{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "ModernBertForScoring(\n",
       "  (model): ModernBertModel(\n",
       "    (embeddings): ModernBertEmbeddings(\n",
       "      (tok_embeddings): Embedding(102400, 512, padding_idx=3)\n",
       "      (norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "      (drop): Dropout(p=0.0, inplace=False)\n",
       "    )\n",
       "    (layers): ModuleList(\n",
       "      (0): ModernBertEncoderLayer(\n",
       "        (attn_norm): Identity()\n",
       "        (attn): ModernBertAttention(\n",
       "          (Wqkv): Linear(in_features=512, out_features=1536, bias=False)\n",
       "          (rotary_emb): ModernBertUnpaddedRotaryEmbedding(dim=64, base=160000.0, scale_base=None)\n",
       "          (Wo): Linear(in_features=512, out_features=512, bias=False)\n",
       "          (out_drop): Identity()\n",
       "        )\n",
       "        (mlp_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): ModernBertMLP(\n",
       "          (Wi): Linear(in_features=512, out_features=4096, bias=False)\n",
       "          (act): GELUActivation()\n",
       "          (drop): Dropout(p=0.0, inplace=False)\n",
       "          (Wo): Linear(in_features=2048, out_features=512, bias=False)\n",
       "        )\n",
       "      )\n",
       "      (1-2): 2 x ModernBertEncoderLayer(\n",
       "        (attn_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (attn): ModernBertAttention(\n",
       "          (Wqkv): Linear(in_features=512, out_features=1536, bias=False)\n",
       "          (rotary_emb): ModernBertUnpaddedRotaryEmbedding(dim=64, base=10000.0, scale_base=None)\n",
       "          (Wo): Linear(in_features=512, out_features=512, bias=False)\n",
       "          (out_drop): Identity()\n",
       "        )\n",
       "        (mlp_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): ModernBertMLP(\n",
       "          (Wi): Linear(in_features=512, out_features=4096, bias=False)\n",
       "          (act): GELUActivation()\n",
       "          (drop): Dropout(p=0.0, inplace=False)\n",
       "          (Wo): Linear(in_features=2048, out_features=512, bias=False)\n",
       "        )\n",
       "      )\n",
       "      (3): ModernBertEncoderLayer(\n",
       "        (attn_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (attn): ModernBertAttention(\n",
       "          (Wqkv): Linear(in_features=512, out_features=1536, bias=False)\n",
       "          (rotary_emb): ModernBertUnpaddedRotaryEmbedding(dim=64, base=160000.0, scale_base=None)\n",
       "          (Wo): Linear(in_features=512, out_features=512, bias=False)\n",
       "          (out_drop): Identity()\n",
       "        )\n",
       "        (mlp_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): ModernBertMLP(\n",
       "          (Wi): Linear(in_features=512, out_features=4096, bias=False)\n",
       "          (act): GELUActivation()\n",
       "          (drop): Dropout(p=0.0, inplace=False)\n",
       "          (Wo): Linear(in_features=2048, out_features=512, bias=False)\n",
       "        )\n",
       "      )\n",
       "      (4-5): 2 x ModernBertEncoderLayer(\n",
       "        (attn_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (attn): ModernBertAttention(\n",
       "          (Wqkv): Linear(in_features=512, out_features=1536, bias=False)\n",
       "          (rotary_emb): ModernBertUnpaddedRotaryEmbedding(dim=64, base=10000.0, scale_base=None)\n",
       "          (Wo): Linear(in_features=512, out_features=512, bias=False)\n",
       "          (out_drop): Identity()\n",
       "        )\n",
       "        (mlp_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): ModernBertMLP(\n",
       "          (Wi): Linear(in_features=512, out_features=4096, bias=False)\n",
       "          (act): GELUActivation()\n",
       "          (drop): Dropout(p=0.0, inplace=False)\n",
       "          (Wo): Linear(in_features=2048, out_features=512, bias=False)\n",
       "        )\n",
       "      )\n",
       "      (6): ModernBertEncoderLayer(\n",
       "        (attn_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (attn): ModernBertAttention(\n",
       "          (Wqkv): Linear(in_features=512, out_features=1536, bias=False)\n",
       "          (rotary_emb): ModernBertUnpaddedRotaryEmbedding(dim=64, base=160000.0, scale_base=None)\n",
       "          (Wo): Linear(in_features=512, out_features=512, bias=False)\n",
       "          (out_drop): Identity()\n",
       "        )\n",
       "        (mlp_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): ModernBertMLP(\n",
       "          (Wi): Linear(in_features=512, out_features=4096, bias=False)\n",
       "          (act): GELUActivation()\n",
       "          (drop): Dropout(p=0.0, inplace=False)\n",
       "          (Wo): Linear(in_features=2048, out_features=512, bias=False)\n",
       "        )\n",
       "      )\n",
       "      (7-8): 2 x ModernBertEncoderLayer(\n",
       "        (attn_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (attn): ModernBertAttention(\n",
       "          (Wqkv): Linear(in_features=512, out_features=1536, bias=False)\n",
       "          (rotary_emb): ModernBertUnpaddedRotaryEmbedding(dim=64, base=10000.0, scale_base=None)\n",
       "          (Wo): Linear(in_features=512, out_features=512, bias=False)\n",
       "          (out_drop): Identity()\n",
       "        )\n",
       "        (mlp_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): ModernBertMLP(\n",
