s\tremaining: 2m 3s\n",
      "136:\tlearn: 0.0404415\ttest: 0.3520270\tbest: 0.3448544 (126)\ttotal: 19.6s\tremaining: 2m 3s\n",
      "137:\tlearn: 0.0401257\ttest: 0.3497002\tbest: 0.3448544 (126)\ttotal: 19.7s\tremaining: 2m 3s\n",
      "138:\tlearn: 0.0394153\ttest: 0.3507382\tbest: 0.3448544 (126)\ttotal: 19.9s\tremaining: 2m 3s\n",
      "139:\tlearn: 0.0393574\ttest: 0.3514009\tbest: 0.3448544 (126)\ttotal: 20s\tremaining: 2m 3s\n",
      "140:\tlearn: 0.0382646\ttest: 0.3522591\tbest: 0.3448544 (126)\ttotal: 20.2s\tremaining: 2m 2s\n",
      "141:\tlearn: 0.0380224\ttest: 0.3531488\tbest: 0.3448544 (126)\ttotal: 20.3s\tremaining: 2m 2s\n",
      "142:\tlearn: 0.0377135\ttest: 0.3563885\tbest: 0.3448544 (126)\ttotal: 20.5s\tremaining: 2m 2s\n",
      "143:\tlearn: 0.0373783\ttest: 0.3561688\tbest: 0.3448544 (126)\ttotal: 20.6s\tremaining: 2m 2s\n",
      "144:\tlearn: 0.0367868\ttest: 0.3547807\tbest: 0.3448544 (126)\ttotal: 20.7s\tremaining: 2m 2s\n",
      "145:\tlearn: 0.0366872\ttest: 0.3553790\tbest: 0.3448544 (126)\ttotal: 20.9s\tremaining: 2m 2s\n",
      "146:\tlearn: 0.0360486\ttest: 0.3549148\tbest: 0.3448544 (126)\ttotal: 21s\tremaining: 2m 1s\n",
      "147:\tlearn: 0.0355952\ttest: 0.3533334\tbest: 0.3448544 (126)\ttotal: 21.2s\tremaining: 2m 1s\n",
      "148:\tlearn: 0.0352339\ttest: 0.3545393\tbest: 0.3448544 (126)\ttotal: 21.3s\tremaining: 2m 1s\n",
      "149:\tlearn: 0.0344265\ttest: 0.3596808\tbest: 0.3448544 (126)\ttotal: 21.5s\tremaining: 2m 1s\n",
      "150:\tlearn: 0.0343406\ttest: 0.3598191\tbest: 0.3448544 (126)\ttotal: 21.6s\tremaining: 2m 1s\n",
      "151:\tlearn: 0.0341099\ttest: 0.3605966\tbest: 0.3448544 (126)\ttotal: 21.7s\tremaining: 2m 1s\n",
      "152:\tlearn: 0.0340523\ttest: 0.3606710\tbest: 0.3448544 (126)\ttotal: 21.9s\tremaining: 2m 1s\n",
      "153:\tlearn: 0.0339750\ttest: 0.3606949\tbest: 0.3448544 (126)\ttotal: 22s\tremaining: 2m\n",
      "154:\tlearn: 0.0335434\ttest: 0.3603426\tbest: 0.3448544 (126)\ttotal: 22.2s\tremaining: 2m\n",
      "155:\tlearn: 0.0331839\ttest: 0.3607169\tbest: 0.3448544 (126)\ttotal: 22.3s\tremaining: 2m\n",
      "156:\tlearn: 0.0326643\ttest: 0.3625024\tbest: 0.3448544 (126)\ttotal: 22.5s\tremaining: 2m\n",
      "157:\tlearn: 0.0324137\ttest: 0.3615694\tbest: 0.3448544 (126)\ttotal: 22.6s\tremaining: 2m\n",
      "158:\tlearn: 0.0319332\ttest: 0.3616328\tbest: 0.3448544 (126)\ttotal: 22.7s\tremaining: 2m\n",
      "159:\tlearn: 0.0317709\ttest: 0.3625759\tbest: 0.3448544 (126)\ttotal: 22.9s\tremaining: 2m\n",
      "160:\tlearn: 0.0315423\ttest: 0.3649858\tbest: 0.3448544 (126)\ttotal: 23s\tremaining: 2m\n",
      "161:\tlearn: 0.0310952\ttest: 0.3665805\tbest: 0.3448544 (126)\ttotal: 23.2s\tremaining: 1m 59s\n",
      "162:\tlearn: 0.0307384\ttest: 0.3686388\tbest: 0.3448544 (126)\ttotal: 23.3s\tremaining: 1m 59s\n",
      "163:\tlearn: 0.0304773\ttest: 0.3672105\tbest: 0.3448544 (126)\ttotal: 23.5s\tremaining: 1m 59s\n",
      "164:\tlearn: 0.0304493\ttest: 0.3673631\tbest: 0.3448544 (126)\ttotal: 23.6s\tremaining: 1m 59s\n",
      "165:\tlearn: 0.0302352\ttest: 0.3692384\tbest: 0.3448544 (126)\ttotal: 23.7s\tremaining: 1m 59s\n",
      "166:\tlearn: 0.0300650\ttest: 0.3692201\tbest: 0.3448544 (126)\ttotal: 23.9s\tremaining: 1m 59s\n",
      "167:\tlearn: 0.0296094\ttest: 0.3734863\tbest: 0.3448544 (126)\ttotal: 24s\tremaining: 1m 58s\n",
      "168:\tlearn: 0.0294202\ttest: 0.3745871\tbest: 0.3448544 (126)\ttotal: 24.2s\tremaining: 1m 58s\n",
      "169:\tlearn: 0.0289706\ttest: 0.3694855\tbest: 0.3448544 (126)\ttotal: 24.3s\tremaining: 1m 58s\n",
      "170:\tlearn: 0.0284406\ttest: 0.3693185\tbest: 0.3448544 (126)\ttotal: 24.5s\tremaining: 1m 58s\n",
      "171:\tlearn: 0.0280195\ttest: 0.3661002\tbest: 0.3448544 (126)\ttotal: 24.6s\tremaining: 1m 58s\n",
      "172:\tlearn: 0.0276198\ttest: 0.3646718\tbest: 0.3448544 (126)\ttotal: 24.7s\tremaining: 1m 58s\n",
      "173:\tlearn: 0.0271507\ttest: 0.3632673\tbest: 0.3448544 (126)\ttotal: 24.9s\tremaining: 1m 58s\n",
      "174:\tlearn: 0.0268972\ttest: 0.3652324\tbest: 0.3448544 (126)\ttotal: 25s\tremaining: 1m 57s\n",
      "175:\tlearn: 0.0266051\ttest: 0.3642983\tbest: 0.3448544 (126)\ttotal: 25.2s\tremaining: 1m 57s\n",
      "176:\tlearn: 0.0261779\ttest: 0.3648232\tbest: 0.3448544 (126)\ttotal: 25.3s\tremaining: 1m 57s\n",
      "177:\tlearn: 0.0256236\ttest: 0.3696824\tbest: 0.3448544 (126)\ttotal: 25.5s\tremaining: 1m 57s\n",
