tuned-debertav3-lgbm-autocorrect_merge.ipynb

{"cells":[{"attachments":{"ff0ac1de-519e-4239-8a78-3386acc3e551.png":{"image/png":"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"}},"cell_type":"markdown","metadata":{},"source":["## Notes\n","This is a slightly tuned version of @nogawanogawa 's work and I have also converted his messages to english here you can find his notebook here https://www.kaggle.com/code/tsunotsuno/updated-debertav3-lgbm-with-spell-autocorrect please give him kudos for sharing his efforts\n","\n","### Things I would expect there to be a number of things that will allow this model to preform better outside of just strategy and more data. I would imagine there are a few more tuning parameters that could help this model go a long way.\n","\n","\n","\n","\n","\n","In this notebook a combonation of Deberta and LGBM is used, pyspellchecker is also used in order to correct some of the spelling mistakes that are discussed in the discussions tab\n","[Discussion Link](https://www.kaggle.com/competitions/commonlit-evaluate-student-summaries/discussion/428941).\n","[my previous notebook](https://www.kaggle.com/code/tsunotsuno/debertav3-lgbm-with-feature-engineering)\n","[Discussion Link](https://www.kaggle.com/competitions/commonlit-evaluate-student-summaries/discussion/428941).\n","\n","The primary goal of this notebook is to enhance the overall score by honing in on the issue of \"misspellings.\"\n","\n","## Main Concept\n","\n","The Transformers model I'm currently utilizing, Deberta, is pretrained on \"correct sentences.\" However, if I were to train and input it with sentences containing misspellings, Deberta's ability to understand meaning might be compromised.\n","\n","From a human evaluator's perspective, detecting misspellings would prompt deductions in scores. After discreetly rectifying the misspelled words, I'd proceed to evaluate other textual facets. If we assume the scoring process aligns with this approach, it's conceivable that tallying and **correcting** misspellings before feeding text into Deberta could enable the model to aptly capture features beyond just misspellings.\n","\n","In this notebook, I will embark on the journey of auto-correcting misspelled words before inputting them into Deberta. The aim is to evaluate the model's performance by distinctly isolating misspellings from other aspects.\n","\n","### Feature Engineering\n","\n","I intend to largely retain the same features as before:\n","\n","- Text Length\n","- Length Ratio\n","- Word Overlap\n","- N-grams Co-occurrence\n","  - Count\n","  - Ratio\n","- Quotes Overlap\n","- Grammar Check\n","  - Spelling: pyspellchecker\n","\n","### Model Architecture\n","\n","I plan to construct a model with the architecture depicted in the following diagram. For the input to Deberta (`text`), I will pre-process by correcting any misspellings. In other aspects of feature engineering, I will utilize the `text` as is.\n","\n","![image.png](attachment:ff0ac1de-519e-4239-8a78-3386acc3e551.png)\n","\n","### References\n","\n","- https://www.kaggle.com/competitions/commonlit-evaluate-student-summaries/discussion/428941\n","\n","### My previous notebooks\n","\n","- https://www.kaggle.com/code/tsunotsuno/debertav3-baseline-content-and-wording-models\n","- https://www.kaggle.com/code/tsunotsuno/debertav3-w-prompt-title-question-fields\n","- https://www.kaggle.com/code/tsunotsuno/debertav3-with-llama2-example\n","- https://www.kaggle.com/code/tsunotsuno/debertav3-lgbm-with-feature-engineering\n"]},{"cell_type":"code","execution_count":1,"metadata":{"_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","trusted":true},"outputs":[],"source":["# !pip install \"./input/autocorrect-2.6.1.zip\"\n","# !pip install \"./input/pyspellchecker-0.7.2-py3-none-any.whl\""]},{"cell_type":"code","execution_count":2,"metadata":{},"outputs":[],"source":["# nltk.download(\"punkt\")\n"]},{"cell_type":"code","execution_count":3,"metadata":{"trusted":true},"outputs":[{"name":"stderr","output_type":"stream","text":["/home/nghiaph/nghiaph_workspace_115/CommonLit/.venv/lib/python3.8/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n","  from .autonotebook import tqdm as notebook_tqdm\n","2023-10-03 18:02:45.360316: I tensorflow/core/util/port.cc:110] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.\n","2023-10-03 18:02:45.361611: I tensorflow/tsl/cuda/cudart_stub.cc:28] Could not find cuda drivers on your machine, GPU will not be used.\n","2023-10-03 18:02:45.383115: I tensorflow/tsl/cuda/cudart_stub.cc:28] Could not find cuda drivers on your machine, GPU will not be used.\n","2023-10-03 18:02:45.383551: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.\n","To enable the following instructions: AVX2 AVX_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.\n","2023-10-03 18:02:45.722306: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT\n","/home/nghiaph/nghiaph_workspace_115/CommonLit/.venv/lib/python3.8/site-packages/tensorflow_io/python/ops/__init__.py:98: UserWarning: unable to load libtensorflow_io_plugins.so: unable to open file: libtensorflow_io_plugins.so, from paths: ['/home/nghiaph/nghiaph_workspace_115/CommonLit/.venv/lib/python3.8/site-packages/tensorflow_io/python/ops/libtensorflow_io_plugins.so']\n","caused by: ['/home/nghiaph/nghiaph_workspace_115/CommonLit/.venv/lib/python3.8/site-packages/tensorflow_io/python/ops/libtensorflow_io_plugins.so: undefined symbol: _ZN3tsl6StatusC1EN10tensorflow5error4CodeESt17basic_string_viewIcSt11char_traitsIcEENS_14SourceLocationE']\n","  warnings.warn(f\"unable to load libtensorflow_io_plugins.so: {e}\")\n","/home/nghiaph/nghiaph_workspace_115/CommonLit/.venv/lib/python3.8/site-packages/tensorflow_io/python/ops/__init__.py:104: UserWarning: file system plugins are not loaded: unable to open file: libtensorflow_io.so, from paths: ['/home/nghiaph/nghiaph_workspace_115/CommonLit/.venv/lib/python3.8/site-packages/tensorflow_io/python/ops/libtensorflow_io.so']\n","caused by: ['/home/nghiaph/nghiaph_workspace_115/CommonLit/.venv/lib/python3.8/site-packages/tensorflow_io/python/ops/libtensorflow_io.so: undefined symbol: _ZTVN10tensorflow13GcsFileSystemE']\n","  warnings.warn(f\"file system plugins are not loaded: {e}\")\n","2023-10-03 18:02:46.888035: I tensorflow/compiler/xla/stream_executor/cuda/cuda_gpu_executor.cc:996] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355\n","2023-10-03 18:02:46.888803: W tensorflow/core/common_runtime/gpu/gpu_device.cc:1956] Cannot dlopen some GPU libraries. Please make sure the missing libraries mentioned above are installed properly if you would like to use GPU. Follow the guide at https://www.tensorflow.org/install/gpu for how to download and setup the required libraries for your platform.\n","Skipping registering GPU devices...\n"]}],"source":["from typing import List\n","import numpy as np\n","import pandas as pd\n","import warnings\n","import logging\n","import os\n","import shutil\n","import json\n","import transformers\n","from transformers import AutoModel, AutoTokenizer, AutoConfig, AutoModelForSequenceClassification\n","from transformers import DataCollatorWithPadding\n","from datasets import Dataset,load_dataset, load_from_disk\n","from transformers import TrainingArguments, Trainer\n","from datasets import load_metric, disable_progress_bar\n","from sklearn.metrics import mean_squared_error\n","import torch\n","from sklearn.model_selection import KFold, GroupKFold\n","from tqdm import tqdm\n","\n","import nltk\n","from nltk.corpus import stopwords\n","from nltk.tokenize import word_tokenize\n","from nltk.tokenize.treebank import TreebankWordDetokenizer\n","from collections import Counter\n","import spacy\n","import re\n","from autocorrect import Speller\n","from spellchecker import SpellChecker\n","import lightgbm as lgb\n","\n","warnings.simplefilter(\"ignore\")\n","logging.disable(logging.ERROR)\n","os.environ[\"TOKENIZERS_PARALLELISM\"] = \"false\"\n","os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3' \n","disable_progress_bar()\n","tqdm.pandas()"]},{"cell_type":"code","execution_count":null,"metadata":{},"outputs":[],"source":[]},{"cell_type":"code","execution_count":4,"metadata":{"trusted":true},"outputs":[],"source":["def seed_everything(seed: int):\n","    import random, os\n","    import numpy as np\n","    import torch\n","    \n","    random.seed(seed)\n","    os.environ['PYTHONHASHSEED'] = str(seed)\n","    np.random.seed(seed)\n","    torch.manual_seed(seed)\n","    torch.cuda.manual_seed(seed)\n","    torch.backends.cudnn.deterministic = True\n","    torch.backends.cudnn.benchmark = True # load seed\n","    \n","seed_everything(seed=42)"]},{"cell_type":"markdown","metadata":{},"source":["## Class CFG"]},{"cell_type":"code","execution_count":5,"metadata":{"trusted":true},"outputs":[],"source":["class CFG:\n","    model_name=\"debertav3large\"\n","    learning_rate=0.000016\n","    weight_decay=0.03\n","    hidden_dropout_prob=0.000\n","    attention_probs_dropout_prob=0.000\n","    num_train_epochs=5\n","    n_splits=4\n","    batch_size= 2\n","    random_seed=46\n","    save_steps=100\n","    max_length= 512\n","    test_mode = True\n","    device = 'GPU'"]},{"cell_type":"code","execution_count":6,"metadata":{},"outputs":[{"name":"stdout","output_type":"stream","text":["cuda\n"]}],"source":["# print device\n","if CFG.device != 'CPU':\n","    device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\") # print device \n","else :\n","    device = torch.device(\"cpu\")\n","print(device)"]},{"cell_type":"markdown","metadata":{},"source":["## Dataload"]},{"cell_type":"code","execution_count":7,"metadata":{"trusted":true},"outputs":[],"source":["DATA_DIR = \"input/datasets/\"\n","\n","prompts_train = pd.read_csv(DATA_DIR + \"prompts_train.csv\")\n","prompts_test = pd.read_csv(DATA_DIR + \"prompts_test.csv\")\n","summaries_train = pd.read_csv(DATA_DIR + \"summaries_train.csv\")\n","summaries_test = pd.read_csv(DATA_DIR + \"summaries_test.csv\")\n","sample_submission = pd.read_csv(DATA_DIR + \"sample_submission.csv\")"]},{"cell_type":"markdown","metadata":{},"source":["# Exploratory Data Analysis"]},{"cell_type":"code","execution_count":8,"metadata":{},"outputs":[{"data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>prompt_id</th>\n","      <th>prompt_question</th>\n","      <th>prompt_title</th>\n","      <th>prompt_text</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>39c16e</td>\n","      <td>Summarize at least 3 elements of an ideal trag...