       "          (Wi): Linear(in_features=512, out_features=4096, bias=False)\n",
       "          (act): GELUActivation()\n",
       "          (drop): Dropout(p=0.0, inplace=False)\n",
       "          (Wo): Linear(in_features=2048, out_features=512, bias=False)\n",
       "        )\n",
       "      )\n",
       "      (9): ModernBertEncoderLayer(\n",
       "        (attn_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (attn): ModernBertAttention(\n",
       "          (Wqkv): Linear(in_features=512, out_features=1536, bias=False)\n",
       "          (rotary_emb): ModernBertUnpaddedRotaryEmbedding(dim=64, base=160000.0, scale_base=None)\n",
       "          (Wo): Linear(in_features=512, out_features=512, bias=False)\n",
       "          (out_drop): Identity()\n",
       "        )\n",
       "        (mlp_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): ModernBertMLP(\n",
       "          (Wi): Linear(in_features=512, out_features=4096, bias=False)\n",
       "          (act): GELUActivation()\n",
       "          (drop): Dropout(p=0.0, inplace=False)\n",
       "          (Wo): Linear(in_features=2048, out_features=512, bias=False)\n",
       "        )\n",
       "      )\n",
       "      (10-11): 2 x ModernBertEncoderLayer(\n",
       "        (attn_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (attn): ModernBertAttention(\n",
       "          (Wqkv): Linear(in_features=512, out_features=1536, bias=False)\n",
       "          (rotary_emb): ModernBertUnpaddedRotaryEmbedding(dim=64, base=10000.0, scale_base=None)\n",
       "          (Wo): Linear(in_features=512, out_features=512, bias=False)\n",
       "          (out_drop): Identity()\n",
       "        )\n",
       "        (mlp_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): ModernBertMLP(\n",
       "          (Wi): Linear(in_features=512, out_features=4096, bias=False)\n",
       "          (act): GELUActivation()\n",
       "          (drop): Dropout(p=0.0, inplace=False)\n",
       "          (Wo): Linear(in_features=2048, out_features=512, bias=False)\n",
       "        )\n",
       "      )\n",
       "      (12): ModernBertEncoderLayer(\n",
       "        (attn_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (attn): ModernBertAttention(\n",
       "          (Wqkv): Linear(in_features=512, out_features=1536, bias=False)\n",
       "          (rotary_emb): ModernBertUnpaddedRotaryEmbedding(dim=64, base=160000.0, scale_base=None)\n",
       "          (Wo): Linear(in_features=512, out_features=512, bias=False)\n",
       "          (out_drop): Identity()\n",
       "        )\n",
       "        (mlp_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): ModernBertMLP(\n",
       "          (Wi): Linear(in_features=512, out_features=4096, bias=False)\n",
       "          (act): GELUActivation()\n",
       "          (drop): Dropout(p=0.0, inplace=False)\n",
       "          (Wo): Linear(in_features=2048, out_features=512, bias=False)\n",
       "        )\n",
       "      )\n",
       "      (13-14): 2 x ModernBertEncoderLayer(\n",
       "        (attn_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (attn): ModernBertAttention(\n",
       "          (Wqkv): Linear(in_features=512, out_features=1536, bias=False)\n",
       "          (rotary_emb): ModernBertUnpaddedRotaryEmbedding(dim=64, base=10000.0, scale_base=None)\n",
       "          (Wo): Linear(in_features=512, out_features=512, bias=False)\n",
       "          (out_drop): Identity()\n",
       "        )\n",
       "        (mlp_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): ModernBertMLP(\n",
       "          (Wi): Linear(in_features=512, out_features=4096, bias=False)\n",
       "          (act): GELUActivation()\n",
       "          (drop): Dropout(p=0.0, inplace=False)\n",
       "          (Wo): Linear(in_features=2048, out_features=512, bias=False)\n",
       "        )\n",
       "      )\n",
       "      (15): ModernBertEncoderLayer(\n",
       "        (attn_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (attn): ModernBertAttention(\n",
       "          (Wqkv): Linear(in_features=512, out_features=1536, bias=False)\n",
       "          (rotary_emb): ModernBertUnpaddedRotaryEmbedding(dim=64, base=160000.0, scale_base=None)\n",
       "          (Wo): Linear(in_features=512, out_features=512, bias=False)\n",
       "          (out_drop): Identity()\n",
       "        )\n",
       "        (mlp_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): ModernBertMLP(\n",
       "          (Wi): Linear(in_features=512, out_features=4096, bias=False)\n",
       "          (act): GELUActivation()\n",
       "          (drop): Dropout(p=0.0, inplace=False)\n",
       "          (Wo): Linear(in_features=2048, out_features=512, bias=False)\n",
       "        )\n",
       "      )\n",
       "      (16-17): 2 x ModernBertEncoderLayer(\n",
       "        (attn_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (attn): ModernBertAttention(\n",
       "          (Wqkv): Linear(in_features=512, out_features=1536, bias=False)\n",
       "          (rotary_emb): ModernBertUnpaddedRotaryEmbedding(dim=64, base=10000.0, scale_base=None)\n",
       "          (Wo): Linear(in_features=512, out_features=512, bias=False)\n",
       "          (out_drop): Identity()\n",
       "        )\n",
       "        (mlp_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): ModernBertMLP(\n",
       "          (Wi): Linear(in_features=512, out_features=4096, bias=False)\n",
       "          (act): GELUActivation()\n",