      "178:\tlearn: 0.0255049\ttest: 0.3666626\tbest: 0.3448544 (126)\ttotal: 25.6s\tremaining: 1m 57s\n",
      "179:\tlearn: 0.0254166\ttest: 0.3669044\tbest: 0.3448544 (126)\ttotal: 25.8s\tremaining: 1m 57s\n",
      "180:\tlearn: 0.0251796\ttest: 0.3656022\tbest: 0.3448544 (126)\ttotal: 25.9s\tremaining: 1m 57s\n",
      "181:\tlearn: 0.0247758\ttest: 0.3665511\tbest: 0.3448544 (126)\ttotal: 26s\tremaining: 1m 57s\n",
      "182:\tlearn: 0.0247259\ttest: 0.3669135\tbest: 0.3448544 (126)\ttotal: 26.2s\tremaining: 1m 56s\n",
      "183:\tlearn: 0.0245322\ttest: 0.3663846\tbest: 0.3448544 (126)\ttotal: 26.3s\tremaining: 1m 56s\n",
      "184:\tlearn: 0.0242701\ttest: 0.3702358\tbest: 0.3448544 (126)\ttotal: 26.5s\tremaining: 1m 56s\n",
      "185:\tlearn: 0.0241735\ttest: 0.3714008\tbest: 0.3448544 (126)\ttotal: 26.6s\tremaining: 1m 56s\n",
      "186:\tlearn: 0.0238884\ttest: 0.3688093\tbest: 0.3448544 (126)\ttotal: 26.8s\tremaining: 1m 56s\n",
      "187:\tlearn: 0.0236289\ttest: 0.3679211\tbest: 0.3448544 (126)\ttotal: 26.9s\tremaining: 1m 56s\n",
      "188:\tlearn: 0.0234358\ttest: 0.3696195\tbest: 0.3448544 (126)\ttotal: 27.1s\tremaining: 1m 56s\n",
      "189:\tlearn: 0.0233721\ttest: 0.3686189\tbest: 0.3448544 (126)\ttotal: 27.2s\tremaining: 1m 55s\n",
      "190:\tlearn: 0.0231932\ttest: 0.3697704\tbest: 0.3448544 (126)\ttotal: 27.3s\tremaining: 1m 55s\n",
      "191:\tlearn: 0.0231608\ttest: 0.3690222\tbest: 0.3448544 (126)\ttotal: 27.5s\tremaining: 1m 55s\n",
      "192:\tlearn: 0.0229560\ttest: 0.3700747\tbest: 0.3448544 (126)\ttotal: 27.6s\tremaining: 1m 55s\n",
      "193:\tlearn: 0.0226678\ttest: 0.3733842\tbest: 0.3448544 (126)\ttotal: 27.8s\tremaining: 1m 55s\n",
      "194:\tlearn: 0.0226221\ttest: 0.3732861\tbest: 0.3448544 (126)\ttotal: 27.9s\tremaining: 1m 55s\n",
      "195:\tlearn: 0.0223794\ttest: 0.3704011\tbest: 0.3448544 (126)\ttotal: 28.1s\tremaining: 1m 55s\n",
      "196:\tlearn: 0.0223232\ttest: 0.3700292\tbest: 0.3448544 (126)\ttotal: 28.2s\tremaining: 1m 54s\n",
      "197:\tlearn: 0.0222307\ttest: 0.3723919\tbest: 0.3448544 (126)\ttotal: 28.3s\tremaining: 1m 54s\n",
      "198:\tlearn: 0.0221210\ttest: 0.3744377\tbest: 0.3448544 (126)\ttotal: 28.5s\tremaining: 1m 54s\n",
      "199:\tlearn: 0.0219492\ttest: 0.3749518\tbest: 0.3448544 (126)\ttotal: 28.6s\tremaining: 1m 54s\n",
      "200:\tlearn: 0.0218949\ttest: 0.3742257\tbest: 0.3448544 (126)\ttotal: 28.8s\tremaining: 1m 54s\n",
      "201:\tlearn: 0.0216103\ttest: 0.3756455\tbest: 0.3448544 (126)\ttotal: 28.9s\tremaining: 1m 54s\n",
      "202:\tlearn: 0.0214196\ttest: 0.3771716\tbest: 0.3448544 (126)\ttotal: 29s\tremaining: 1m 54s\n",
      "203:\tlearn: 0.0213283\ttest: 0.3763125\tbest: 0.3448544 (126)\ttotal: 29.2s\tremaining: 1m 53s\n",
      "204:\tlearn: 0.0211683\ttest: 0.3751221\tbest: 0.3448544 (126)\ttotal: 29.3s\tremaining: 1m 53s\n",
      "205:\tlearn: 0.0211430\ttest: 0.3745698\tbest: 0.3448544 (126)\ttotal: 29.5s\tremaining: 1m 53s\n",
      "206:\tlearn: 0.0208778\ttest: 0.3744857\tbest: 0.3448544 (126)\ttotal: 29.6s\tremaining: 1m 53s\n",
      "207:\tlearn: 0.0205427\ttest: 0.3767459\tbest: 0.3448544 (126)\ttotal: 29.8s\tremaining: 1m 53s\n",
      "208:\tlearn: 0.0204766\ttest: 0.3755550\tbest: 0.3448544 (126)\ttotal: 29.9s\tremaining: 1m 53s\n",
      "209:\tlearn: 0.0203363\ttest: 0.3756777\tbest: 0.3448544 (126)\ttotal: 30s\tremaining: 1m 53s\n",
      "210:\tlearn: 0.0200706\ttest: 0.3765168\tbest: 0.3448544 (126)\ttotal: 30.2s\tremaining: 1m 52s\n",
      "211:\tlearn: 0.0199669\ttest: 0.3772875\tbest: 0.3448544 (126)\ttotal: 30.3s\tremaining: 1m 52s\n",
      "212:\tlearn: 0.0198926\ttest: 0.3784816\tbest: 0.3448544 (126)\ttotal: 30.5s\tremaining: 1m 52s\n",
      "213:\tlearn: 0.0198323\ttest: 0.3789297\tbest: 0.3448544 (126)\ttotal: 30.6s\tremaining: 1m 52s\n",
      "214:\tlearn: 0.0196294\ttest: 0.3777333\tbest: 0.3448544 (126)\ttotal: 30.8s\tremaining: 1m 52s\n",
      "215:\tlearn: 0.0194148\ttest: 0.3795734\tbest: 0.3448544 (126)\ttotal: 30.9s\tremaining: 1m 52s\n",
      "216:\tlearn: 0.0192520\ttest: 0.3773848\tbest: 0.3448544 (126)\ttotal: 31s\tremaining: 1m 52s\n",
      "217:\tlearn: 0.0190563\ttest: 0.3780820\tbest: 0.3448544 (126)\ttotal: 31.2s\tremaining: 1m 51s\n",
      "218:\tlearn: 0.0190055\ttest: 0.3781532\tbest: 0.3448544 (126)\ttotal: 31.3s\tremaining: 1m 51s\n",