</td>\n","      <td>On Tragedy</td>\n","      <td>Chapter 13 \\r\\nAs the sequel to what has alrea...</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>3b9047</td>\n","      <td>In complete sentences, summarize the structure...</td>\n","      <td>Egyptian Social Structure</td>\n","      <td>Egyptian society was structured like a pyramid...</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>814d6b</td>\n","      <td>Summarize how the Third Wave developed over su...</td>\n","      <td>The Third Wave</td>\n","      <td>Background \\r\\nThe Third Wave experiment took ...</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>ebad26</td>\n","      <td>Summarize the various ways the factory would u...</td>\n","      <td>Excerpt from The Jungle</td>\n","      <td>With one member trimming beef in a cannery, an...</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["  prompt_id                                    prompt_question  \\\n","0    39c16e  Summarize at least 3 elements of an ideal trag...   \n","1    3b9047  In complete sentences, summarize the structure...   \n","2    814d6b  Summarize how the Third Wave developed over su...   \n","3    ebad26  Summarize the various ways the factory would u...   \n","\n","                prompt_title  \\\n","0                 On Tragedy   \n","1  Egyptian Social Structure   \n","2             The Third Wave   \n","3    Excerpt from The Jungle   \n","\n","                                         prompt_text  \n","0  Chapter 13 \\r\\nAs the sequel to what has alrea...  \n","1  Egyptian society was structured like a pyramid...  \n","2  Background \\r\\nThe Third Wave experiment took ...  \n","3  With one member trimming beef in a cannery, an...  "]},"execution_count":8,"metadata":{},"output_type":"execute_result"}],"source":["\n","prompts_train.head()"]},{"cell_type":"code","execution_count":9,"metadata":{},"outputs":[{"data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>prompt_id</th>\n","      <th>prompt_question</th>\n","      <th>prompt_title</th>\n","      <th>prompt_text</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>39c16e</td>\n","      <td>Summarize at least 3 elements of an ideal trag...</td>\n","      <td>On Tragedy</td>\n","      <td>Chapter 13 \\r\\nAs the sequel to what has alrea...</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>3b9047</td>\n","      <td>In complete sentences, summarize the structure...</td>\n","      <td>Egyptian Social Structure</td>\n","      <td>Egyptian society was structured like a pyramid...</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>814d6b</td>\n","      <td>Summarize how the Third Wave developed over su...</td>\n","      <td>The Third Wave</td>\n","      <td>Background \\r\\nThe Third Wave experiment took ...</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>ebad26</td>\n","      <td>Summarize the various ways the factory would u...</td>\n","      <td>Excerpt from The Jungle</td>\n","      <td>With one member trimming beef in a cannery, an...</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["  prompt_id                                    prompt_question  \\\n","0    39c16e  Summarize at least 3 elements of an ideal trag...   \n","1    3b9047  In complete sentences, summarize the structure...   \n","2    814d6b  Summarize how the Third Wave developed over su...   \n","3    ebad26  Summarize the various ways the factory would u...   \n","\n","                prompt_title  \\\n","0                 On Tragedy   \n","1  Egyptian Social Structure   \n","2             The Third Wave   \n","3    Excerpt from The Jungle   \n","\n","                                         prompt_text  \n","0  Chapter 13 \\r\\nAs the sequel to what has alrea...  \n","1  Egyptian society was structured like a pyramid...  \n","2  Background \\r\\nThe Third Wave experiment took ...  \n","3  With one member trimming beef in a cannery, an...  "]},"execution_count":9,"metadata":{},"output_type":"execute_result"}],"source":["prompts_train.head()"]},{"cell_type":"code","execution_count":null,"metadata":{},"outputs":[],"source":[]},{"cell_type":"markdown","metadata":{},"source":["## Preprocess\n","\n","[Using features]\n","\n","- Text Length\n","- Length Ratio\n","- Word Overlap\n","- N-grams Co-occurrence\n","  - count\n","  - ratio\n","- Quotes Overlap\n","- Grammar Check\n","  - spelling: pyspellchecker\n"]},{"cell_type":"code","execution_count":10,"metadata":{"trusted":true},"outputs":[],"source":["class Preprocessor:\n","    def __init__(self, \n","                model_name: str,\n","                ) -> None:\n","        self.tokenizer = AutoTokenizer.from_pretrained(f\"./input/{model_name}\")\n","        self.twd = TreebankWordDetokenizer()\n","        self.STOP_WORDS = set(stopwords.words('english'))\n","        \n","        self.spacy_ner_model = spacy.load('en_core_web_sm',)\n","        self.speller = Speller(lang='en')\n","        self.spellchecker = SpellChecker() \n","        \n","    def word_overlap_count(self, row):\n","        \"\"\" intersection(prompt_text, text) \"\"\"        \n","        def check_is_stop_word(word):\n","            return word in self.STOP_WORDS\n","        \n","        prompt_words = row['prompt_tokens']\n","        summary_words = row['summary_tokens']\n","        if self.STOP_WORDS:\n","            prompt_words = list(filter(check_is_stop_word, prompt_words))\n","            summary_words = list(filter(check_is_stop_word, summary_words))\n","        return len(set(prompt_words).intersection(set(summary_words)))\n","            \n","    def ngrams(self, token, n):\n","        # Use the zip function to help us generate n-grams\n","        # Concatentate the tokens into ngrams and return\n","        ngrams = zip(*[token[i:] for i in range(n)])\n","        return [\" \".join(ngram) for ngram in ngrams]\n","\n","    def ngram_co_occurrence(self, row, n: int) -> int:\n","        # Tokenize the original text and summary into words\n","        original_tokens = row['prompt_tokens']\n","        summary_tokens = row['summary_tokens']\n","\n","        # Generate n-grams for the original text and summary\n","        original_ngrams = set(self.ngrams(original_tokens, n))\n","        summary_ngrams = set(self.ngrams(summary_tokens, n))\n","\n","        # Calculate the number of common n-grams\n","        common_ngrams = original_ngrams.intersection(summary_ngrams)\n","        return len(common_ngrams)\n","    \n","    def ner_overlap_count(self, row, mode:str):\n","        model = self.spacy_ner_model\n","        def clean_ners(ner_list):\n","            return set([(ner[0].lower(), ner[1]) for ner in ner_list])\n","        prompt = model(row['prompt_text'])\n","        summary = model(row['text'])\n","\n","        if \"spacy\" in str(model):\n","            prompt_ner = set([(token.text, token.label_) for token in prompt.ents])\n","            summary_ner = set([(token.text, token.label_) for token in summary.ents])\n","        elif \"stanza\" in str(model):\n","            prompt_ner = set([(token.text, token.type) for token in prompt.ents])\n","            summary_ner = set([(token.text, token.type) for token in summary.ents])\n","        else:\n","            raise Exception(\"Model not supported\")\n","\n","        prompt_ner = clean_ners(prompt_ner)\n","        summary_ner = clean_ners(summary_ner)\n","\n","        intersecting_ners = prompt_ner.intersection(summary_ner)\n","        \n","        ner_dict = dict(Counter([ner[1] for ner in intersecting_ners]))\n","        \n","        if mode == \"train\":\n","            return ner_dict\n","        elif mode == \"test\":\n","            return {key: ner_dict.get(key) for key in self.ner_keys}\n","\n","    \n","    def quotes_count(self, row):\n","        summary = row['text']\n","        text = row['prompt_text']\n","        quotes_from_summary = re.findall(r'\"([^\"]*)\"', summary)\n","        if len(quotes_from_summary)>0:\n","            return [quote in text for quote in quotes_from_summary].count(True)\n","        else:\n","            return 0\n","\n","    def spelling(self, text):\n","        \n","        wordlist=text.split()\n","        amount_miss = len(list(self.spellchecker.unknown(wordlist)))\n","\n","        