       "          (drop): Dropout(p=0.0, inplace=False)\n",
       "          (Wo): Linear(in_features=2048, out_features=512, bias=False)\n",
       "        )\n",
       "      )\n",
       "      (18): ModernBertEncoderLayer(\n",
       "        (attn_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (attn): ModernBertAttention(\n",
       "          (Wqkv): Linear(in_features=512, out_features=1536, bias=False)\n",
       "          (rotary_emb): ModernBertUnpaddedRotaryEmbedding(dim=64, base=160000.0, scale_base=None)\n",
       "          (Wo): Linear(in_features=512, out_features=512, bias=False)\n",
       "          (out_drop): Identity()\n",
       "        )\n",
       "        (mlp_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): ModernBertMLP(\n",
       "          (Wi): Linear(in_features=512, out_features=4096, bias=False)\n",
       "          (act): GELUActivation()\n",
       "          (drop): Dropout(p=0.0, inplace=False)\n",
       "          (Wo): Linear(in_features=2048, out_features=512, bias=False)\n",
       "        )\n",
       "      )\n",
       "    )\n",
       "    (final_norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "  )\n",
       "  (head): ModernBertPredictionHead(\n",
       "    (dense): Linear(in_features=512, out_features=512, bias=False)\n",
       "    (act): GELUActivation()\n",
       "    (norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n",
       "  )\n",
       "  (drop): Dropout(p=0.0, inplace=False)\n",
       "  (classifier): Linear(in_features=512, out_features=1, bias=True)\n",
       "  (sigmoid): Sigmoid()\n",
       ")"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import torch\n",
    "from transformers import AutoTokenizer\n",
    "\n",
    "# カスタムクラスが必要な場合はそちらを import\n",
    "# from your_module import ModernBertForScoring\n",
    "\n",
    "MODEL_DIR = \"./modernbert_jamt_finetune_ckpt_49\"  # 実際のパスに置き換えてください\n",
    "\n",
    "# もし学習時のクラスがカスタムクラス ModernBertForScoring なら\n",
    "# model = ModernBertForScoring.from_pretrained(MODEL_DIR)\n",
    "\n",
    "# もし学習時に ModernBertForSequenceClassification などを使ったなら（config.jsonを修正済み）\n",
    "# from transformers import AutoModelForSequenceClassification\n",
    "# model = AutoModelForSequenceClassification.from_pretrained(MODEL_DIR)\n",
    "\n",
    "# 例：カスタムクラス ModernBertForScoring の場合\n",
    "from train_jmtb_v6 import ModernBertForScoring\n",
    "model = ModernBertForScoring.from_pretrained(MODEL_DIR)\n",
    "\n",
    "tokenizer = AutoTokenizer.from_pretrained(MODEL_DIR)\n",
    "\n",
    "# GPU利用する場合\n",
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
    "model.to(device)\n",
    "model.eval()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Predicted score: 0.3882\n"
     ]
    }
   ],
   "source": [
    "def predict_score(text: str, model, tokenizer, device):\n",
    "    \"\"\"\n",
    "    1つのテキストに対し、学習済みモデルで 0.0~1.0 の推定スコアを返す\n",
    "    (ModernBertForScoring で Sigmoidがかかっている想定)\n",
    "    \"\"\"\n",
    "    # トークナイズ\n",
    "    inputs = tokenizer(\n",
    "        text,\n",
    "        return_tensors=\"pt\",\n",
    "        truncation=True,\n",
    "        max_length=512\n",
    "    )\n",
    "    # GPUへ移動\n",
    "    inputs = {k: v.to(device) for k, v in inputs.items()}\n",
    "\n",
    "    # 推論\n",
    "    with torch.no_grad():\n",
    "        outputs = model(**inputs)\n",
    "        # ModernBertForScoring なら outputs.logits が [batch_size,1]\n",
    "        score = outputs.logits.squeeze().item()  # floatに変換\n",
    "\n",
    "    return score\n",
    "\n",
    "# ------------------------\n",
    "# 推論テスト\n",
    "# ------------------------\n",
    "example_text = \"これはテスト入力です。BERTに対するテストを行います。\"\n",
    "pred_score = predict_score(example_text, model, tokenizer, device)\n",
    "print(f\"Predicted score: {pred_score:.4f}\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 学習時に保存したデータセットpickle (floatラベル)\n",
    "with open(r\"/media/kurogane/kioxia1/dataset/sss/pixiv/modernbert_jamt_finetune_ckpt_49/dataset_dict_float.pkl\", \"rb\") as file:\n",
    "    dataset_dict = pickle.load(file)\n",
    "\n",
    "# テストセットだけ取り出す (train/validation も必要なら適宜呼び出す)\n",
    "test_dataset = dataset_dict[\"test\"]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Dataset({\n",
       "    features: ['input_text', 'label'],\n",
       "    num_rows: 648\n",
       "})"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tqdm import tqdm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 648/648 [00:04<00:00, 132.07it/s]\n"
     ]
    }
   ],
   "source": [
    "l_estimate_scores = []\n",
    "for i_dataset in tqdm(test_dataset):\n",
    "    # print(i_dataset)\n",
    "    f_estimate_score = predict_score(i_dataset['input_text'], model, tokenizer, device)\n",
    "    l_estimate_scores.append([f_estimate_score, i_dataset[\"label\"]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[[0.8853596448898315, 0.9],\n",
       " [0.7726119756698608, 0.9],\n",
       " [0.9444791674613953, 0.8],\n",
       " [0.8277913928031921, 0.9],\n",
       " [0.650458574295044, 0.6],\n",
       " [0.9936065673828125, 1.0],\n",
       " [0.8900719881057739, 1.0],\n",
       " [0.9954805374145508, 0.9],\n",
       " [0.8674108386039734, 1.0],\n",
       " [0.6612706184387207, 0.6],\n",
       " [0.7883831262588501, 0.9],\n",
       " [0.8626026511192322, 0.9],\n",
       " [0.8753176927566528, 0.9],\n",
       " [0.8415157794952393, 1.0],\n",
       " [0.8576846718788147, 0.9],\n",
       " [0.8369491100311279, 0.8],\n",