      "219:\tlearn: 0.0189099\ttest: 0.3764366\tbest: 0.3448544 (126)\ttotal: 31.5s\tremaining: 1m 51s\n",
      "220:\tlearn: 0.0185800\ttest: 0.3750160\tbest: 0.3448544 (126)\ttotal: 31.6s\tremaining: 1m 51s\n",
      "221:\tlearn: 0.0184689\ttest: 0.3745573\tbest: 0.3448544 (126)\ttotal: 31.8s\tremaining: 1m 51s\n",
      "222:\tlearn: 0.0184394\ttest: 0.3745767\tbest: 0.3448544 (126)\ttotal: 31.9s\tremaining: 1m 51s\n",
      "223:\tlearn: 0.0183614\ttest: 0.3738731\tbest: 0.3448544 (126)\ttotal: 32s\tremaining: 1m 51s\n",
      "224:\tlearn: 0.0183011\ttest: 0.3740826\tbest: 0.3448544 (126)\ttotal: 32.2s\tremaining: 1m 50s\n",
      "225:\tlearn: 0.0181639\ttest: 0.3723586\tbest: 0.3448544 (126)\ttotal: 32.3s\tremaining: 1m 50s\n",
      "226:\tlearn: 0.0179600\ttest: 0.3732018\tbest: 0.3448544 (126)\ttotal: 32.5s\tremaining: 1m 50s\n",
      "bestTest = 0.3448544301\n",
      "bestIteration = 126\n",
      "Shrink model to first 127 iterations.\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.87      0.83      0.85        47\n",
      "           1       0.65      0.71      0.68        21\n",
      "\n",
      "    accuracy                           0.79        68\n",
      "   macro avg       0.76      0.77      0.76        68\n",
      "weighted avg       0.80      0.79      0.80        68\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# performance of same embedder where literals are encoded as regular nodes\n",
    "train_embs = embedder.embeddings(\n",
    "    data.edge_index,\n",
    "    data.edge_type,\n",
    "    idx=train_idx,\n",
    ")\n",
    "val_embs = embedder.embeddings(\n",
    "    data.edge_index,\n",
    "    data.edge_type,\n",
    "    idx=val_idx,\n",
    ")\n",
    "test_embs = embedder.embeddings(\n",
    "    data.edge_index,\n",
    "    data.edge_type,\n",
    "    idx=data.test_idx,\n",
    ")\n",
    "task_type = \"GPU\" if torch.cuda.is_available() else \"CPU\"\n",
    "clf = CatBoostClassifier(\n",
    "    iterations=1000,\n",
    "    early_stopping_rounds=100,\n",
    "    task_type=task_type,\n",
    "    random_seed=42,\n",
    "    use_best_model=True,\n",
    "    auto_class_weights=\"Balanced\",\n",
    ")\n",
    "clf = clf.fit(\n",
    "    train_embs.cpu().numpy(),\n",
    "    y_train.cpu().numpy(),\n",
    "    eval_set=(val_embs.cpu().numpy(), y_val.cpu().numpy()),\n",
    ")\n",
    "print(classification_report(clf.predict(test_embs.cpu().numpy()), data.test_y.cpu().numpy()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1/1 [00:12<00:00, 12.54s/it]\n",
      "100%|██████████| 1/1 [00:12<00:00, 12.55s/it]\n",
      "100%|██████████| 1/1 [00:12<00:00, 12.58s/it]\n",
      "Warning: less than 75% gpu memory available for training. Free: 6383.4375 Total: 11178.5\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Learning rate set to 0.080278\n",
      "0:\tlearn: 0.6609449\ttest: 0.6597681\tbest: 0.6597681 (0)\ttotal: 146ms\tremaining: 2m 25s\n",
      "1:\tlearn: 0.6357253\ttest: 0.6465873\tbest: 0.6465873 (1)\ttotal: 282ms\tremaining: 2m 20s\n",
      "2:\tlearn: 0.6027870\ttest: 0.6190348\tbest: 0.6190348 (2)\ttotal: 419ms\tremaining: 2m 19s\n",
      "3:\tlearn: 0.5868718\ttest: 0.6099773\tbest: 0.6099773 (3)\ttotal: 555ms\tremaining: 2m 18s\n",
      "4:\tlearn: 0.5571536\ttest: 0.6033397\tbest: 0.6033397 (4)\ttotal: 699ms\tremaining: 2m 19s\n",
      "5:\tlearn: 0.5379230\ttest: 0.5891080\tbest: 0.5891080 (5)\ttotal: 845ms\tremaining: 2m 20s\n",
      "6:\tlearn: 0.5200257\ttest: 0.5608741\tbest: 0.5608741 (6)\ttotal: 981ms\tremaining: 2m 19s\n",
      "7:\tlearn: 0.5018957\ttest: 0.5454596\tbest: 0.5454596 (7)\ttotal: 1.13s\tremaining: 2m 19s\n",
      "8:\tlearn: 0.4798202\ttest: 0.5418535\tbest: 0.5418535 (8)\ttotal: 1.26s\tremaining: 2m 19s\n",
      "9:\tlearn: 0.4711153\ttest: 0.5369572\tbest: 0.5369572 (9)\ttotal: 1.4s\tremaining: 2m 18s\n",
      "10:\tlearn: 0.4580398\ttest: 0.5212490\tbest: 0.5212490 (10)\ttotal: 1.54s\tremaining: 2m 18s\n",
      "11:\tlearn: 0.4402080\ttest: 0.5211169\tbest: 0.5211169 (11)\ttotal: 1.67s\tremaining: 2m 17s\n",
      "12:\tlearn: 0.4271016\ttest: 0.5049584\tbest: 0.5049584 (12)\ttotal: 1.81s\tremaining: 2m 17s\n",
      "13:\tlearn: 0.4133213\ttest: 0.4928203\tbest: 0.4928203 (13)\ttotal: 1.95s\tremaining: 2m 17s\n",
      "14:\tlearn: 0.4063767\ttest: 0.4873508\tbest: 0.4873508 (14)\ttotal: 2.08s\tremaining: 2m 16s\n",
      "15:\tlearn: 0.3935111\ttest: 0.4843923\tbest: 0.4843923 (15)\ttotal: 2.22s\tremaining: 2m 16s\n",