return amount_miss\n","    \n","    def add_spelling_dictionary(self, tokens: List[str]) -> List[str]:\n","        \"\"\"dictionary update for pyspell checker and autocorrect\"\"\"\n","        self.spellchecker.word_frequency.load_words(tokens)\n","        self.speller.nlp_data.update({token:1000 for token in tokens})\n","    \n","    def run(self, \n","            prompts: pd.DataFrame,\n","            summaries:pd.DataFrame,\n","            mode:str\n","        ) -> pd.DataFrame:\n","        \n","        # before merge preprocess\n","        prompts[\"prompt_length\"] = prompts[\"prompt_text\"].apply(\n","            lambda x: len(word_tokenize(x))\n","        )\n","        prompts[\"prompt_tokens\"] = prompts[\"prompt_text\"].apply(\n","            lambda x: word_tokenize(x)\n","        )\n","\n","        summaries[\"summary_length\"] = summaries[\"text\"].apply(\n","            lambda x: len(word_tokenize(x))\n","        )\n","        summaries[\"summary_tokens\"] = summaries[\"text\"].apply(\n","            lambda x: word_tokenize(x)\n","        )\n","        \n","        # Add prompt tokens into spelling checker dictionary\n","        prompts[\"prompt_tokens\"].apply(\n","            lambda x: self.add_spelling_dictionary(x)\n","        )\n","        \n","        #         from IPython.core.debugger import Pdb; Pdb().set_trace()\n","        # fix misspelling\n","        summaries[\"fixed_summary_text\"] = summaries[\"text\"].progress_apply(\n","            lambda x: self.speller(x)\n","        ) # fix mission spelling\n","        \n","        # count misspelling\n","        summaries[\"splling_err_num\"] = summaries[\"text\"].progress_apply(self.spelling) # count number of misspelling \n","        \n","        # merge prompts and summaries\n","        input_df = summaries.merge(prompts, how=\"left\", on=\"prompt_id\") \n","\n","        # after merge preprocess\n","        # input_df['length_ratio'] = input_df['summary_length'] / input_df['prompt_length']\n","        \n","        input_df['word_overlap_count'] = input_df.progress_apply(self.word_overlap_count, axis=1)\n","        input_df['bigram_overlap_count'] = input_df.progress_apply(\n","            self.ngram_co_occurrence,args=(2,), axis=1 \n","        )\n","        input_df['bigram_overlap_ratio'] = input_df['bigram_overlap_count'] / (input_df['summary_length'] - 1)\n","        \n","        input_df['trigram_overlap_count'] = input_df.progress_apply(\n","            self.ngram_co_occurrence, args=(3,), axis=1\n","        )\n","        input_df['trigram_overlap_ratio'] = input_df['trigram_overlap_count'] / (input_df['summary_length'] - 2)\n","        \n","        input_df['quotes_count'] = input_df.progress_apply(self.quotes_count, axis=1)\n","        \n","        return input_df.drop(columns=[\"summary_tokens\", \"prompt_tokens\"])\n","    \n","preprocessor = Preprocessor(model_name=CFG.model_name)"]},{"cell_type":"markdown","metadata":{},"source":["## Create the train and test sets\n"]},{"cell_type":"code","execution_count":11,"metadata":{},"outputs":[],"source":["if CFG.test_mode : \n","    prompts_train = prompts_train[:100]\n","    prompts_test = prompts_test[:100]\n","    summaries_train = summaries_train[:100]\n","    summaries_test = summaries_test[:100]"]},{"cell_type":"code","execution_count":12,"metadata":{"trusted":true},"outputs":[{"name":"stderr","output_type":"stream","text":["100%|██████████| 100/100 [00:01<00:00, 50.65it/s]\n","100%|██████████| 100/100 [00:00<00:00, 20504.03it/s]\n","100%|██████████| 100/100 [00:00<00:00, 19527.46it/s]\n","100%|██████████| 100/100 [00:00<00:00, 12560.43it/s]\n","100%|██████████| 100/100 [00:00<00:00, 11009.83it/s]\n","100%|██████████| 100/100 [00:00<00:00, 100824.62it/s]\n","100%|██████████| 4/4 [00:00<00:00, 23865.17it/s]\n","100%|██████████| 4/4 [00:00<00:00, 30229.22it/s]\n","100%|██████████| 4/4 [00:00<00:00, 10356.31it/s]\n","100%|██████████| 4/4 [00:00<00:00, 11781.75it/s]\n","100%|██████████| 4/4 [00:00<00:00, 9436.00it/s]\n","100%|██████████| 4/4 [00:00<00:00, 11155.06it/s]\n"]},{"data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>student_id</th>\n","      <th>prompt_id</th>\n","      <th>text</th>\n","      <th>content</th>\n","      <th>wording</th>\n","      <th>summary_length</th>\n","      <th>fixed_summary_text</th>\n","      <th>splling_err_num</th>\n","      <th>prompt_question</th>\n","      <th>prompt_title</th>\n","      <th>prompt_text</th>\n","      <th>prompt_length</th>\n","      <th>word_overlap_count</th>\n","      <th>bigram_overlap_count</th>\n","      <th>bigram_overlap_ratio</th>\n","      <th>trigram_overlap_count</th>\n","      <th>trigram_overlap_ratio</th>\n","      <th>quotes_count</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>000e8c3c7ddb</td>\n","      <td>814d6b</td>\n","      <td>The third wave was an experimentto see how peo...</td>\n","      <td>0.205683</td>\n","      <td>0.380538</td>\n","      <td>64</td>\n","      <td>The third wave was an experimental see how peo...</td>\n","      <td>5</td>\n","      <td>Summarize how the Third Wave developed over su...</td>\n","      <td>The Third Wave</td>\n","      <td>Background \\r\\nThe Third Wave experiment took ...</td>\n","      <td>660</td>\n","      <td>14</td>\n","      <td>4</td>\n","      <td>0.063492</td>\n","      <td>0</td>\n","      <td>0.000000</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>0020ae56ffbf</td>\n","      <td>ebad26</td>\n","      <td>They would rub it up with soda to make the sme...</td>\n","      <td>-0.548304</td>\n","      <td>0.506755</td>\n","      <td>54</td>\n","      <td>They would rub it up with soda to make the sme...</td>\n","      <td>2</td>\n","      <td>Summarize the various ways the factory would u...</td>\n","      <td>Excerpt from The Jungle</td>\n","      <td>With one member trimming beef in a cannery, an...</td>\n","      <td>1076</td>\n","      <td>18</td>\n","      <td>22</td>\n","      <td>0.415094</td>\n","      <td>10</td>\n","      <td>0.192308</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>004e978e639e</td>\n","      <td>3b9047</td>\n","      <td>In Egypt, there were many occupations and soci...</td>\n","      <td>3.128928</td>\n","      <td>4.231226</td>\n","      <td>269</td>\n","      <td>In Egypt, there were many occupations and soci...</td>\n","      <td>32</td>\n","      <td>In complete sentences, summarize the structure...</td>\n","      <td>Egyptian Social Structure</td>\n","      <td>Egyptian society was structured like a pyramid...</td>\n","      <td>625</td>\n","      <td>22</td>\n","      <td>52</td>\n","      <td>0.194030</td>\n","      <td>23</td>\n","      <td>0.086142</td>\n","      <td>2</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>005ab0199905</td>\n","      <td>3b9047</td>\n","      <td>The highest class was Pharaohs these people we...</td>\n","      <td>-0.210614</td>\n","      <td>-0.471415</td>\n","      <td>28</td>\n","      <td>The highest class was Pharaohs these people we...</td>\n","      <td>5</td>\n","      <td>In complete sentences, summarize the structure...</td>\n","      <td>Egyptian Social Structure</td>\n","      <td>Egyptian society was structured like a pyramid...</td>\n","      <td>625</td>\n","      <td>6</td>\n","      <td>6</td>\n","      <td>0.222222</td>\n","      <td>5</td>\n","      <td>0.192308</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>0070c9e7af47</td>\n","      <td>814d6b</td>\n","      <td>The Third Wave developed  rapidly because the ...</td>\n","      <td>3.272894</td>\n","      <td>3.219757</td>\n","      <td>232</td>\n","      <td>The Third Wave developed  rapidly because the ...</td>\n","      <td>29</td>\n","      <td>Summarize how the Third Wave developed over su...</td>\n","      <td>The Third Wave</td>\n","      <td>Background \\r\\nThe Third Wave experiment took ...</td>\n","      <td>660</td>\n","      <td>23</td>\n","      <td>27</td>\n","      <td>0.116883</td>\n","      <td>5</td>\n","      <td>0.021739</td>\n","      <td>4</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["     student_id prompt_id                                               text  \\\n","0  000e8c3c7ddb    814d6b  The third wave was an experimentto see how peo...   \n","1  0020ae56ffbf    ebad26  They would rub it up with soda to make the sme...   \n","2  004e978e639e    3b9047  In Egypt, there were many occupations and soci...   \n","3  005ab0199905    3b9047  The highest class was Pharaohs these people we...   \n","4  0070c9e7af47    814d6b  The Third Wave developed  rapidly because the ...   \n","\n","    content   wording  summary_length  \\\n","0  0.205683  0.380538              64   \n","1 -0.548304  0.506755              54   \n","2  3.128928  4.231226             269   \n","3 -0.210614 -0.471415              28   \n","4  3.272894  3.219757             232   \n","\n","                                  fixed_summary_text  splling_err_num  \\\n","0  The third wave was an experimental see how peo...                5   \n","1  They would rub it up with soda to make the sme...                2   \n","2  In Egypt, there were many occupations and soci...               32   \n","3  The highest class was Pharaohs these people we...                5   \n","4  The Third Wave developed  rapidly because the ...               