       " [0.6891637444496155, 0.4],\n",
       " [0.5401517152786255, 0.8],\n",
       " [0.6221821308135986, 0.7],\n",
       " [0.7067455053329468, 0.9],\n",
       " [0.862845778465271, 0.9],\n",
       " [0.754692554473877, 0.9],\n",
       " [0.8646848797798157, 0.9],\n",
       " [0.8190110325813293, 0.7],\n",
       " [0.8598576784133911, 0.7],\n",
       " [0.1510585993528366, 0.2],\n",
       " [0.3246677815914154, 0.2],\n",
       " [0.47619491815567017, 0.7],\n",
       " [0.6976843476295471, 0.7],\n",
       " [0.7661383152008057, 0.8],\n",
       " [0.8208702802658081, 0.9],\n",
       " [0.8893846869468689, 0.7],\n",
       " [0.7436974048614502, 0.8],\n",
       " [0.8706310987472534, 0.9],\n",
       " [0.7577768564224243, 0.6],\n",
       " [0.4159798324108124, 0.6],\n",
       " [0.8147375583648682, 0.9],\n",
       " [0.9518447518348694, 1.0],\n",
       " [0.7909210920333862, 0.7],\n",
       " [0.5652756094932556, 0.2],\n",
       " [0.885291337966919, 0.9],\n",
       " [0.5614107847213745, 0.4],\n",
       " [0.9521855711936951, 1.0],\n",
       " [0.9538584351539612, 0.9],\n",
       " [0.7246905565261841, 0.8],\n",
       " [0.752802312374115, 0.3],\n",
       " [0.9999985694885254, 1.0],\n",
       " [0.8176718950271606, 0.9],\n",
       " [0.29216688871383667, 0.2],\n",
       " [0.7777050137519836, 0.9],\n",
       " [0.7092275619506836, 0.9],\n",
       " [0.3888046443462372, 0.2],\n",
       " [0.671532154083252, 0.6],\n",
       " [0.9784291386604309, 1.0],\n",
       " [0.7344419956207275, 0.9],\n",
       " [0.9504329562187195, 1.0],\n",
       " [0.29299959540367126, 0.4],\n",
       " [0.8813745379447937, 0.9],\n",
       " [0.9393790364265442, 1.0],\n",
       " [0.7882702946662903, 0.9],\n",
       " [0.3024066388607025, 0.2],\n",
       " [0.8905518054962158, 0.9],\n",
       " [0.31458771228790283, 0.4],\n",
       " [0.831453800201416, 0.9],\n",
       " [0.6851061582565308, 0.9],\n",
       " [0.8689720034599304, 0.9],\n",
       " [0.7875602841377258, 0.9],\n",
       " [0.9903738498687744, 1.0],\n",
       " [0.8902719616889954, 0.9],\n",
       " [0.6511611342430115, 0.9],\n",
       " [0.9400674104690552, 0.9],\n",
       " [0.8891795873641968, 1.0],\n",
       " [0.9117794632911682, 0.9],\n",
       " [0.5624850988388062, 0.4],\n",
       " [0.8355247378349304, 0.7],\n",
       " [0.5644713640213013, 0.5],\n",
       " [0.8942336440086365, 1.0],\n",
       " [0.5728762745857239, 0.3],\n",
       " [0.6248719692230225, 0.6],\n",
       " [0.8402083516120911, 0.9],\n",
       " [0.9225605726242065, 0.8],\n",
       " [0.7299030423164368, 0.9],\n",
       " [0.8318969011306763, 0.2],\n",
       " [0.7699995040893555, 0.7],\n",
       " [0.9013778567314148, 0.9],\n",
       " [0.8981260061264038, 0.9],\n",
       " [0.94044429063797, 1.0],\n",
       " [0.5691388845443726, 0.8],\n",
       " [0.906032145023346, 0.9],\n",
       " [0.7258855104446411, 0.9],\n",
       " [0.6072960495948792, 0.7],\n",
       " [0.8223610520362854, 0.9],\n",
       " [0.8334646821022034, 0.9],\n",
       " [0.7919225096702576, 0.9],\n",
       " [0.6191745400428772, 0.3],\n",
       " [0.8917948007583618, 1.0],\n",
       " [0.9037709832191467, 1.0],\n",
       " [0.9426612854003906, 0.9],\n",
       " [0.9898667335510254, 1.0],\n",
       " [0.8706862926483154, 0.9],\n",
       " [0.9408425092697144, 0.9],\n",
       " [0.547015905380249, 0.3],\n",
       " [0.6246976852416992, 0.9],\n",
       " [0.5377495288848877, 0.6],\n",
       " [0.7105527520179749, 0.6],\n",
       " [0.8361542820930481, 0.9],\n",
       " [0.854382336139679, 0.9],\n",
       " [0.7632260322570801, 0.9],\n",
       " [0.8267722129821777, 0.9],\n",
       " [0.7572315335273743, 0.9],\n",
       " [0.5597705245018005, 0.5],\n",
       " [0.5241197347640991, 0.5],\n",
       " [0.7364503145217896, 0.6],\n",
       " [0.8915094137191772, 0.9],\n",
       " [0.8340743184089661, 0.9],\n",
       " [0.8814294338226318, 0.9],\n",
       " [0.8407534956932068, 0.8],\n",
       " [0.8628779053688049, 0.9],\n",
       " [0.6497765779495239, 0.4],\n",
       " [0.8453640937805176, 0.9],\n",
       " [0.6019569635391235, 0.4],\n",
       " [0.7613986730575562, 0.9],\n",
       " [0.7194490432739258, 0.4],\n",
       " [0.7249951958656311, 0.9],\n",
       " [0.8339079022407532, 0.6],\n",
       " [0.6795671582221985, 0.9],\n",
       " [0.5414535403251648, 0.2],\n",
       " [0.8997176289558411, 0.6],\n",
       " [0.8898103833198547, 0.8],\n",
       " [0.8188425302505493, 0.9],\n",
       " [0.859573245048523, 0.9],\n",
       " [0.8335742950439453, 0.8],\n",
       " [0.35358649492263794, 0.3],\n",
       " [0.9607169032096863, 0.9],\n",
       " [0.8315537571907043, 0.5],\n",
       " [0.8553935885429382, 0.9],\n",
       " [0.31317272782325745, 0.3],\n",
       " [0.7185221910476685, 0.8],\n",
       " [0.811720073223114, 0.9],\n",
       " [0.41012346744537354, 0.9],\n",
       " [0.678536593914032, 0.7],\n",
       " [0.8444895148277283, 0.9],\n",
       " [0.9015213251113892, 0.9],\n",
       " [0.1751689314842224, 0.2],\n",
       " [0.5616938471794128, 0.5],\n",
       " [0.7608761191368103, 0.9],\n",
       " [0.8352195620536804, 0.7],\n",
       " [0.7250229716300964, 0.8],\n",
       " [0.9551438093185425, 1.0],\n",
       " [0.8680942058563232, 1.0],\n",
       " [0.3837340176105499, 0.4],\n",
       " [0.8416690230369568, 0.9],\n",
       " [0.3323519825935364, 0.3],\n",
       " [0.9254946708679199, 0.9],\n",
       " [0.41063499450683594, 0.3],\n",
       " [0.9265263080596924, 0.9],\n",