      "16:\tlearn: 0.3867904\ttest: 0.4772219\tbest: 0.4772219 (16)\ttotal: 2.36s\tremaining: 2m 16s\n",
      "17:\tlearn: 0.3737827\ttest: 0.4745867\tbest: 0.4745867 (17)\ttotal: 2.5s\tremaining: 2m 16s\n",
      "18:\tlearn: 0.3629140\ttest: 0.4712717\tbest: 0.4712717 (18)\ttotal: 2.65s\tremaining: 2m 16s\n",
      "19:\tlearn: 0.3564670\ttest: 0.4719194\tbest: 0.4712717 (18)\ttotal: 2.78s\tremaining: 2m 16s\n",
      "20:\tlearn: 0.3508575\ttest: 0.4608128\tbest: 0.4608128 (20)\ttotal: 2.92s\tremaining: 2m 15s\n",
      "21:\tlearn: 0.3442729\ttest: 0.4618843\tbest: 0.4608128 (20)\ttotal: 3.05s\tremaining: 2m 15s\n",
      "22:\tlearn: 0.3399293\ttest: 0.4558768\tbest: 0.4558768 (22)\ttotal: 3.18s\tremaining: 2m 15s\n",
      "23:\tlearn: 0.3303717\ttest: 0.4663949\tbest: 0.4558768 (22)\ttotal: 3.32s\tremaining: 2m 14s\n",
      "24:\tlearn: 0.3226281\ttest: 0.4582116\tbest: 0.4558768 (22)\ttotal: 3.46s\tremaining: 2m 15s\n",
      "25:\tlearn: 0.3156867\ttest: 0.4549059\tbest: 0.4549059 (25)\ttotal: 3.6s\tremaining: 2m 14s\n",
      "26:\tlearn: 0.3060828\ttest: 0.4513810\tbest: 0.4513810 (26)\ttotal: 3.74s\tremaining: 2m 14s\n",
      "27:\tlearn: 0.2967078\ttest: 0.4423264\tbest: 0.4423264 (27)\ttotal: 3.88s\tremaining: 2m 14s\n",
      "28:\tlearn: 0.2888803\ttest: 0.4423296\tbest: 0.4423264 (27)\ttotal: 4.02s\tremaining: 2m 14s\n",
      "29:\tlearn: 0.2805198\ttest: 0.4464589\tbest: 0.4423264 (27)\ttotal: 4.16s\tremaining: 2m 14s\n",
      "30:\tlearn: 0.2749897\ttest: 0.4455594\tbest: 0.4423264 (27)\ttotal: 4.29s\tremaining: 2m 14s\n",
      "31:\tlearn: 0.2700319\ttest: 0.4434993\tbest: 0.4423264 (27)\ttotal: 4.43s\tremaining: 2m 14s\n",
      "32:\tlearn: 0.2613537\ttest: 0.4455976\tbest: 0.4423264 (27)\ttotal: 4.58s\tremaining: 2m 14s\n",
      "33:\tlearn: 0.2535033\ttest: 0.4394539\tbest: 0.4394539 (33)\ttotal: 4.72s\tremaining: 2m 14s\n",
      "34:\tlearn: 0.2470991\ttest: 0.4369953\tbest: 0.4369953 (34)\ttotal: 4.86s\tremaining: 2m 13s\n",
      "35:\tlearn: 0.2415285\ttest: 0.4286683\tbest: 0.4286683 (35)\ttotal: 5s\tremaining: 2m 13s\n",
      "36:\tlearn: 0.2360315\ttest: 0.4262785\tbest: 0.4262785 (36)\ttotal: 5.14s\tremaining: 2m 13s\n",
      "37:\tlearn: 0.2289108\ttest: 0.4337074\tbest: 0.4262785 (36)\ttotal: 5.29s\tremaining: 2m 13s\n",
      "38:\tlearn: 0.2223986\ttest: 0.4313884\tbest: 0.4262785 (36)\ttotal: 5.42s\tremaining: 2m 13s\n",
      "39:\tlearn: 0.2196059\ttest: 0.4272099\tbest: 0.4262785 (36)\ttotal: 5.56s\tremaining: 2m 13s\n",
      "40:\tlearn: 0.2158858\ttest: 0.4242253\tbest: 0.4242253 (40)\ttotal: 5.7s\tremaining: 2m 13s\n",
      "41:\tlearn: 0.2098088\ttest: 0.4196658\tbest: 0.4196658 (41)\ttotal: 5.84s\tremaining: 2m 13s\n",
      "42:\tlearn: 0.2064021\ttest: 0.4176153\tbest: 0.4176153 (42)\ttotal: 5.98s\tremaining: 2m 13s\n",
      "43:\tlearn: 0.2012524\ttest: 0.4182722\tbest: 0.4176153 (42)\ttotal: 6.12s\tremaining: 2m 12s\n",
      "44:\tlearn: 0.1989411\ttest: 0.4148391\tbest: 0.4148391 (44)\ttotal: 6.26s\tremaining: 2m 12s\n",
      "45:\tlearn: 0.1960886\ttest: 0.4136754\tbest: 0.4136754 (45)\ttotal: 6.39s\tremaining: 2m 12s\n",
      "46:\tlearn: 0.1935123\ttest: 0.4114234\tbest: 0.4114234 (46)\ttotal: 6.53s\tremaining: 2m 12s\n",
      "47:\tlearn: 0.1882394\ttest: 0.4071764\tbest: 0.4071764 (47)\ttotal: 6.67s\tremaining: 2m 12s\n",
      "48:\tlearn: 0.1858382\ttest: 0.4098165\tbest: 0.4071764 (47)\ttotal: 6.81s\tremaining: 2m 12s\n",
      "49:\tlearn: 0.1827270\ttest: 0.4088656\tbest: 0.4071764 (47)\ttotal: 6.95s\tremaining: 2m 12s\n",
      "50:\tlearn: 0.1789270\ttest: 0.4055469\tbest: 0.4055469 (50)\ttotal: 7.09s\tremaining: 2m 12s\n",
      "51:\tlearn: 0.1756243\ttest: 0.4073622\tbest: 0.4055469 (50)\ttotal: 7.23s\tremaining: 2m 11s\n",
      "52:\tlearn: 0.1736462\ttest: 0.4054993\tbest: 0.4054993 (52)\ttotal: 7.37s\tremaining: 2m 11s\n",
      "53:\tlearn: 0.1715007\ttest: 0.4049261\tbest: 0.4049261 (53)\ttotal: 7.5s\tremaining: 2m 11s\n",
      "54:\tlearn: 0.1701367\ttest: 0.4052684\tbest: 0.4049261 (53)\ttotal: 7.64s\tremaining: 2m 11s\n",
      "55:\tlearn: 0.1683188\ttest: 0.4014093\tbest: 0.4014093 (55)\ttotal: 7.78s\tremaining: 2m 11s\n",
      "56:\tlearn: 0.1658131\ttest: 0.4080793\tbest: 0.4014093 (55)\ttotal: 7.92s\tremaining: 2m 11s\n",
      "57:\tlearn: 0.1595091\ttest: 0.4012116\tbest: 0.4012116 (57)\ttotal: 8.07s\tremaining: 2m 11s\n",