29   \n","\n","                                     prompt_question  \\\n","0  Summarize how the Third Wave developed over su...   \n","1  Summarize the various ways the factory would u...   \n","2  In complete sentences, summarize the structure...   \n","3  In complete sentences, summarize the structure...   \n","4  Summarize how the Third Wave developed over su...   \n","\n","                prompt_title  \\\n","0             The Third Wave   \n","1    Excerpt from The Jungle   \n","2  Egyptian Social Structure   \n","3  Egyptian Social Structure   \n","4             The Third Wave   \n","\n","                                         prompt_text  prompt_length  \\\n","0  Background \\r\\nThe Third Wave experiment took ...            660   \n","1  With one member trimming beef in a cannery, an...           1076   \n","2  Egyptian society was structured like a pyramid...            625   \n","3  Egyptian society was structured like a pyramid...            625   \n","4  Background \\r\\nThe Third Wave experiment took ...            660   \n","\n","   word_overlap_count  bigram_overlap_count  bigram_overlap_ratio  \\\n","0                  14                     4              0.063492   \n","1                  18                    22              0.415094   \n","2                  22                    52              0.194030   \n","3                   6                     6              0.222222   \n","4                  23                    27              0.116883   \n","\n","   trigram_overlap_count  trigram_overlap_ratio  quotes_count  \n","0                      0               0.000000             0  \n","1                     10               0.192308             0  \n","2                     23               0.086142             2  \n","3                      5               0.192308             0  \n","4                      5               0.021739             4  "]},"execution_count":12,"metadata":{},"output_type":"execute_result"}],"source":["train = preprocessor.run(prompts_train, summaries_train, mode=\"train\")\n","test = preprocessor.run(prompts_test, summaries_test, mode=\"test\")\n","\n","train.head()"]},{"cell_type":"code","execution_count":13,"metadata":{"trusted":true},"outputs":[{"data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>student_id</th>\n","      <th>prompt_id</th>\n","      <th>text</th>\n","      <th>content</th>\n","      <th>wording</th>\n","      <th>summary_length</th>\n","      <th>fixed_summary_text</th>\n","      <th>splling_err_num</th>\n","      <th>prompt_question</th>\n","      <th>prompt_title</th>\n","      <th>prompt_text</th>\n","      <th>prompt_length</th>\n","      <th>word_overlap_count</th>\n","      <th>bigram_overlap_count</th>\n","      <th>bigram_overlap_ratio</th>\n","      <th>trigram_overlap_count</th>\n","      <th>trigram_overlap_ratio</th>\n","      <th>quotes_count</th>\n","      <th>fold</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>000e8c3c7ddb</td>\n","      <td>814d6b</td>\n","      <td>The third wave was an experimentto see how peo...</td>\n","      <td>0.205683</td>\n","      <td>0.380538</td>\n","      <td>64</td>\n","      <td>The third wave was an experimental see how peo...</td>\n","      <td>5</td>\n","      <td>Summarize how the Third Wave developed over su...</td>\n","      <td>The Third Wave</td>\n","      <td>Background \\r\\nThe Third Wave experiment took ...</td>\n","      <td>660</td>\n","      <td>14</td>\n","      <td>4</td>\n","      <td>0.063492</td>\n","      <td>0</td>\n","      <td>0.000000</td>\n","      <td>0</td>\n","      <td>0.0</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>0020ae56ffbf</td>\n","      <td>ebad26</td>\n","      <td>They would rub it up with soda to make the sme...</td>\n","      <td>-0.548304</td>\n","      <td>0.506755</td>\n","      <td>54</td>\n","      <td>They would rub it up with soda to make the sme...</td>\n","      <td>2</td>\n","      <td>Summarize the various ways the factory would u...</td>\n","      <td>Excerpt from The Jungle</td>\n","      <td>With one member trimming beef in a cannery, an...</td>\n","      <td>1076</td>\n","      <td>18</td>\n","      <td>22</td>\n","      <td>0.415094</td>\n","      <td>10</td>\n","      <td>0.192308</td>\n","      <td>0</td>\n","      <td>1.0</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>004e978e639e</td>\n","      <td>3b9047</td>\n","      <td>In Egypt, there were many occupations and soci...</td>\n","      <td>3.128928</td>\n","      <td>4.231226</td>\n","      <td>269</td>\n","      <td>In Egypt, there were many occupations and soci...</td>\n","      <td>32</td>\n","      <td>In complete sentences, summarize the structure...</td>\n","      <td>Egyptian Social Structure</td>\n","      <td>Egyptian society was structured like a pyramid...</td>\n","      <td>625</td>\n","      <td>22</td>\n","      <td>52</td>\n","      <td>0.194030</td>\n","      <td>23</td>\n","      <td>0.086142</td>\n","      <td>2</td>\n","      <td>1.0</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>005ab0199905</td>\n","      <td>3b9047</td>\n","      <td>The highest class was Pharaohs these people we...</td>\n","      <td>-0.210614</td>\n","      <td>-0.471415</td>\n","      <td>28</td>\n","      <td>The highest class was Pharaohs these people we...</td>\n","      <td>5</td>\n","      <td>In complete sentences, summarize the structure...</td>\n","      <td>Egyptian Social Structure</td>\n","      <td>Egyptian society was structured like a pyramid...</td>\n","      <td>625</td>\n","      <td>6</td>\n","      <td>6</td>\n","      <td>0.222222</td>\n","      <td>5</td>\n","      <td>0.192308</td>\n","      <td>0</td>\n","      <td>1.0</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>0070c9e7af47</td>\n","      <td>814d6b</td>\n","      <td>The Third Wave developed  rapidly because the ...</td>\n","      <td>3.272894</td>\n","      <td>3.219757</td>\n","      <td>232</td>\n","      <td>The Third Wave developed  rapidly because the ...</td>\n","      <td>29</td>\n","      <td>Summarize how the Third Wave developed over su...</td>\n","      <td>The Third Wave</td>\n","      <td>Background \\r\\nThe Third Wave experiment took ...</td>\n","      <td>660</td>\n","      <td>23</td>\n","      <td>27</td>\n","      <td>0.116883</td>\n","      <td>5</td>\n","      <td>0.021739</td>\n","      <td>4</td>\n","      <td>0.0</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["     student_id prompt_id                                               text  \\\n","0  000e8c3c7ddb    814d6b  The third wave was an experimentto see how peo...   \n","1  0020ae56ffbf    ebad26  They would rub it up with soda to make the sme...   \n","2  004e978e639e    3b9047  In Egypt, there were many occupations and soci...   \n","3  005ab0199905    3b9047  The highest class was Pharaohs these people we...   \n","4  0070c9e7af47    814d6b  The Third Wave developed  rapidly because the ...   \n","\n","    content   wording  summary_length  \\\n","0  0.205683  0.380538              64   \n","1 -0.548304  0.506755              54   \n","2  3.128928  4.231226             269   \n","3 -0.210614 -0.471415              28   \n","4  3.272894  3.219757             232   \n","\n","                                  fixed_summary_text  splling_err_num  \\\n","0  The third wave was an experimental see how peo...                5   \n","1  They would rub it up with soda to make the sme...                2   \n","2  In Egypt, there were many occupations and soci...               32   \n","3  The highest class was Pharaohs these people we...                5   \n","4  The Third Wave developed  rapidly because the ...               29   \n","\n","                                     prompt_question  \\\n","0  Summarize how the Third Wave developed over su...   \n","1  Summarize the various ways the factory would u...   \n","2  In complete sentences, summarize the structure...   \n","3  In complete sentences, summarize the structure...   \n","4  Summarize how the Third Wave developed over su...   \n","\n","                prompt_title  \\\n","0             The Third Wave   \n","1    Excerpt from The Jungle   \n","2  Egyptian Social Structure   \n","3  Egyptian Social Structure   \n","4             The Third Wave   \n","\n","                                         prompt_text  prompt_length  \\\n","0  Background \\r\\nThe Third Wave experiment took ...            660   \n","1  With one member trimming beef in a cannery, an...           1076   \n","2  Egyptian society was structured like a pyramid...            625   \n","3  Egyptian society was structured like a pyramid...            625   \n","4  Background \\r\\nThe Third Wave experiment took ...            