       " [0.8754938244819641, 0.9],\n",
       " [0.35498619079589844, 0.3],\n",
       " [0.6464210152626038, 0.9],\n",
       " [0.6440871357917786, 0.8],\n",
       " [0.7724695801734924, 0.9],\n",
       " [0.8986650705337524, 0.9],\n",
       " [0.6043409109115601, 0.8],\n",
       " [0.8680981993675232, 0.9],\n",
       " [0.6971868872642517, 0.6],\n",
       " [0.7866945266723633, 0.9],\n",
       " [0.7488465309143066, 0.9],\n",
       " [0.8502185940742493, 0.9],\n",
       " [0.6744739413261414, 0.9],\n",
       " [0.7877558469772339, 0.9],\n",
       " [0.7066866159439087, 0.6],\n",
       " [0.8486088514328003, 0.9],\n",
       " [0.8751844763755798, 0.8],\n",
       " [0.6643693447113037, 0.6],\n",
       " [0.9658104777336121, 1.0],\n",
       " [0.8607232570648193, 0.9],\n",
       " [0.9032878875732422, 0.9],\n",
       " [0.7011781334877014, 0.4],\n",
       " [0.6084978580474854, 0.6],\n",
       " [0.612901508808136, 0.2],\n",
       " [0.8223897218704224, 0.9],\n",
       " [0.8220482468605042, 0.7],\n",
       " [0.8246190547943115, 0.9],\n",
       " [0.9188733696937561, 1.0],\n",
       " [0.8889643549919128, 1.0],\n",
       " [0.5127625465393066, 0.3],\n",
       " [0.9323657155036926, 0.9],\n",
       " [0.8109257221221924, 0.9],\n",
       " [0.9101774096488953, 1.0],\n",
       " [0.8433371782302856, 0.9],\n",
       " [0.7009791731834412, 0.7],\n",
       " [0.4037330448627472, 0.3],\n",
       " [0.8095818758010864, 0.9],\n",
       " [0.8199410438537598, 0.9],\n",
       " [0.9239128828048706, 0.9],\n",
       " [0.9458503723144531, 0.9],\n",
       " [0.8803860545158386, 0.9],\n",
       " [0.9318424463272095, 0.9],\n",
       " [0.44887277483940125, 0.4],\n",
       " [0.870177149772644, 0.9],\n",
       " [0.6904446482658386, 0.4],\n",
       " [0.8616786599159241, 1.0],\n",
       " [0.8151728510856628, 1.0],\n",
       " [0.8659726977348328, 0.9],\n",
       " [0.704562246799469, 0.7],\n",
       " [0.8409744501113892, 0.9],\n",
       " [0.7970026135444641, 0.9],\n",
       " [0.6209415793418884, 0.4],\n",
       " [0.4070623815059662, 0.8],\n",
       " [0.4036594331264496, 0.5],\n",
       " [0.3024316430091858, 0.3],\n",
       " [0.7340905070304871, 0.6],\n",
       " [0.57145094871521, 0.6],\n",
       " [0.8337589502334595, 0.9],\n",
       " [0.36238765716552734, 0.2],\n",
       " [0.9250513315200806, 0.8],\n",
       " [0.7166903018951416, 0.3],\n",
       " [0.7488646507263184, 0.4],\n",
       " [0.731031060218811, 0.1],\n",
       " [0.7825756072998047, 0.9],\n",
       " [0.44814005494117737, 0.7],\n",
       " [0.841301679611206, 0.9],\n",
       " [0.9161314368247986, 0.9],\n",
       " [0.6954988241195679, 0.8],\n",
       " [0.858526349067688, 0.9],\n",
       " [0.715857207775116, 0.4],\n",
       " [0.8710260391235352, 0.9],\n",
       " [0.8597891926765442, 0.9],\n",
       " [0.8327909111976624, 0.8],\n",
       " [0.7761644124984741, 0.6],\n",
       " [0.8617022633552551, 0.8],\n",
       " [0.6540952920913696, 0.4],\n",
       " [0.6021221876144409, 0.8],\n",
       " [0.9648029804229736, 0.9],\n",
       " [0.945438027381897, 0.9],\n",
       " [0.7100310921669006, 0.6],\n",
       " [0.8940247297286987, 0.9],\n",
       " [0.6955048441886902, 0.6],\n",
       " [0.7425771951675415, 0.9],\n",
       " [0.785810112953186, 0.9],\n",
       " [0.7673643231391907, 0.9],\n",
       " [0.6945856809616089, 0.6],\n",
       " [0.8823869228363037, 0.9],\n",
       " [0.8154327869415283, 0.9],\n",
       " [0.40487372875213623, 0.9],\n",
       " [0.9968383312225342, 1.0],\n",
       " [0.8875244855880737, 0.9],\n",
       " [0.7641423344612122, 0.8],\n",
       " [0.14845283329486847, 0.2],\n",
       " [0.9186547994613647, 1.0],\n",
       " [0.898697018623352, 0.9],\n",
       " [0.7675024271011353, 0.9],\n",
       " [0.7242623567581177, 0.6],\n",
       " [0.22218134999275208, 0.2],\n",
       " [0.2889435589313507, 0.9],\n",
       " [0.8731667995452881, 0.9],\n",
       " [0.33617374300956726, 0.4],\n",
       " [0.8252063393592834, 0.1],\n",
       " [0.9084322452545166, 0.9],\n",
       " [0.573447585105896, 0.3],\n",
       " [0.999987006187439, 1.0],\n",
       " [0.8201974630355835, 0.8],\n",
       " [0.46674323081970215, 0.4],\n",
       " [0.7789996862411499, 0.6],\n",
       " [0.6326496005058289, 0.7],\n",
       " [0.7323980927467346, 0.9],\n",
       " [0.8065733313560486, 0.9],\n",
       " [0.9422905445098877, 0.9],\n",
       " [0.39810624718666077, 0.7],\n",
       " [0.26497283577919006, 0.2],\n",
       " [0.8232632279396057, 0.9],\n",
       " [0.8560084700584412, 0.9],\n",
       " [0.9182361364364624, 0.9],\n",
       " [0.3261430561542511, 0.3],\n",
       " [0.6424864530563354, 0.4],\n",
       " [0.705535888671875, 0.8],\n",
       " [0.37012979388237, 0.3],\n",
       " [0.879560649394989, 0.9],\n",
       " [0.8578589558601379, 0.9],\n",
       " [0.7684881687164307, 0.9],\n",
       " [0.3636164367198944, 0.4],\n",
       " [0.5062754154205322, 0.4],\n",
       " [0.7284942269325256, 0.8],\n",
       " [0.9060500264167786, 0.9],\n",
       " [0.6004676222801208, 0.5],\n",
       " [0.8459793925285339, 0.9],\n",
       " [0.8385589122772217, 0.9],\n",
       " [0.5945008397102356, 0.8],\n",
       " [0.6309816837310791, 0.6],\n",
       " [0.9385666847229004, 0.9],\n",
       " [0.6159519553184509, 0.5],\n",
       " [0.655017077922821, 0.8],\n",
       " [0.8311317563056946, 0.9],\n",
       " [0.4965081214904785, 0.6],\n",
       " [0.8611951470375061, 0.9],\n",
       " [0.8639079332351685, 1.0],\n",