      "58:\tlearn: 0.1583976\ttest: 0.4003534\tbest: 0.4003534 (58)\ttotal: 8.21s\tremaining: 2m 10s\n",
      "59:\tlearn: 0.1559042\ttest: 0.4016497\tbest: 0.4003534 (58)\ttotal: 8.35s\tremaining: 2m 10s\n",
      "60:\tlearn: 0.1537176\ttest: 0.3994527\tbest: 0.3994527 (60)\ttotal: 8.49s\tremaining: 2m 10s\n",
      "61:\tlearn: 0.1519104\ttest: 0.3995475\tbest: 0.3994527 (60)\ttotal: 8.63s\tremaining: 2m 10s\n",
      "62:\tlearn: 0.1496215\ttest: 0.3965086\tbest: 0.3965086 (62)\ttotal: 8.77s\tremaining: 2m 10s\n",
      "63:\tlearn: 0.1454742\ttest: 0.3959734\tbest: 0.3959734 (63)\ttotal: 8.92s\tremaining: 2m 10s\n",
      "64:\tlearn: 0.1439244\ttest: 0.3975754\tbest: 0.3959734 (63)\ttotal: 9.06s\tremaining: 2m 10s\n",
      "65:\tlearn: 0.1421353\ttest: 0.3984353\tbest: 0.3959734 (63)\ttotal: 9.2s\tremaining: 2m 10s\n",
      "66:\tlearn: 0.1392330\ttest: 0.3982582\tbest: 0.3959734 (63)\ttotal: 9.34s\tremaining: 2m 10s\n",
      "67:\tlearn: 0.1365833\ttest: 0.4002090\tbest: 0.3959734 (63)\ttotal: 9.48s\tremaining: 2m 9s\n",
      "68:\tlearn: 0.1336547\ttest: 0.3999883\tbest: 0.3959734 (63)\ttotal: 9.62s\tremaining: 2m 9s\n",
      "69:\tlearn: 0.1321430\ttest: 0.4003080\tbest: 0.3959734 (63)\ttotal: 9.76s\tremaining: 2m 9s\n",
      "70:\tlearn: 0.1294556\ttest: 0.4004156\tbest: 0.3959734 (63)\ttotal: 9.91s\tremaining: 2m 9s\n",
      "71:\tlearn: 0.1281537\ttest: 0.3971451\tbest: 0.3959734 (63)\ttotal: 10s\tremaining: 2m 9s\n",
      "72:\tlearn: 0.1278107\ttest: 0.3961264\tbest: 0.3959734 (63)\ttotal: 10.2s\tremaining: 2m 9s\n",
      "73:\tlearn: 0.1247339\ttest: 0.3947753\tbest: 0.3947753 (73)\ttotal: 10.3s\tremaining: 2m 9s\n",
      "74:\tlearn: 0.1224500\ttest: 0.3968383\tbest: 0.3947753 (73)\ttotal: 10.5s\tremaining: 2m 9s\n",
      "75:\tlearn: 0.1207901\ttest: 0.3993956\tbest: 0.3947753 (73)\ttotal: 10.6s\tremaining: 2m 8s\n",
      "76:\tlearn: 0.1198747\ttest: 0.3972085\tbest: 0.3947753 (73)\ttotal: 10.7s\tremaining: 2m 8s\n",
      "77:\tlearn: 0.1180949\ttest: 0.4001745\tbest: 0.3947753 (73)\ttotal: 10.9s\tremaining: 2m 8s\n",
      "78:\tlearn: 0.1149209\ttest: 0.4025906\tbest: 0.3947753 (73)\ttotal: 11s\tremaining: 2m 8s\n",
      "79:\tlearn: 0.1135849\ttest: 0.4030137\tbest: 0.3947753 (73)\ttotal: 11.2s\tremaining: 2m 8s\n",
      "80:\tlearn: 0.1107245\ttest: 0.4019041\tbest: 0.3947753 (73)\ttotal: 11.3s\tremaining: 2m 8s\n",
      "81:\tlearn: 0.1064976\ttest: 0.3997678\tbest: 0.3947753 (73)\ttotal: 11.5s\tremaining: 2m 8s\n",
      "82:\tlearn: 0.1050894\ttest: 0.3997844\tbest: 0.3947753 (73)\ttotal: 11.6s\tremaining: 2m 8s\n",
      "83:\tlearn: 0.1040771\ttest: 0.4001465\tbest: 0.3947753 (73)\ttotal: 11.7s\tremaining: 2m 7s\n",
      "84:\tlearn: 0.1027069\ttest: 0.3991570\tbest: 0.3947753 (73)\ttotal: 11.9s\tremaining: 2m 7s\n",
      "85:\tlearn: 0.1008158\ttest: 0.4009875\tbest: 0.3947753 (73)\ttotal: 12s\tremaining: 2m 7s\n",
      "86:\tlearn: 0.0998588\ttest: 0.3978148\tbest: 0.3947753 (73)\ttotal: 12.2s\tremaining: 2m 7s\n",
      "87:\tlearn: 0.0975498\ttest: 0.3990047\tbest: 0.3947753 (73)\ttotal: 12.3s\tremaining: 2m 7s\n",
      "88:\tlearn: 0.0965340\ttest: 0.3989523\tbest: 0.3947753 (73)\ttotal: 12.4s\tremaining: 2m 7s\n",
      "89:\tlearn: 0.0949141\ttest: 0.3958819\tbest: 0.3947753 (73)\ttotal: 12.6s\tremaining: 2m 7s\n",
      "90:\tlearn: 0.0939399\ttest: 0.3985226\tbest: 0.3947753 (73)\ttotal: 12.7s\tremaining: 2m 7s\n",
      "91:\tlearn: 0.0925952\ttest: 0.3997340\tbest: 0.3947753 (73)\ttotal: 12.9s\tremaining: 2m 7s\n",
      "92:\tlearn: 0.0894785\ttest: 0.4025316\tbest: 0.3947753 (73)\ttotal: 13s\tremaining: 2m 6s\n",
      "93:\tlearn: 0.0886350\ttest: 0.4025443\tbest: 0.3947753 (73)\ttotal: 13.2s\tremaining: 2m 6s\n",
      "94:\tlearn: 0.0880650\ttest: 0.4021414\tbest: 0.3947753 (73)\ttotal: 13.3s\tremaining: 2m 6s\n",
      "95:\tlearn: 0.0871816\ttest: 0.4004217\tbest: 0.3947753 (73)\ttotal: 13.4s\tremaining: 2m 6s\n",
      "96:\tlearn: 0.0859827\ttest: 0.3979005\tbest: 0.3947753 (73)\ttotal: 13.6s\tremaining: 2m 6s\n",
      "97:\tlearn: 0.0841231\ttest: 0.3954839\tbest: 0.3947753 (73)\ttotal: 13.7s\tremaining: 2m 6s\n",
      "98:\tlearn: 0.0833144\ttest: 0.3915131\tbest: 0.3915131 (98)\ttotal: 13.9s\tremaining: 2m 6s\n",
      "99:\tlearn: 0.0816937\ttest: 0.3918314\tbest: 0.3915131 (98)\ttotal: 14s\tremaining: 2m 6s\n",
      "100:\tlearn: 0.0799029\ttest: 0.3858075\tbest: 0.3858075 (100)\ttotal: 14.2s\tremaining: 2m 6s\n",