660   \n","\n","   word_overlap_count  bigram_overlap_count  bigram_overlap_ratio  \\\n","0                  14                     4              0.063492   \n","1                  18                    22              0.415094   \n","2                  22                    52              0.194030   \n","3                   6                     6              0.222222   \n","4                  23                    27              0.116883   \n","\n","   trigram_overlap_count  trigram_overlap_ratio  quotes_count  fold  \n","0                      0               0.000000             0   0.0  \n","1                     10               0.192308             0   1.0  \n","2                     23               0.086142             2   1.0  \n","3                      5               0.192308             0   1.0  \n","4                      5               0.021739             4   0.0  "]},"execution_count":13,"metadata":{},"output_type":"execute_result"}],"source":["gkf = GroupKFold(n_splits=CFG.n_splits)\n","\n","for i, (_, val_index) in enumerate(gkf.split(train, groups=train[\"prompt_id\"])):\n","    train.loc[val_index, \"fold\"] = i\n","\n","train.head()"]},{"cell_type":"markdown","metadata":{},"source":["## Model Function Definition"]},{"cell_type":"code","execution_count":14,"metadata":{"trusted":true},"outputs":[],"source":["def compute_metrics(eval_pred):\n","    predictions, labels = eval_pred\n","    rmse = mean_squared_error(labels, predictions, squared=False)\n","    return {\"rmse\": rmse}\n","\n","def compute_mcrmse(eval_pred):\n","    \"\"\"\n","    Calculates mean columnwise root mean squared error\n","    https://www.kaggle.com/competitions/commonlit-evaluate-student-summaries/overview/evaluation\n","    \"\"\"\n","    preds, labels = eval_pred\n","\n","    col_rmse = np.sqrt(np.mean((preds - labels) ** 2, axis=0))\n","    mcrmse = np.mean(col_rmse)\n","\n","    return {\n","        \"content_rmse\": col_rmse[0],\n","        \"wording_rmse\": col_rmse[1],\n","        \"mcrmse\": mcrmse,\n","    }\n","\n","def compt_score(content_true, content_pred, wording_true, wording_pred):\n","    content_score = mean_squared_error(content_true, content_pred)**(1/2)\n","    wording_score = mean_squared_error(wording_true, wording_pred)**(1/2)\n","    \n","    return (content_score + wording_score)/2"]},{"cell_type":"markdown","metadata":{},"source":["## Deberta Regressor"]},{"cell_type":"code","execution_count":15,"metadata":{"trusted":true},"outputs":[],"source":["class ContentScoreRegressor:\n","    def __init__(self, \n","                model_name: str,\n","                model_dir: str,\n","                target: str,\n","                hidden_dropout_prob: float,\n","                attention_probs_dropout_prob: float,\n","                max_length: int,\n","                ):\n","        self.inputs = [\"prompt_text\", \"prompt_title\", \"prompt_question\", \"fixed_summary_text\"] # fix summary text have prompt text in it \n","        self.input_col = \"input\"\n","        \n","        self.text_cols = [self.input_col] \n","        self.target = target\n","        self.target_cols = [target]\n","\n","        self.model_name = model_name\n","        self.model_dir = model_dir\n","        self.max_length = max_length\n","        \n","        self.tokenizer = AutoTokenizer.from_pretrained(f\"./input/{model_name}\")\n","        self.model_config = AutoConfig.from_pretrained(f\"./input/{model_name}\")\n","        \n","        self.model_config.update({\n","            \"hidden_dropout_prob\": hidden_dropout_prob,\n","            \"attention_probs_dropout_prob\": attention_probs_dropout_prob,\n","            \"num_labels\": 1,\n","            \"problem_type\": \"regression\",\n","        })\n","        \n","        seed_everything(seed=42)\n","\n","        self.data_collator = DataCollatorWithPadding(\n","            tokenizer=self.tokenizer\n","        )\n","\n","\n","    def tokenize_function(self, examples: pd.DataFrame):\n","        labels = [examples[self.target]]\n","        tokenized = self.tokenizer(examples[self.input_col],\n","                         padding=False,\n","                         truncation=True,\n","                         max_length=self.max_length)\n","        return {\n","            **tokenized,\n","            \"labels\": labels,\n","        }\n","    \n","    def tokenize_function_test(self, examples: pd.DataFrame):\n","        tokenized = self.tokenizer(examples[self.input_col],\n","                         padding=False,\n","                         truncation=True,\n","                         max_length=self.max_length)\n","        return tokenized\n","        \n","    def train(self, \n","            fold: int,\n","            train_df: pd.DataFrame,\n","            valid_df: pd.DataFrame,\n","            batch_size: int,\n","            learning_rate: float,\n","            weight_decay: float,\n","            num_train_epochs: float,\n","            save_steps: int,\n","        ) -> None:\n","        \"\"\"fine-tuning\"\"\"\n","        \n","        sep = self.tokenizer.sep_token\n","        train_df[self.input_col] = (\n","                    train_df[\"prompt_title\"] + sep \n","                    + train_df[\"prompt_question\"] + sep \n","                    + train_df[\"fixed_summary_text\"]\n","                  )\n","\n","        valid_df[self.input_col] = (\n","                    valid_df[\"prompt_title\"] + sep \n","                    + valid_df[\"prompt_question\"] + sep \n","                    + valid_df[\"fixed_summary_text\"]\n","                  )\n","        print(train_df.shape, valid_df.shape)\n","        # filter train_df with input_col have more than 5000 tokens\n","\n","        train_df = train_df[[self.input_col] + self.target_cols]\n","        valid_df = valid_df[[self.input_col] + self.target_cols]\n","        model_content = AutoModelForSequenceClassification.from_pretrained(\n","            f\"./input/{self.model_name}\", \n","            config=self.model_config\n","        )\n","\n","        train_dataset = Dataset.from_pandas(train_df, preserve_index=False) \n","        val_dataset = Dataset.from_pandas(valid_df, preserve_index=False) \n","        \n","        train_tokenized_datasets = train_dataset.map(self.tokenize_function, batched=False)\n","        val_tokenized_datasets = val_dataset.map(self.tokenize_function, batched=False)\n","\n","        # eg. \"bert/fold_0/\"\n","        model_fold_dir = os.path.join(self.model_dir, str(fold)) \n","        training_args = TrainingArguments(\n","            output_dir=model_fold_dir,\n","            load_best_model_at_end=True, # select best model\n","            learning_rate=learning_rate,\n","            per_device_train_batch_size=batch_size,\n","            per_device_eval_batch_size=8,\n","            num_train_epochs=num_train_epochs,\n","            weight_decay=weight_decay,\n","            report_to='none',\n","            greater_is_better=False,\n","            save_strategy=\"steps\",\n","            evaluation_strategy=\"steps\",\n","            eval_steps=save_steps,\n","            save_steps=save_steps,\n","            metric_for_best_model=\"rmse\",\n","            save_total_limit=1\n","        )\n","\n","        trainer = Trainer(\n","            model=model_content,\n","            args=training_args,\n","            train_dataset=train_tokenized_datasets,\n","            eval_dataset=val_tokenized_datasets,\n","            tokenizer=self.tokenizer,\n","            compute_metrics=compute_metrics,\n","            data_collator=self.data_collator\n","        )\n","\n","        trainer.train()\n","        \n","        model_content.save_pretrained(self.model_dir)\n","        self.tokenizer.save_pretrained(self.model_dir)\n","\n","        \n","    def predict(self, \n","                test_df: pd.DataFrame,\n","                fold: int,\n","               ):\n","        \"\"\"predict content score\"\"\"\n","        \n","        sep = self.tokenizer.sep_token\n","        in_text = (\n","                    test_df[\"prompt_title\"] + sep \n","                    + test_df[\"prompt_question\"] + sep \n","                    + test_df[\"fixed_summary_text\"]\n","                  )\n","        test_df[self.input_col] = in_text\n","\n","        test_ = test_df[[self.input_col]]\n","    \n","        test_dataset = Dataset.from_pandas(test_, preserve_index=False) \n","        test_tokenized_dataset = test_dataset.map(self.tokenize_function_test, batched=False)\n","\n","        model_content = AutoModelForSequenceClassification.from_pretrained(f\"{self.model_dir}\")\n","        model_content.eval()\n","        \n","        # eg. \"bert/fold_0/\"\n","        model_fold_dir = os.path.join(self.model_dir, str(fold)) \n","\n","        test_args = TrainingArguments(\n","            output_dir=model_fold_dir,\n","            do_train = False,\n","            do_predict = True,\n","            per_device_eval_batch_size = 4,   \n","            dataloader_drop_last = False,\n","        )\n","\n","        # init trainer\n","        infer_content = Trainer(\n","                      model = model_content, \n","                      tokenizer=self.tokenizer,\n","                      data_collator=self.data_collator,\n","                      args = test_args)\n","\n","        preds = infer_content.predict(test_tokenized_dataset)[0]\n","\n","        return preds"]},{"cell_type":"markdown","metadata":{},"source":["## Train by fold function\n"]},{"cell_type":"code","execution_count":16,"metadata":{"trusted":true},"outputs":[],"source":["def train_by_fold(\n","        train_df: pd.DataFrame,\n","        model_name: str,\n","        target:str,\n","        save_each_model: bool,\n","        n_splits: int,\n","        batch_size: int,\n","        learning_rate: int,\n","        hidden_dropout_prob: float,\n","        attention_probs_dropout_prob: float,\n","        weight_decay: float,\n","        num_train_epochs: int,\n","        save_steps: int,\n","        max_length:int\n","    ):\n","\n","    # delete old model files\n","    if os.path.exists(model_name):\n","        shutil.rmtree(model_name)\n","    \n","    os.mkdir(model_name)\n","        \n","    for fold in range(CFG.n_splits):\n","        print(f\"fold {fold}:\")\n","        \n","        train_data = train_df[train_df[\"fold\"] != fold]\n","        valid_data = train_df[train_df[\"fold\"] == fold]\n","        \n","        if save_each_model == True:\n","            model_dir =  f\"{target}/{model_name}/fold_{fold}\"\n","        else: \n","            model_dir =  f\"{model_name}/fold_{fold}\"\n","\n","        csr = ContentScoreRegressor(\n","            model_name=model_name,\n","            target=target,\n","            model_dir = model_dir, \n","            hidden_dropout_prob=hidden_dropout_prob,\n","            attention_probs_dropout_prob=attention_probs_dropout_prob,\n","            max_length=max_length,\n","           )\n","        \n","        csr.train(\n","            fold=fold,\n","            train_df=train_data,\n","            valid_df=valid_data, \n","            batch_size=batch_size,\n","            learning_rate=learning_rate,\n","            weight_decay=weight_decay,\n","            num_train_epochs=num_train_epochs,\n","            save_steps=save_steps,\n","        )\n","\n","def validate(\n","    train_df: pd.DataFrame,\n","    target:str,\n","    save_each_model: bool,\n","    model_name: str,\n","    hidden_dropout_prob: float,\n","    attention_probs_dropout_prob: float,\n","    max_length : int\n","    ) -> pd.DataFrame:\n","    \"\"\"predict oof data\"\"\"\n","    for fold in range(CFG.n_splits):\n","        print(f\"fold {fold}:\")\n","        \n","        valid_data = train_df[train_df[\"fold\"] == fold]\n","        \n","        if save_each_model == True:\n","            model_dir =  f\"{target}/{model_name}/fold_{fold}\"\n","        else: \n","            model_dir =  f\"{model_name}/fold_{fold}\"\n","        csr = ContentScoreRegressor(\n","            model_name=model_name,\n","            target=target,\n","            model_dir = model_dir,\n","            hidden_dropout_prob=hidden_dropout_prob,\n","            attention_probs_dropout_prob=attention_probs_dropout_prob,\n","            max_length=max_length,\n","           )\n","        \n","        pred = csr.predict(\n","            test_df=valid_data, \n","            fold=fold\n","        )\n","        \n","        train_df.loc[valid_data.index, f\"{target}_pred\"] = pred\n","\n","    return train_df\n","    \n","def predict(\n","    test_df: pd.DataFrame,\n","    target:str,\n","    save_each_model: bool,\n","    model_name: str,\n","    hidden_dropout_prob: float,\n","    attention_probs_dropout_prob: float,\n","    max_length : int\n","    ):\n","    \"\"\"predict using mean folds\"\"\"\n","\n","    for fold in range(CFG.n_splits):\n","        print(f\"fold {fold}:\")\n","        \n","        if save_each_model == True:\n","            model_dir =  f\"{target}/{model_name}/fold_{fold}\"\n","        else: \n","            model_dir =  f\"{model_name}/fold_{fold}\"\n","        csr = ContentScoreRegressor(\n","            model_name=model_name,\n","            target=target,\n","            model_dir = model_dir, \n","            hidden_dropout_prob=hidden_dropout_prob,\n","            attention_probs_dropout_prob=attention_probs_dropout_prob,\n","            max_length=max_length,\n","           )\n","        \n","        pred = csr.predict(\n","            test_df=test_df, \n","            fold=fold\n","        )\n","        \n","        test_df[f\"{target}_pred_{fold}\"] = pred\n","    \n","    test_df[f\"{target}\"] = test_df[[f\"{target}_pred_{fold}\" for fold in range(CFG.n_splits)]].mean(axis=1)\n","\n","    return test_df"]},{"cell_type":"markdown","metadata":{},"source":["## Train the model"]},{"cell_type":"code","execution_count":17,"metadata":{"trusted":true},"outputs":[{"name":"stdout","output_type":"stream","text":["fold 0:\n","(56, 20) (44, 20)\n"]},{"data":{"text/html":["\n","    <div>\n","      \n","      <progress value='20' max='20' style='width:300px; height:20px; vertical-align: middle;'></progress>\n","      [20/20 00:04, Epoch 5/5]\n","    </div>\n","    <table border=\"1\" class=\"dataframe\">\n","  <thead>\n"," <tr style=\"text-align: left;\">\n","      <th>Step</th>\n","      <th>Training Loss</th>\n","      <th>Validation Loss</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","  </tbody>\n","</table><p>"],"text/plain":["<IPython.core.display.HTML object>"]},"metadata":{},"output_type":"display_data"},{"name":"stdout","output_type":"stream","text":["fold 1:\n","(44, 20) (56, 20)\n"]},{"data":{"text/html":["\n","    <div>\n","      \n","      <progress value='15' max='15' style='width:300px; height:20px; vertical-align: middle;'></progress>\n","      [15/15 00:02, Epoch 5/5]\n","    </div>\n","    <table border=\"1\" class=\"dataframe\">\n","  <thead>\n"," <tr style=\"text-align: left;\">\n","      <th>Step</th>\n","      <th>Training Loss</th>\n","      <th>Validation Loss</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","  </tbody>\n","</table><p>"],"text/plain":["<IPython.core.display.HTML object>"]},"metadata":{},"output_type":"display_data"},{"name":"stdout","output_type":"stream","text":["fold 0:\n"]},{"data":{"text/html":[],"text/plain":["<IPython.core.display.HTML object>"]},"metadata":{},"output_type":"display_data"},{"name":"stdout","output_type":"stream","text":["fold 1:\n"]},{"data":{"text/html":[],"text/plain":["<IPython.core.display.HTML object>"]},"metadata":{},"output_type":"display_data"},{"name":"stdout","output_type":"stream","text":["cv content rmse: 0.9942135378303695\n","fold 0:\n"]},{"data":{"text/html":[],"text/plain":["<IPython.core.display.HTML object>"]},"metadata":{},"output_type":"display_data"},{"name":"stdout","output_type":"stream","text":["fold 1:\n"]},{"data":{"text/html":[],"text/plain":["<IPython.core.display.HTML object>"]},"metadata":{},"output_type":"display_data"},{"name":"stdout","output_type":"stream","text":["fold 0:\n","(56, 21) (44, 21)\n"]},{"data":{"text/html":["\n","    <div>\n","      \n","      <progress value='20' max='20' style='width:300px; height:20px; vertical-align: middle;'></progress>\n","      [20/20 00:04, Epoch 5/5]\n","    </div>\n","    <table border=\"1\" class=\"dataframe\">\n","  <thead>\n"," <tr style=\"text-align: left;\">\n","      <th>Step</th>\n","      <th>Training Loss</th>\n","      <th>Validation Loss</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","  </tbody>\n","</table><p>"],"text/plain":["<IPython.core.display.HTML object>"]},"metadata":{},"output_type":"display_data"},{"name":"stdout","output_type":"stream","text":["fold 1:\n","(44, 21) (56, 21)\n"]},{"data":{"text/html":["\n","    <div>\n","      \n","      <progress value='15' max='15' style='width:300px; height:20px; vertical-align: middle;'></progress>\n","      [15/15 00:02, Epoch 5/5]\n","    </div>\n","    <table border=\"1\" class=\"dataframe\">\n","  <thead>\n"," <tr style=\"text-align: left;\">\n","      <th>Step</th>\n","      <th>Training Loss</th>\n","      <th>Validation Loss</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","  </tbody>\n","</table><p>"],"text/plain":["<IPython.core.display.HTML object>"]},"metadata":{},"output_type":"display_data"},{"name":"stdout","output_type":"stream","text":["fold 0:\n"]},{"data":{"text/html":[],"text/plain":["<IPython.core.display.HTML object>"]},"metadata":{},"output_type":"display_data"},{"name":"stdout","output_type":"stream","text":["fold 1:\n"]},{"data":{"text/html":[],"text/plain":["<IPython.core.display.HTML object>"]},"metadata":{},"output_type":"display_data"},{"name":"stdout","output_type":"stream","text":["cv wording rmse: 1.1213127881154004\n","fold 0:\n"]},{"data":{"text/html":[],"text/plain":["<IPython.core.display.HTML object>"]},"metadata":{},"output_type":"display_data"},{"name":"stdout","output_type":"stream","text":["fold 1:\n"]},{"data":{"text/html":[],"text/plain":["<IPython.core.display.HTML object>"]},"metadata":{},"output_type":"display_data"}],"source":["save_each_model = True\n","for target in [\"content\", \"wording\"]:\n","    train_by_fold(\n","        train, # this is train dataset\n","        model_name=CFG.model_name,\n","        save_each_model=save_each_model,\n","        target=target,\n","        learning_rate=CFG.learning_rate,\n","        hidden_dropout_prob=CFG.hidden_dropout_prob,\n","        attention_probs_dropout_prob=CFG.attention_probs_dropout_prob,\n","        weight_decay=CFG.weight_decay,\n","        num_train_epochs=CFG.num_train_epochs,\n","        n_splits=CFG.n_splits,\n","        batch_size=CFG.batch_size,\n","        save_steps=CFG.save_steps,\n","        max_length=CFG.max_length\n","    )\n","    \n","    \n","    train = validate(\n","        train,\n","        target=target,\n","        save_each_model=save_each_model,\n","        model_name=CFG.model_name,\n","        hidden_dropout_prob=CFG.hidden_dropout_prob,\n","        attention_probs_dropout_prob=CFG.attention_probs_dropout_prob,\n","        max_length=CFG.max_length\n","    )\n","\n","    rmse = mean_squared_error(train[target], train[f\"{target}_pred\"], squared=False)\n","    print(f\"cv {target} rmse: {rmse}\")\n","\n","    test = predict(\n","        test,\n","        target=target,\n","        save_each_model=save_each_model,\n","        model_name=CFG.model_name,\n","        hidden_dropout_prob=CFG.hidden_dropout_prob,\n","        attention_probs_dropout_prob=CFG.attention_probs_dropout_prob,\n","        max_length=CFG.max_length\n","    )"]},{"cell_type":"markdown","metadata":{},"source":["## LGBM model"]},{"cell_type":"code","execution_count":18,"metadata":{"trusted":true},"outputs":[],"source":["targets = [\"content\", \"wording\"]\n","\n","drop_columns = [\"fold\", \"student_id\", \"prompt_id\", \"text\", \"fixed_summary_text\",\n","                \"prompt_question\", \"prompt_title\", \n","                \"prompt_text\"\n","               ] + targets"]},{"cell_type":"code","execution_count":19,"metadata":{"trusted":true},"outputs":[{"name":"stdout","output_type":"stream","text":["[LightGBM] [Info] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000045 seconds.