       " [0.666800320148468, 0.7],\n",
       " [0.7406517267227173, 0.6],\n",
       " [0.8740969300270081, 0.9],\n",
       " [0.7778376340866089, 0.8],\n",
       " [0.31210464239120483, 0.1],\n",
       " [0.6217873692512512, 0.6],\n",
       " [0.6040461659431458, 0.6],\n",
       " [0.36912548542022705, 0.4],\n",
       " [0.4077532887458801, 0.4],\n",
       " [0.8767481446266174, 0.9],\n",
       " [0.6639301776885986, 0.7],\n",
       " [0.48675811290740967, 0.2],\n",
       " [0.7451918125152588, 0.9],\n",
       " [0.660937488079071, 0.7],\n",
       " [0.7976565957069397, 0.9],\n",
       " [0.4495948851108551, 0.6],\n",
       " [0.9202705025672913, 1.0],\n",
       " [0.7145339250564575, 0.5],\n",
       " [0.24536927044391632, 0.1],\n",
       " [0.7468162775039673, 0.9],\n",
       " [0.303520143032074, 0.2],\n",
       " [0.831900954246521, 0.9],\n",
       " [0.6076899766921997, 0.6],\n",
       " [0.8617318272590637, 0.9],\n",
       " [0.6391803622245789, 0.6],\n",
       " [0.9131392240524292, 0.9],\n",
       " [0.6205729842185974, 0.8],\n",
       " [0.6970980763435364, 0.8],\n",
       " [0.5966942310333252, 0.7],\n",
       " [0.7513420581817627, 0.7],\n",
       " [0.7164792418479919, 0.8],\n",
       " [0.9224633574485779, 0.9],\n",
       " [0.9898439645767212, 1.0],\n",
       " [0.8730711936950684, 1.0],\n",
       " [0.9467040300369263, 0.9],\n",
       " [0.8748565316200256, 1.0],\n",
       " [0.6697064638137817, 0.9],\n",
       " [0.6062734723091125, 0.3],\n",
       " [0.8608449697494507, 0.8],\n",
       " [0.8742120265960693, 0.9],\n",
       " [0.23517559468746185, 0.1],\n",
       " [0.7231286764144897, 0.9],\n",
       " [0.7599864602088928, 0.8],\n",
       " [0.7403525710105896, 0.8],\n",
       " [0.4421011805534363, 0.3],\n",
       " [0.9046536087989807, 0.9],\n",
       " [0.850109338760376, 0.8],\n",
       " [0.9594632387161255, 0.9],\n",
       " [0.9095196723937988, 0.9],\n",
       " [0.8575534820556641, 0.9],\n",
       " [0.840995728969574, 0.9],\n",
       " [0.38247016072273254, 0.2],\n",
       " [0.8575950264930725, 0.9],\n",
       " [0.43700307607650757, 0.3],\n",
       " [0.7925073504447937, 0.9],\n",
       " [0.9442921280860901, 1.0],\n",
       " [0.6393008232116699, 0.6],\n",
       " [0.7051029205322266, 0.6],\n",
       " [0.9170321226119995, 0.9],\n",
       " [0.630299985408783, 0.6],\n",
       " [0.7872487902641296, 1.0],\n",
       " [0.7631149291992188, 0.4],\n",
       " [0.5498858094215393, 0.9],\n",
       " [0.8421579599380493, 0.9],\n",
       " [0.4347347021102905, 0.5],\n",
       " [0.681331217288971, 0.9],\n",
       " [0.6790593862533569, 0.8],\n",
       " [0.6480871438980103, 0.6],\n",
       " [0.9049367308616638, 1.0],\n",
       " [0.8846441507339478, 0.8],\n",
       " [0.7245463728904724, 0.7],\n",
       " [0.9535670876502991, 1.0],\n",
       " [0.4036366939544678, 0.5],\n",
       " [0.6277033090591431, 0.5],\n",
       " [0.8284041881561279, 0.5],\n",
       " [0.897631049156189, 0.9],\n",
       " [0.9063036441802979, 0.9],\n",
       " [0.8816997408866882, 0.9],\n",
       " [0.5361436605453491, 0.3],\n",
       " [0.436718612909317, 0.2],\n",
       " [0.34977638721466064, 0.3],\n",
       " [0.9161922931671143, 0.9],\n",
       " [0.7956700921058655, 0.8],\n",
       " [0.47582751512527466, 0.8],\n",
       " [0.7620000839233398, 0.9],\n",
       " [0.7394476532936096, 0.6],\n",
       " [0.28235864639282227, 0.1],\n",
       " [0.5476358532905579, 0.9],\n",
       " [0.865868091583252, 0.9],\n",
       " [0.8919934630393982, 0.9],\n",
       " [0.8730195164680481, 0.9],\n",
       " [0.827759325504303, 0.9],\n",
       " [0.6775012612342834, 0.9],\n",
       " [0.6291446685791016, 0.3],\n",
       " [0.36084914207458496, 0.3],\n",
       " [0.9138197898864746, 0.9],\n",
       " [0.9265753030776978, 1.0],\n",
       " [0.8643822073936462, 1.0],\n",
       " [0.46074220538139343, 0.2],\n",
       " [0.7956123352050781, 0.7],\n",
       " [0.7552251219749451, 0.3],\n",
       " [0.9272438287734985, 0.9],\n",
       " [0.8851864337921143, 0.9],\n",
       " [0.8505227565765381, 0.9],\n",
       " [0.4472144842147827, 0.5],\n",
       " [0.6631287336349487, 0.7],\n",
       " [0.810291051864624, 0.9],\n",
       " [0.8809759616851807, 0.9],\n",
       " [0.727469801902771, 0.8],\n",
       " [0.831512451171875, 0.8],\n",
       " [0.4537806808948517, 0.4],\n",
       " [0.8270776867866516, 0.6],\n",
       " [0.6123011112213135, 0.3],\n",
       " [0.7847217321395874, 0.9],\n",
       " [0.6133781671524048, 0.7],\n",
       " [0.8344851136207581, 0.9],\n",
       " [0.6837958693504333, 0.9],\n",
       " [0.6095199584960938, 0.1],\n",
       " [0.7856889367103577, 0.8],\n",
       " [0.58282870054245, 0.5],\n",
       " [0.7601823210716248, 1.0],\n",
       " [0.2562515139579773, 0.1],\n",
       " [0.3906874358654022, 0.3],\n",
       " [0.7979496121406555, 0.9],\n",
       " [0.7281751036643982, 0.9],\n",
       " [0.9426548480987549, 0.9],\n",
       " [0.7257517576217651, 0.9],\n",
       " [0.8553199172019958, 0.9],\n",
       " [0.706281840801239, 0.6],\n",
       " [0.5235117673873901, 0.4],\n",
       " [0.5437600612640381, 0.4],\n",
       " [0.8904375433921814, 0.9],\n",
       " [0.9263536930084229, 0.9],\n",
       " [0.8902406692504883, 0.9],\n",
       " [0.5807684659957886, 0.2],\n",
       " [0.8884558081626892, 0.9],\n",
       " [0.45907241106033325, 0.3],\n",
       " [0.8150802850723267, 0.9],\n",
       " [0.7494222521781921, 0.8],\n",