      "101:\tlearn: 0.0777708\ttest: 0.3850508\tbest: 0.3850508 (101)\ttotal: 14.3s\tremaining: 2m 6s\n",
      "102:\tlearn: 0.0752729\ttest: 0.3896078\tbest: 0.3850508 (101)\ttotal: 14.5s\tremaining: 2m 5s\n",
      "103:\tlearn: 0.0737178\ttest: 0.3852495\tbest: 0.3850508 (101)\ttotal: 14.6s\tremaining: 2m 5s\n",
      "104:\tlearn: 0.0731786\ttest: 0.3867755\tbest: 0.3850508 (101)\ttotal: 14.7s\tremaining: 2m 5s\n",
      "105:\tlearn: 0.0714601\ttest: 0.3843389\tbest: 0.3843389 (105)\ttotal: 14.9s\tremaining: 2m 5s\n",
      "106:\tlearn: 0.0695166\ttest: 0.3861488\tbest: 0.3843389 (105)\ttotal: 15s\tremaining: 2m 5s\n",
      "107:\tlearn: 0.0683655\ttest: 0.3891950\tbest: 0.3843389 (105)\ttotal: 15.2s\tremaining: 2m 5s\n",
      "108:\tlearn: 0.0670973\ttest: 0.3910245\tbest: 0.3843389 (105)\ttotal: 15.3s\tremaining: 2m 5s\n",
      "109:\tlearn: 0.0663421\ttest: 0.3897996\tbest: 0.3843389 (105)\ttotal: 15.5s\tremaining: 2m 5s\n",
      "110:\tlearn: 0.0658352\ttest: 0.3905714\tbest: 0.3843389 (105)\ttotal: 15.6s\tremaining: 2m 4s\n",
      "111:\tlearn: 0.0649574\ttest: 0.3898744\tbest: 0.3843389 (105)\ttotal: 15.8s\tremaining: 2m 4s\n",
      "112:\tlearn: 0.0639151\ttest: 0.3888693\tbest: 0.3843389 (105)\ttotal: 15.9s\tremaining: 2m 4s\n",
      "113:\tlearn: 0.0635841\ttest: 0.3882823\tbest: 0.3843389 (105)\ttotal: 16s\tremaining: 2m 4s\n",
      "114:\tlearn: 0.0626335\ttest: 0.3927130\tbest: 0.3843389 (105)\ttotal: 16.2s\tremaining: 2m 4s\n",
      "115:\tlearn: 0.0613150\ttest: 0.3968487\tbest: 0.3843389 (105)\ttotal: 16.3s\tremaining: 2m 4s\n",
      "116:\tlearn: 0.0600303\ttest: 0.4015576\tbest: 0.3843389 (105)\ttotal: 16.5s\tremaining: 2m 4s\n",
      "117:\tlearn: 0.0590471\ttest: 0.4061342\tbest: 0.3843389 (105)\ttotal: 16.6s\tremaining: 2m 4s\n",
      "118:\tlearn: 0.0576792\ttest: 0.4049975\tbest: 0.3843389 (105)\ttotal: 16.8s\tremaining: 2m 4s\n",
      "119:\tlearn: 0.0566166\ttest: 0.4002456\tbest: 0.3843389 (105)\ttotal: 16.9s\tremaining: 2m 4s\n",
      "120:\tlearn: 0.0557546\ttest: 0.3990316\tbest: 0.3843389 (105)\ttotal: 17.1s\tremaining: 2m 3s\n",
      "121:\tlearn: 0.0547767\ttest: 0.4019894\tbest: 0.3843389 (105)\ttotal: 17.2s\tremaining: 2m 3s\n",
      "122:\tlearn: 0.0534029\ttest: 0.4012113\tbest: 0.3843389 (105)\ttotal: 17.4s\tremaining: 2m 3s\n",
      "123:\tlearn: 0.0519063\ttest: 0.4030012\tbest: 0.3843389 (105)\ttotal: 17.5s\tremaining: 2m 3s\n",
      "124:\tlearn: 0.0513194\ttest: 0.4043040\tbest: 0.3843389 (105)\ttotal: 17.6s\tremaining: 2m 3s\n",
      "125:\tlearn: 0.0504922\ttest: 0.4027534\tbest: 0.3843389 (105)\ttotal: 17.8s\tremaining: 2m 3s\n",
      "126:\tlearn: 0.0496262\ttest: 0.4008210\tbest: 0.3843389 (105)\ttotal: 17.9s\tremaining: 2m 3s\n",
      "127:\tlearn: 0.0489144\ttest: 0.4039476\tbest: 0.3843389 (105)\ttotal: 18.1s\tremaining: 2m 3s\n",
      "128:\tlearn: 0.0478402\ttest: 0.4070064\tbest: 0.3843389 (105)\ttotal: 18.2s\tremaining: 2m 3s\n",
      "129:\tlearn: 0.0470903\ttest: 0.4067584\tbest: 0.3843389 (105)\ttotal: 18.4s\tremaining: 2m 2s\n",
      "130:\tlearn: 0.0464567\ttest: 0.4084214\tbest: 0.3843389 (105)\ttotal: 18.5s\tremaining: 2m 2s\n",
      "131:\tlearn: 0.0459546\ttest: 0.4079765\tbest: 0.3843389 (105)\ttotal: 18.7s\tremaining: 2m 2s\n",
      "132:\tlearn: 0.0454967\ttest: 0.4073639\tbest: 0.3843389 (105)\ttotal: 18.8s\tremaining: 2m 2s\n",
      "133:\tlearn: 0.0453015\ttest: 0.4076331\tbest: 0.3843389 (105)\ttotal: 19s\tremaining: 2m 2s\n",
      "134:\tlearn: 0.0441959\ttest: 0.4099799\tbest: 0.3843389 (105)\ttotal: 19.1s\tremaining: 2m 2s\n",
      "135:\tlearn: 0.0433695\ttest: 0.4054330\tbest: 0.3843389 (105)\ttotal: 19.2s\tremaining: 2m 2s\n",
      "136:\tlearn: 0.0426820\ttest: 0.4056745\tbest: 0.3843389 (105)\ttotal: 19.4s\tremaining: 2m 2s\n",
      "137:\tlearn: 0.0421301\ttest: 0.4047625\tbest: 0.3843389 (105)\ttotal: 19.5s\tremaining: 2m 2s\n",
      "138:\tlearn: 0.0415568\ttest: 0.4023589\tbest: 0.3843389 (105)\ttotal: 19.7s\tremaining: 2m 2s\n",
      "139:\tlearn: 0.0408221\ttest: 0.4019357\tbest: 0.3843389 (105)\ttotal: 19.8s\tremaining: 2m 1s\n",
      "140:\tlearn: 0.0404135\ttest: 0.4024242\tbest: 0.3843389 (105)\ttotal: 20s\tremaining: 2m 1s\n",
      "141:\tlearn: 0.0400588\ttest: 0.4018562\tbest: 0.3843389 (105)\ttotal: 20.1s\tremaining: 2m 1s\n",
      "142:\tlearn: 0.0397866\ttest: 0.4023637\tbest: 0.3843389 (105)\ttotal: 20.3s\tremaining: 2m 1s\n",