\n","You can set `force_row_wise=true` to remove the overhead.\n","And if memory is not enough, you can set `force_col_wise=true`.\n","[LightGBM] [Info] Total Bins 154\n","[LightGBM] [Info] Number of data points in the train set: 56, number of used features: 10\n","[LightGBM] [Info] Start training from score -0.199268\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","Training until validation scores don't improve for 30 rounds\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[100]\ttrain's rmse: 0.777087\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[200]\ttrain's rmse: 0.732947\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[300]\ttrain's rmse: 0.728191\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","Early stopping, best iteration is:\n","[358]\ttrain's rmse: 0.725847\n","[LightGBM] [Info] Auto-choosing col-wise multi-threading, the overhead of testing was 0.000026 seconds.\n","You can set `force_col_wise=true` to remove the overhead.\n","[LightGBM] [Info] Total Bins 126\n","[LightGBM] [Info] Number of data points in the train set: 44, number of used features: 9\n","[LightGBM] [Info] Start training from score -0.115339\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","Training until validation scores don't improve for 30 rounds\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[100]\ttrain's rmse: 0.96308\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[200]\ttrain's rmse: 0.948092\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[300]\ttrain's rmse: 0.938319\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[400]\ttrain's rmse: 0.920593\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[500]\ttrain's rmse: 0.905319\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[600]\ttrain's rmse: 0.892141\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[700]\ttrain's rmse: 0.882282\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[800]\ttrain's rmse: 0.873737\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[900]\ttrain's rmse: 0.867206\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[1000]\ttrain's rmse: 0.863666\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[1100]\ttrain's rmse: 0.861414\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[1200]\ttrain's rmse: 0.860224\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[1300]\ttrain's rmse: 0.859333\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[1400]\ttrain's rmse: 0.858269\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","Early stopping, best iteration is:\n","[1469]\ttrain's rmse: 0.857558\n","[LightGBM] [Info] Auto-choosing col-wise multi-threading, the overhead of testing was 0.000027 seconds.\n","You can set `force_col_wise=true` to remove the overhead.\n","[LightGBM] [Info] Total Bins 154\n","[LightGBM] [Info] Number of data points in the train set: 56, number of used features: 10\n","[LightGBM] [Info] Start training from score -0.243706\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","Training until validation scores don't improve for 30 rounds\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[100]\ttrain's rmse: 1.01834\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","Early stopping, best iteration is:\n","[111]\ttrain's rmse: 1.01663\n","[LightGBM] [Info] Auto-choosing col-wise multi-threading, the overhead of testing was 0.000025 seconds.\n","You can set `force_col_wise=true` to remove the overhead.\n","[LightGBM] [Info] Total Bins 126\n","[LightGBM] [Info] Number of data points in the train set: 44, number of used features: 9\n","[LightGBM] [Info] Start training from score -0.146694\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","Training until validation scores don't improve for 30 rounds\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[100]\ttrain's rmse: 1.04248\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n","Early stopping, best iteration is:\n","[140]\ttrain's rmse: 1.03655\n"]}],"source":["model_dict = {}\n","\n","for target in targets:\n","    models = []\n","    \n","    for fold in range(CFG.n_splits):\n","        # if target == targets[1]:\n","        #   drop_columns.remove(targets[0])\n","        X_train_cv = train[train[\"fold\"] != fold].drop(columns=drop_columns)\n","        y_train_cv = train[train[\"fold\"] != fold][target]\n","\n","        X_eval_cv = train[train[\"fold\"] == fold].drop(columns=drop_columns)\n","        y_eval_cv = train[train[\"fold\"] == fold][target]\n","\n","        dtrain = lgb.Dataset(X_train_cv, label=y_train_cv)\n","        dval = lgb.Dataset(X_eval_cv, label=y_eval_cv)\n","          \n","        params = {\n","            'boosting_type': 'gbdt',\n","            'random_state': 42,\n","            'objective': 'regression',\n","            'metric': 'rmse',\n","            'learning_rate': 0.048,\n","            'max_depth': 3,\n","            'lambda_l1': 0.0,\n","            'lambda_l2': 0.011\n","        }\n","\n","        evaluation_results = {}\n","        model = lgb.train(params,\n","                          num_boost_round=10000,\n","                            #categorical_feature = categorical_features,\n","                          valid_names=['train', 'valid'],\n","                          train_set=dtrain,\n","                          valid_sets=dval,\n","                          callbacks=[\n","                              lgb.early_stopping(stopping_rounds=30, verbose=True),\n","                               lgb.log_evaluation(100),\n","                              lgb.callback.record_evaluation(evaluation_results)\n","                            ],\n","                          )\n","        models.append(model)\n","    \n","    model_dict[target] = models"]},{"cell_type":"code","execution_count":20,"metadata":{},"outputs":[],"source":["# save models\n","import pickle\n","with open(f\"{CFG.model_name}_model_dict.pickle\", 'wb') as f:\n","    pickle.dump(model_dict, f)\n","    "]},{"cell_type":"markdown","metadata":{},"source":["## CV Score"]},{"cell_type":"code","execution_count":21,"metadata":{"trusted":true},"outputs":[{"name":"stdout","output_type":"stream","text":["content_rmse : 0.8022735635840902\n","wording_rmse : 1.0278321561411683\n","mcrmse : 0.9150528598626293\n","__module__ __main__\n","model_name debertav3base\n","learning_rate 1.6e-05\n","weight_decay 0.03\n","hidden_dropout_prob 0.0\n","attention_probs_dropout_prob 0.0\n","num_train_epochs 5\n","n_splits 2\n","batch_size 18\n","random_seed 46\n","save_steps 100\n","max_length 512\n","test_mode True\n","device GPU\n","__dict__ {'__module__': '__main__', 'model_name': 'debertav3base', 'learning_rate': 1.6e-05, 'weight_decay': 0.03, 'hidden_dropout_prob': 0.0, 'attention_probs_dropout_prob': 0.0, 'num_train_epochs': 5, 'n_splits': 2, 'batch_size': 18, 'random_seed': 46, 'save_steps': 100, 'max_length': 512, 'test_mode': True, 'device': 'GPU', '__dict__': <attribute '__dict__' of 'CFG' objects>, '__weakref__': <attribute '__weakref__' of 'CFG' objects>, '__doc__': None}\n","__weakref__ <attribute '__weakref__' of 'CFG' objects>\n","__doc__ None\n"]}],"source":["# cv\n","rmses = []\n","\n","for target in targets:\n","    models = model_dict[target]\n","\n","    preds = []\n","    trues = []\n","    \n","    for fold, model in enumerate(models):\n","        X_eval_cv = train[train[\"fold\"] == fold].drop(columns=drop_columns)\n","        y_eval_cv = train[train[\"fold\"] == fold][target]\n","\n","        pred = model.predict(X_eval_cv)\n","\n","        trues.extend(y_eval_cv)\n","        preds.extend(pred)\n","        \n","    rmse = np.sqrt(mean_squared_error(trues, preds))\n","    print(f\"{target}_rmse : {rmse}\")\n","    rmses = rmses + [rmse]\n","\n","print(f\"mcrmse : {sum(rmses) / len(rmses)}\")\n","# deberta large mcrmse : 0.8319814894700264\n","# print CFG config \n","for x in CFG.