       " [0.9023911952972412, 1.0],\n",
       " [0.8728761076927185, 0.8],\n",
       " [0.6842396855354309, 0.4],\n",
       " [0.7816420793533325, 0.6],\n",
       " [0.9999998807907104, 1.0],\n",
       " [0.36503979563713074, 0.4],\n",
       " [0.22098341584205627, 0.2],\n",
       " [0.8546743392944336, 0.9],\n",
       " [0.5210355520248413, 0.3],\n",
       " [0.5562800765037537, 0.9],\n",
       " [0.139817014336586, 0.2],\n",
       " [0.963597297668457, 0.9],\n",
       " [0.9265448451042175, 0.9],\n",
       " [0.7304396629333496, 0.9],\n",
       " [0.6448819041252136, 0.2],\n",
       " [0.6585232019424438, 0.5],\n",
       " [0.6637713313102722, 0.8],\n",
       " [0.4105170965194702, 0.6],\n",
       " [0.870437502861023, 1.0],\n",
       " [0.8676131367683411, 0.9],\n",
       " [0.9441999197006226, 0.9],\n",
       " [0.35093289613723755, 0.2],\n",
       " [0.6546303033828735, 0.9],\n",
       " [0.5971137285232544, 0.6],\n",
       " [0.5721868872642517, 0.3],\n",
       " [0.4270297884941101, 0.4],\n",
       " [0.7565173506736755, 0.9],\n",
       " [0.5655442476272583, 0.8],\n",
       " [0.8798771500587463, 0.9],\n",
       " [0.6331707239151001, 0.7],\n",
       " [0.5986428260803223, 0.6],\n",
       " [0.9033551216125488, 0.9],\n",
       " [0.62423175573349, 0.4],\n",
       " [0.8692666292190552, 0.6],\n",
       " [0.6969449520111084, 0.9],\n",
       " [0.28058552742004395, 0.3],\n",
       " [0.895095944404602, 1.0],\n",
       " [0.6236510276794434, 0.7],\n",
       " [0.1668677181005478, 0.3],\n",
       " [0.7578088045120239, 0.9],\n",
       " [0.8314532041549683, 0.7],\n",
       " [0.7570028305053711, 0.7],\n",
       " [0.7946324348449707, 0.9],\n",
       " [0.48063868284225464, 0.3],\n",
       " [0.7300595045089722, 0.9],\n",
       " [0.6662213802337646, 0.8],\n",
       " [0.7028450965881348, 0.9],\n",
       " [0.5379695296287537, 0.4],\n",
       " [0.8979285955429077, 1.0],\n",
       " [0.8592706918716431, 0.9],\n",
       " [0.9018276929855347, 0.9],\n",
       " [0.20990478992462158, 0.2],\n",
       " [0.45436862111091614, 0.6],\n",
       " [0.8687785863876343, 0.9],\n",
       " [0.9446025490760803, 0.9],\n",
       " [0.7154167294502258, 0.7],\n",
       " [0.8810001015663147, 0.9],\n",
       " [0.3845926821231842, 0.9],\n",
       " [0.836503267288208, 0.9],\n",
       " [0.9457764625549316, 0.9],\n",
       " [0.8793256282806396, 1.0],\n",
       " [0.8247732520103455, 0.9],\n",
       " [0.8680576682090759, 0.9],\n",
       " [0.9203128814697266, 0.9],\n",
       " [0.7456535696983337, 0.9],\n",
       " [0.1374431997537613, 0.1],\n",
       " [0.9480844140052795, 0.9],\n",
       " [0.7856678366661072, 0.9],\n",
       " [0.8737369775772095, 0.9],\n",
       " [0.771739661693573, 0.9],\n",
       " [0.8470485210418701, 0.7],\n",
       " [0.8958970308303833, 0.9],\n",
       " [0.882662296295166, 0.9],\n",
       " [0.8658181428909302, 0.9],\n",
       " [0.8478298187255859, 1.0],\n",
       " [0.6869574785232544, 0.8],\n",
       " [0.9119142293930054, 0.9],\n",
       " [0.8124718070030212, 0.9],\n",
       " [0.28570201992988586, 0.2],\n",
       " [0.47605857253074646, 0.4],\n",
       " [0.7397007346153259, 0.7],\n",
       " [0.8669880628585815, 0.9],\n",
       " [0.6908926367759705, 0.4],\n",
       " [0.8385718464851379, 0.8],\n",
       " [0.6881493330001831, 0.6],\n",
       " [0.4708886444568634, 0.2],\n",
       " [0.6998989582061768, 0.4],\n",
       " [0.6612530946731567, 0.4],\n",
       " [0.6099846959114075, 0.5],\n",
       " [0.37454453110694885, 0.2],\n",
       " [0.8589801788330078, 0.9],\n",
       " [0.7200120091438293, 0.9],\n",
       " [0.7728976011276245, 0.9],\n",
       " [0.9183526635169983, 1.0],\n",
       " [0.4759904444217682, 0.3],\n",
       " [0.8585455417633057, 0.9],\n",
       " [0.9012464284896851, 0.9],\n",
       " [0.8849640488624573, 0.9],\n",
       " [0.8484570980072021, 0.9],\n",
       " [0.8841190338134766, 0.8],\n",
       " [0.8012551665306091, 0.8],\n",
       " [0.3492189645767212, 0.3],\n",
       " [0.9154704809188843, 0.9],\n",
       " [0.5742915868759155, 0.6],\n",
       " [0.6070501208305359, 0.4],\n",
       " [0.9244760870933533, 0.9],\n",
       " [0.8266920447349548, 0.7],\n",
       " [0.8408470749855042, 0.9],\n",
       " [0.8546991348266602, 0.9],\n",
       " [0.6945802569389343, 0.4],\n",
       " [0.832909882068634, 0.8],\n",
       " [0.8912072777748108, 0.9],\n",
       " [0.4542817175388336, 0.4],\n",
       " [0.4879341125488281, 0.4],\n",
       " [0.9397506713867188, 0.9],\n",
       " [0.8234472870826721, 0.9],\n",
       " [0.8152168989181519, 0.9],\n",
       " [0.8133808970451355, 0.9],\n",
       " [0.9458497166633606, 1.0],\n",
       " [0.9251167178153992, 0.9],\n",
       " [0.8833156228065491, 0.9],\n",
       " [0.7805896401405334, 0.9],\n",
       " [0.2121143341064453, 0.3],\n",
       " [0.7307424545288086, 0.9],\n",
       " [0.8751575946807861, 0.9],\n",
       " [0.183512344956398, 0.2],\n",
       " [0.9652208685874939, 0.9],\n",
       " [0.5947028398513794, 0.4],\n",
       " [0.717076301574707, 0.9],\n",
       " [0.18104188144207, 0.2],\n",
       " [0.8794432282447815, 0.8],\n",
       " [0.7900682687759399, 0.9],\n",
       " [0.863516092300415, 0.9],\n",
       " [0.8091879487037659, 0.9],\n",
       " [0.6893913745880127, 0.3],\n",
       " [0.844683825969696, 0.8],\n",
       " [0.5584215521812439, 0.9],\n",
       " [0.8559276461601257, 0.1],\n",
       " [0.8497210741043091, 0.9],\n",
       " [0.8662698268890381, 0.8],\n",