      "143:\tlearn: 0.0386183\ttest: 0.4005384\tbest: 0.3843389 (105)\ttotal: 20.4s\tremaining: 2m 1s\n",
      "144:\tlearn: 0.0378003\ttest: 0.4018517\tbest: 0.3843389 (105)\ttotal: 20.6s\tremaining: 2m 1s\n",
      "145:\tlearn: 0.0375216\ttest: 0.4020302\tbest: 0.3843389 (105)\ttotal: 20.7s\tremaining: 2m 1s\n",
      "146:\tlearn: 0.0369140\ttest: 0.4022135\tbest: 0.3843389 (105)\ttotal: 20.9s\tremaining: 2m\n",
      "147:\tlearn: 0.0360925\ttest: 0.3990781\tbest: 0.3843389 (105)\ttotal: 21s\tremaining: 2m\n",
      "148:\tlearn: 0.0354709\ttest: 0.3977188\tbest: 0.3843389 (105)\ttotal: 21.1s\tremaining: 2m\n",
      "149:\tlearn: 0.0348247\ttest: 0.4023609\tbest: 0.3843389 (105)\ttotal: 21.3s\tremaining: 2m\n",
      "150:\tlearn: 0.0342623\ttest: 0.3984784\tbest: 0.3843389 (105)\ttotal: 21.4s\tremaining: 2m\n",
      "151:\tlearn: 0.0337834\ttest: 0.3985696\tbest: 0.3843389 (105)\ttotal: 21.6s\tremaining: 2m\n",
      "152:\tlearn: 0.0331165\ttest: 0.4004205\tbest: 0.3843389 (105)\ttotal: 21.7s\tremaining: 2m\n",
      "153:\tlearn: 0.0323409\ttest: 0.4045566\tbest: 0.3843389 (105)\ttotal: 21.9s\tremaining: 2m\n",
      "154:\tlearn: 0.0318213\ttest: 0.4021748\tbest: 0.3843389 (105)\ttotal: 22s\tremaining: 2m\n",
      "155:\tlearn: 0.0316678\ttest: 0.4020823\tbest: 0.3843389 (105)\ttotal: 22.2s\tremaining: 1m 59s\n",
      "156:\tlearn: 0.0313310\ttest: 0.4035092\tbest: 0.3843389 (105)\ttotal: 22.3s\tremaining: 1m 59s\n",
      "157:\tlearn: 0.0309096\ttest: 0.4058470\tbest: 0.3843389 (105)\ttotal: 22.4s\tremaining: 1m 59s\n",
      "158:\tlearn: 0.0306735\ttest: 0.4067345\tbest: 0.3843389 (105)\ttotal: 22.6s\tremaining: 1m 59s\n",
      "159:\tlearn: 0.0303202\ttest: 0.4083656\tbest: 0.3843389 (105)\ttotal: 22.7s\tremaining: 1m 59s\n",
      "160:\tlearn: 0.0299271\ttest: 0.4095085\tbest: 0.3843389 (105)\ttotal: 22.9s\tremaining: 1m 59s\n",
      "161:\tlearn: 0.0296625\ttest: 0.4107958\tbest: 0.3843389 (105)\ttotal: 23s\tremaining: 1m 59s\n",
      "162:\tlearn: 0.0291154\ttest: 0.4090643\tbest: 0.3843389 (105)\ttotal: 23.2s\tremaining: 1m 58s\n",
      "163:\tlearn: 0.0289097\ttest: 0.4110341\tbest: 0.3843389 (105)\ttotal: 23.3s\tremaining: 1m 58s\n",
      "164:\tlearn: 0.0286343\ttest: 0.4144092\tbest: 0.3843389 (105)\ttotal: 23.4s\tremaining: 1m 58s\n",
      "165:\tlearn: 0.0284987\ttest: 0.4130613\tbest: 0.3843389 (105)\ttotal: 23.6s\tremaining: 1m 58s\n",
      "166:\tlearn: 0.0283774\ttest: 0.4138348\tbest: 0.3843389 (105)\ttotal: 23.7s\tremaining: 1m 58s\n",
      "167:\tlearn: 0.0280271\ttest: 0.4138611\tbest: 0.3843389 (105)\ttotal: 23.9s\tremaining: 1m 58s\n",
      "168:\tlearn: 0.0278946\ttest: 0.4150920\tbest: 0.3843389 (105)\ttotal: 24s\tremaining: 1m 58s\n",
      "169:\tlearn: 0.0276534\ttest: 0.4153124\tbest: 0.3843389 (105)\ttotal: 24.2s\tremaining: 1m 57s\n",
      "170:\tlearn: 0.0273200\ttest: 0.4137479\tbest: 0.3843389 (105)\ttotal: 24.3s\tremaining: 1m 57s\n",
      "171:\tlearn: 0.0271320\ttest: 0.4161173\tbest: 0.3843389 (105)\ttotal: 24.4s\tremaining: 1m 57s\n",
      "172:\tlearn: 0.0267899\ttest: 0.4141186\tbest: 0.3843389 (105)\ttotal: 24.6s\tremaining: 1m 57s\n",
      "173:\tlearn: 0.0264997\ttest: 0.4144558\tbest: 0.3843389 (105)\ttotal: 24.7s\tremaining: 1m 57s\n",
      "174:\tlearn: 0.0258862\ttest: 0.4124561\tbest: 0.3843389 (105)\ttotal: 24.9s\tremaining: 1m 57s\n",
      "175:\tlearn: 0.0254954\ttest: 0.4123339\tbest: 0.3843389 (105)\ttotal: 25s\tremaining: 1m 57s\n",
      "176:\tlearn: 0.0252093\ttest: 0.4118595\tbest: 0.3843389 (105)\ttotal: 25.2s\tremaining: 1m 57s\n",
      "177:\tlearn: 0.0250927\ttest: 0.4118997\tbest: 0.3843389 (105)\ttotal: 25.3s\tremaining: 1m 56s\n",
      "178:\tlearn: 0.0250248\ttest: 0.4110626\tbest: 0.3843389 (105)\ttotal: 25.5s\tremaining: 1m 56s\n",
      "179:\tlearn: 0.0249271\ttest: 0.4110763\tbest: 0.3843389 (105)\ttotal: 25.6s\tremaining: 1m 56s\n",
      "180:\tlearn: 0.0248512\ttest: 0.4105943\tbest: 0.3843389 (105)\ttotal: 25.7s\tremaining: 1m 56s\n",
      "181:\tlearn: 0.0246099\ttest: 0.4117096\tbest: 0.3843389 (105)\ttotal: 25.9s\tremaining: 1m 56s\n",
      "182:\tlearn: 0.0243584\ttest: 0.4098165\tbest: 0.3843389 (105)\ttotal: 26s\tremaining: 1m 56s\n",
      "183:\tlearn: 0.0239702\ttest: 0.4107594\tbest: 0.3843389 (105)\ttotal: 26.2s\tremaining: 1m 56s\n",
      "184:\tlearn: 0.0237405\ttest: 0.4089869\tbest: 0.3843389 (105)\ttotal: 26.3s\tremaining: 1m 55s\n",