__dict__:\n","    print(x, getattr(CFG, x))"]},{"cell_type":"markdown","metadata":{},"source":["## Predict"]},{"cell_type":"code","execution_count":22,"metadata":{"trusted":true},"outputs":[],"source":["drop_columns = [\n","                #\"fold\", \n","                \"student_id\", \"prompt_id\", \"text\", \"fixed_summary_text\",\n","                \"prompt_question\", \"prompt_title\", \n","                \"prompt_text\",\n","                \"input\"\n","               ] + [\n","                f\"content_pred_{i}\" for i in range(CFG.n_splits)\n","                ] + [\n","                f\"wording_pred_{i}\" for i in range(CFG.n_splits)\n","                ]"]},{"cell_type":"code","execution_count":23,"metadata":{"trusted":true},"outputs":[],"source":["pred_dict = {}\n","for target in targets:\n","    models = model_dict[target]\n","    preds = []\n","\n","    for fold, model in enumerate(models):\n","        X_eval_cv = test.drop(columns=drop_columns)\n","\n","        pred = model.predict(X_eval_cv)\n","        preds.append(pred)\n","    \n","    pred_dict[target] = preds"]},{"cell_type":"code","execution_count":24,"metadata":{"trusted":true},"outputs":[],"source":["for target in targets:\n","    preds = pred_dict[target]\n","    for i, pred in enumerate(preds):\n","        test[f\"{target}_pred_{i}\"] = pred\n","\n","    test[target] = test[[f\"{target}_pred_{fold}\" for fold in range(CFG.n_splits)]].mean(axis=1)"]},{"cell_type":"code","execution_count":25,"metadata":{"trusted":true},"outputs":[{"data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>student_id</th>\n","      <th>prompt_id</th>\n","      <th>text</th>\n","      <th>summary_length</th>\n","      <th>fixed_summary_text</th>\n","      <th>splling_err_num</th>\n","      <th>prompt_question</th>\n","      <th>prompt_title</th>\n","      <th>prompt_text</th>\n","      <th>prompt_length</th>\n","      <th>...</th>\n","      <th>trigram_overlap_count</th>\n","      <th>trigram_overlap_ratio</th>\n","      <th>quotes_count</th>\n","      <th>input</th>\n","      <th>content_pred_0</th>\n","      <th>content_pred_1</th>\n","      <th>content</th>\n","      <th>wording_pred_0</th>\n","      <th>wording_pred_1</th>\n","      <th>wording</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>000000ffffff</td>\n","      <td>abc123</td>\n","      <td>Example text 1</td>\n","      <td>3</td>\n","      <td>Example text 1</td>\n","      <td>0</td>\n","      <td>Summarize...</td>\n","      <td>Example Title 1</td>\n","      <td>Heading\\nText...</td>\n","      <td>3</td>\n","      <td>...</td>\n","      <td>0</td>\n","      <td>0.0</td>\n","      <td>0</td>\n","      <td>Example Title 1[SEP]Summarize...[SEP]Example t...</td>\n","      <td>-0.924872</td>\n","      <td>-0.324712</td>\n","      <td>-0.624792</td>\n","      <td>-0.543148</td>\n","      <td>0.010063</td>\n","      <td>-0.266543</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>111111eeeeee</td>\n","      <td>def789</td>\n","      <td>Example text 2</td>\n","      <td>3</td>\n","      <td>Example text 2</td>\n","      <td>0</td>\n","      <td>Summarize...</td>\n","      <td>Example Title 2</td>\n","      <td>Heading\\nText...</td>\n","      <td>3</td>\n","      <td>...</td>\n","      <td>0</td>\n","      <td>0.0</td>\n","      <td>0</td>\n","      <td>Example Title 2[SEP]Summarize...[SEP]Example t...</td>\n","      <td>-0.924872</td>\n","      <td>-0.324712</td>\n","      <td>-0.624792</td>\n","      <td>-0.543148</td>\n","      <td>0.010063</td>\n","      <td>-0.266543</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>222222cccccc</td>\n","      <td>abc123</td>\n","      <td>Example text 3</td>\n","      <td>3</td>\n","      <td>Example text 3</td>\n","      <td>0</td>\n","      <td>Summarize...</td>\n","      <td>Example Title 1</td>\n","      <td>Heading\\nText...</td>\n","      <td>3</td>\n","      <td>...</td>\n","      <td>0</td>\n","      <td>0.0</td>\n","      <td>0</td>\n","      <td>Example Title 1[SEP]Summarize...[SEP]Example t...</td>\n","      <td>-0.924872</td>\n","      <td>-0.324712</td>\n","      <td>-0.624792</td>\n","      <td>-0.543148</td>\n","      <td>0.010063</td>\n","      <td>-0.266543</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>333333dddddd</td>\n","      <td>def789</td>\n","      <td>Example text 4</td>\n","      <td>3</td>\n","      <td>Example text 4</td>\n","      <td>0</td>\n","      <td>Summarize...</td>\n","      <td>Example Title 2</td>\n","      <td>Heading\\nText...</td>\n","      <td>3</td>\n","      <td>...</td>\n","      <td>0</td>\n","      <td>0.0</td>\n","      <td>0</td>\n","      <td>Example Title 2[SEP]Summarize...[SEP]Example t...</td>\n","      <td>-0.924872</td>\n","      <td>-0.324712</td>\n","      <td>-0.624792</td>\n","      <td>-0.543148</td>\n","      <td>0.010063</td>\n","      <td>-0.266543</td>\n","    </tr>\n","  </tbody>\n","</table>\n","<p>4 rows × 23 columns</p>\n","</div>"],"text/plain":["     student_id prompt_id            text  summary_length fixed_summary_text  \\\n","0  000000ffffff    abc123  Example text 1               3     Example text 1   \n","1  111111eeeeee    def789  Example text 2               3     Example text 2   \n","2  222222cccccc    abc123  Example text 3               3     Example text 3   \n","3  333333dddddd    def789  Example text 4               3     Example text 4   \n","\n","   splling_err_num prompt_question     prompt_title       prompt_text  \\\n","0                0    Summarize...  Example Title 1  Heading\\nText...   \n","1                0    Summarize...  Example Title 2  Heading\\nText...   \n","2                0    Summarize...  Example Title 1  Heading\\nText...   \n","3                0    Summarize...  Example Title 2  Heading\\nText...   \n","\n","   prompt_length  ...  trigram_overlap_count  trigram_overlap_ratio  \\\n","0              3  ...                      0                    0.0   \n","1              3  ...                      0                    0.0   \n","2              3  ...                      0                    0.0   \n","3              3  ...                      0                    0.0   \n","\n","   quotes_count                                              input  \\\n","0             0  Example Title 1[SEP]Summarize...[SEP]Example t...   \n","1             0  Example Title 2[SEP]Summarize...[SEP]Example t...   \n","2             0  Example Title 1[SEP]Summarize...[SEP]Example t...   \n","3             0  Example Title 2[SEP]Summarize...[SEP]Example t...   \n","\n","   content_pred_0  content_pred_1   content  wording_pred_0  wording_pred_1  \\\n","0       -0.924872       -0.324712 -0.624792       -0.543148        0.010063   \n","1       -0.924872       -0.324712 -0.624792       -0.543148        0.010063   \n","2       -0.924872       -0.324712 -0.624792       -0.543148        0.010063   \n","3       -0.924872       -0.324712 -0.624792       -0.543148        0.010063   \n","\n","    wording  \n","0 -0.266543  \n","1 -0.266543  \n","2 -0.266543  \n","3 -0.266543  \n","\n","[4 rows x 23 columns]"]},"execution_count":25,"metadata":{},"output_type":"execute_result"}],"source":["test"]},{"cell_type":"markdown","metadata":{},"source":["## Create Submission file"]},{"cell_type":"code","execution_count":26,"metadata":{"trusted":true},"outputs":[{"data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>student_id</th>\n","      <th>content</th>\n","      <th>wording</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>000000ffffff</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>111111eeeeee</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>222222cccccc</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>333333dddddd</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["     student_id  content  wording\n","0  000000ffffff      0.0      0.0\n","1  111111eeeeee      0.0      0.0\n","2  222222cccccc      0.0      0.0\n","3  333333dddddd      0.0      0.0"]},"execution_count":26,"metadata":{},"output_type":"execute_result"}],"source":["sample_submission"]},{"cell_type":"code","execution_count":27,"metadata":{"trusted":true},"outputs":[{"ename":"","evalue":"","output_type":"error","traceback":["\u001b[1;31mThe Kernel crashed while executing code in the the current cell or a previous cell. Please review the code in the cell(s) to identify a possible cause of the failure. Click <a href='https://aka.ms/vscodeJupyterKernelCrash'>here</a> for more info. View Jupyter <a href='command:jupyter.viewOutput'>log</a> for further details."]}],"source":["test[[\"student_id\", \"content\", \"wording\"]].to_csv(\"submission.csv\", index=False)"]},{"cell_type":"markdown","metadata":{},"source":["## Summary\n","\n","CV result is like this.\n","\n","| | content rmse |wording rmse | mcrmse | LB| |\n","| -- | -- | -- | -- | -- | -- |\n","|baseline| 0.494 | 0.630 | 0.562 | 0.509 | [link](https://www.kaggle.com/code/tsunotsuno/debertav3-baseline-content-and-wording-models)|\n","| use title and question field | 0.476| 0.619 | 0.548 | 0.508 | [link](https://www.kaggle.com/code/tsunotsuno/debertav3-w-prompt-title-question-fields) |\n","| Debertav3 + LGBM | 0.451 | 0.591 | 0.521 | 0.461 | [link](https://www.kaggle.com/code/tsunotsuno/debertav3-lgbm-with-feature-engineering) |\n","| Debertav3 + LGBM with spell autocorrect | 0.448 | 0.581 | 0.514 | 0.459 |nogawanogawa's original code\n","| Debertav3 + LGBM with spell autocorrect and tuning | 0.442 | 0.566 | 0.504 | 0.453 | this notebook |\n","\n","The CV values improved slightly, and the LB value is improved."]}],"metadata":{"kernelspec":{"display_name":"Python 3","language":"python","name":"python3"},"language_info":{"codemirror_mode":{"name":"ipython","version":3},"file_extension":".py","mimetype":"text/x-python","name":"python","nbconvert_exporter":"python","pygments_lexer":"ipython3","version":"3.8.10"}},"nbformat":4,"nbformat_minor":4}