      "185:\tlearn: 0.0236303\ttest: 0.4092501\tbest: 0.3843389 (105)\ttotal: 26.5s\tremaining: 1m 55s\n",
      "186:\tlearn: 0.0232427\ttest: 0.4109088\tbest: 0.3843389 (105)\ttotal: 26.6s\tremaining: 1m 55s\n",
      "187:\tlearn: 0.0230771\ttest: 0.4078348\tbest: 0.3843389 (105)\ttotal: 26.8s\tremaining: 1m 55s\n",
      "188:\tlearn: 0.0229278\ttest: 0.4086210\tbest: 0.3843389 (105)\ttotal: 26.9s\tremaining: 1m 55s\n",
      "189:\tlearn: 0.0226941\ttest: 0.4056709\tbest: 0.3843389 (105)\ttotal: 27s\tremaining: 1m 55s\n",
      "190:\tlearn: 0.0226281\ttest: 0.4062257\tbest: 0.3843389 (105)\ttotal: 27.2s\tremaining: 1m 55s\n",
      "191:\tlearn: 0.0225078\ttest: 0.4068192\tbest: 0.3843389 (105)\ttotal: 27.3s\tremaining: 1m 54s\n",
      "192:\tlearn: 0.0223946\ttest: 0.4060624\tbest: 0.3843389 (105)\ttotal: 27.5s\tremaining: 1m 54s\n",
      "193:\tlearn: 0.0219287\ttest: 0.4012742\tbest: 0.3843389 (105)\ttotal: 27.6s\tremaining: 1m 54s\n",
      "194:\tlearn: 0.0218453\ttest: 0.4009433\tbest: 0.3843389 (105)\ttotal: 27.8s\tremaining: 1m 54s\n",
      "195:\tlearn: 0.0216331\ttest: 0.4009329\tbest: 0.3843389 (105)\ttotal: 27.9s\tremaining: 1m 54s\n",
      "196:\tlearn: 0.0215903\ttest: 0.4024761\tbest: 0.3843389 (105)\ttotal: 28.1s\tremaining: 1m 54s\n",
      "197:\tlearn: 0.0215247\ttest: 0.4030228\tbest: 0.3843389 (105)\ttotal: 28.2s\tremaining: 1m 54s\n",
      "198:\tlearn: 0.0213553\ttest: 0.4020844\tbest: 0.3843389 (105)\ttotal: 28.3s\tremaining: 1m 54s\n",
      "199:\tlearn: 0.0212972\ttest: 0.4025592\tbest: 0.3843389 (105)\ttotal: 28.5s\tremaining: 1m 53s\n",
      "200:\tlearn: 0.0212067\ttest: 0.4033112\tbest: 0.3843389 (105)\ttotal: 28.6s\tremaining: 1m 53s\n",
      "201:\tlearn: 0.0208695\ttest: 0.4036664\tbest: 0.3843389 (105)\ttotal: 28.8s\tremaining: 1m 53s\n",
      "202:\tlearn: 0.0206913\ttest: 0.4028833\tbest: 0.3843389 (105)\ttotal: 28.9s\tremaining: 1m 53s\n",
      "203:\tlearn: 0.0206104\ttest: 0.4024938\tbest: 0.3843389 (105)\ttotal: 29.1s\tremaining: 1m 53s\n",
      "204:\tlearn: 0.0204873\ttest: 0.4016031\tbest: 0.3843389 (105)\ttotal: 29.2s\tremaining: 1m 53s\n",
      "205:\tlearn: 0.0203924\ttest: 0.4017501\tbest: 0.3843389 (105)\ttotal: 29.4s\tremaining: 1m 53s\n",
      "bestTest = 0.3843388585\n",
      "bestIteration = 105\n",
      "Shrink model to first 106 iterations.\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.87      0.80      0.83        49\n",
      "           1       0.57      0.68      0.62        19\n",
      "\n",
      "    accuracy                           0.76        68\n",
      "   macro avg       0.72      0.74      0.72        68\n",
      "weighted avg       0.78      0.76      0.77        68\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# performance of same embedder where literals are removed from the graph\n",
    "literals = filtered_node_features[0][0].squeeze().to(device)\n",
    "mask = torch.isin(data.edge_index[0], literals) | torch.isin(data.edge_index[1], literals)\n",
    "edge_index = data.edge_index[:, ~mask]\n",
    "edge_type = data.edge_type[~mask]\n",
    "\n",
    "train_embs = embedder.embeddings(\n",
    "    edge_index,\n",
    "    edge_type,\n",
    "    idx=train_idx,\n",
    ")\n",
    "val_embs = embedder.embeddings(\n",
    "    edge_index,\n",
    "    edge_type,\n",
    "    idx=val_idx,\n",
    ")\n",
    "test_embs = embedder.embeddings(\n",
    "    edge_index,\n",
    "    edge_type,\n",
    "    idx=data.test_idx,\n",
    ")\n",
    "task_type = \"GPU\" if torch.cuda.is_available() else \"CPU\"\n",
    "clf = CatBoostClassifier(\n",
    "    iterations=1000,\n",
    "    early_stopping_rounds=100,\n",
    "    task_type=task_type,\n",
    "    random_seed=42,\n",
    "    use_best_model=True,\n",
    "    auto_class_weights=\"Balanced\",\n",
    ")\n",
    "clf = clf.fit(\n",
    "    train_embs.cpu().numpy(),\n",
    "    y_train.cpu().numpy(),\n",
    "    eval_set=(val_embs.cpu().numpy(), y_val.cpu().numpy()),\n",
    ")\n",
    "print(classification_report(clf.predict(test_embs.cpu().numpy()), data.test_y.cpu().numpy()))"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "d4d1e4263499bec80672ea0156c357c1ee493ec2b1c70f0acce89fc37c4a6abe"
  },
  "kernelspec": {
   "display_name": "Python 3.9.7 ('base')",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.7"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}