       " [0.8164659738540649, 0.9],\n",
       " [0.408670037984848, 0.5],\n",
       " [0.6202747225761414, 0.4],\n",
       " [0.8251261711120605, 0.9],\n",
       " [0.7874932289123535, 0.4],\n",
       " [0.5404496788978577, 0.6],\n",
       " [0.8296866416931152, 0.9],\n",
       " [0.8826070427894592, 0.9],\n",
       " [0.9776808023452759, 0.9],\n",
       " [0.7339064478874207, 0.9],\n",
       " [0.9289330244064331, 1.0],\n",
       " [0.29415950179100037, 0.2],\n",
       " [0.22822527587413788, 0.2],\n",
       " [0.0848986804485321, 0.1],\n",
       " [0.912237823009491, 0.9],\n",
       " [0.9197036623954773, 0.9],\n",
       " [0.9009316563606262, 1.0],\n",
       " [0.5669426321983337, 0.6],\n",
       " [0.5724276900291443, 0.8],\n",
       " [0.8282648324966431, 0.6],\n",
       " [0.4727105498313904, 0.3],\n",
       " [0.4200384020805359, 0.6],\n",
       " [0.5819321870803833, 0.6],\n",
       " [0.999990701675415, 1.0],\n",
       " [0.9016520380973816, 1.0],\n",
       " [0.8751431703567505, 0.9],\n",
       " [0.9731988310813904, 1.0],\n",
       " [0.8932197690010071, 0.9],\n",
       " [0.4709314703941345, 0.4],\n",
       " [0.867012083530426, 0.9],\n",
       " [0.34234946966171265, 0.8],\n",
       " [0.7872167229652405, 1.0],\n",
       " [0.8228144645690918, 0.9],\n",
       " [0.5241886377334595, 0.3],\n",
       " [0.7403218150138855, 0.5],\n",
       " [0.5080766677856445, 0.3],\n",
       " [0.24471229314804077, 0.2],\n",
       " [0.8673152327537537, 0.9],\n",
       " [0.8606200218200684, 1.0],\n",
       " [0.958793580532074, 0.9],\n",
       " [0.7127995491027832, 0.8],\n",
       " [0.954348087310791, 0.9],\n",
       " [0.8698477149009705, 0.9],\n",
       " [0.909246563911438, 0.9],\n",
       " [0.5925514101982117, 0.1],\n",
       " [0.8720778226852417, 0.9],\n",
       " [0.8127439022064209, 0.6],\n",
       " [0.8452993035316467, 0.9],\n",
       " [0.798759937286377, 0.7],\n",
       " [0.6484208106994629, 0.7],\n",
       " [0.809238076210022, 0.7],\n",
       " [0.9439582824707031, 1.0],\n",
       " [0.3885458707809448, 0.3],\n",
       " [0.9496456980705261, 1.0],\n",
       " [0.8439157009124756, 0.9],\n",
       " [0.8469123244285583, 0.9],\n",
       " [0.6921724677085876, 0.8],\n",
       " [0.924028217792511, 0.9],\n",
       " [0.7298551201820374, 0.8],\n",
       " [0.7920827865600586, 0.9],\n",
       " [0.8530755043029785, 0.9],\n",
       " [0.6965543031692505, 0.8],\n",
       " [0.7397118806838989, 0.9],\n",
       " [0.8060211539268494, 0.9],\n",
       " [0.5099717378616333, 0.7],\n",
       " [0.8076223731040955, 0.9],\n",
       " [0.17613975703716278, 0.2],\n",
       " [0.6921284794807434, 0.6],\n",
       " [0.7126461863517761, 0.9],\n",
       " [0.7388619184494019, 0.9],\n",
       " [0.4796067476272583, 0.2],\n",
       " [0.4193134903907776, 0.4],\n",
       " [0.9194392561912537, 1.0]]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "l_estimate_scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MSE : 0.023660880108473913\n",
      "RMSE: 0.1538209352086832\n",
      "MAE : 0.10959695647068231\n",
      "R^2 : 0.6317264634099204\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 500x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 500x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAxYAAAHqCAYAAACZcdjsAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8hTgPZAAAACXBIWXMAAA9hAAAPYQGoP6dpAABEOUlEQVR4nO3deVgVdf//8ddhB1HRUExFMTWXUkxM0zS1KGyxXCqyRcWlsrhzaZPKLVPaNLPbonKt9FYz8zY1u4070tRut/S2ckmFtK+KYLlhgsL8/ujHuT2yCMzhzOHwfFwXl56Zz8y8z3zOMq/Zjs0wDEMAAAAAYIKX1QUAAAAAqPgIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAc2m03jx4+3ugzLdevWTd26dbM/TktLk81m09y5cy2r6VKX1ggAViJYAEA5evfdd2Wz2dShQ4cyz+Pw4cMaP368tm/f7rzC3FxKSopsNpv9z9fXV1dddZX69++vAwcOWF1eqWzYsEHjx4/XiRMnrC4FAMqVj9UFAIAnmz9/viIiIrRp0ybt27dPTZo0KfU8Dh8+rAkTJigiIkJt2rRxfpFu7KmnntL111+v8+fPa9u2bfrggw+0cuVK7dy5U3Xr1nVpLQ0bNtSff/4pX1/fUk23YcMGTZgwQQMHDlRISEj5FAcAboAjFgBQTlJTU7VhwwZNnTpVtWrV0vz5860uqcLp0qWLHn74YcXFxemdd97Rm2++qd9//13z5s0rcpqsrKxyqcVmsykgIEDe3t7lMn8AqOgIFgBQTubPn68aNWrozjvv1L333ltksDhx4oRGjhypiIgI+fv7q379+urfv78yMzOVkpKi66+/XpIUFxdnPzUo/zz/iIgIDRw4sMA8Lz33PicnR2PHjlVUVJSqV6+uKlWqqEuXLvrmm29K/bzS09Pl4+OjCRMmFBi3Z88e2Ww2/f3vf5cknT9/XhMmTFDTpk0VEBCgK664Qp07d9aaNWtKvVxJuvnmmyX9Fdokafz48bLZbPr555/14IMPqkaNGurcubO9/SeffKKoqCgFBgaqZs2aeuCBB3To0KEC8/3ggw/UuHFjBQYGqn379lq3bl2BNkVdY7F7927df//9qlWrlgIDA9WsWTO9+OKL9vqeffZZSVKjRo3s/ZeWllYuNQKAlTgVCgDKyfz589WnTx/5+fmpX79+eu+997R582Z7UJCkM2fOqEuXLtq1a5cGDRqktm3bKjMzU8uXL9dvv/2mFi1a6OWXX9bYsWP16KOPqkuXLpKkTp06laqWU6dOaebMmerXr5+GDh2q06dPa9asWYqJidGmTZtKdYpVWFiYunbtqsWLF2vcuHEO4xYtWiRvb2/dd999kv7asE5MTNSQIUPUvn17nTp1Slu2bNG2bdt06623luo5SNL+/fslSVdccYXD8Pvuu09NmzbV5MmTZRiGJGnSpEkaM2aM7r//fg0ZMkQZGRl65513dNNNN+mHH36wn5Y0a9YsPfbYY+rUqZNGjBihAwcO6O6771bNmjUVHh5ebD3//e9/1aVLF/n6+urRRx9VRESE9u/fry+++EKTJk1Snz59tHfvXv3jH//QW2+9pdDQUElSrVq1XFYjALiMAQBwui1bthiSjDVr1hiGYRh5eXlG/fr1jeHDhzu0Gzt2rCHJWLp0aYF55OXlGYZhGJs3bzYkGXPmzCnQpmHDhsaAAQMKDO/atavRtWtX++MLFy4Y2dnZDm3++OMPIywszBg0aJDDcEnGuHHjin1+77//viHJ2Llzp8Pwli1bGjfffLP9cWRkpHHnnXcWO6/CfPPNN4YkY/bs2UZGRoZx+PBhY+XKlUZERIRhs9mMzZs3G4ZhGOPGjTMkGf369XOYPi0tzfD29jYmTZrkMHznzp2Gj4+PfXhOTo5Ru3Zto02bNg7r54MPPjAkOazD1NTUAv1w0003GVWrVjV+/fVXh+Xk951hGMYbb7xhSDJSU1PLvUYAsBKnQgFAOZg/f77CwsLUvXt3SX+dnx8bG6uFCxcqNzfX3u6zzz5TZGSkevfuXWAeNpvNafV4e3vLz89PkpSXl6fff/9dFy5cULt27bRt27ZSz69Pnz7y8fHRokWL7MN+/PFH/fzzz4qNjbUPCwkJ0U8//aRffvmlTHUPGjRItWrVUt26dXXnnXcqKytL8+bNU7t27RzaPf744w6Ply5dqry8PN1///3KzMy0/9WpU0dNmza1nwK2ZcsWHTt2TI8//rh9/UjSwIEDVb169WJry8jI0Nq1azVo0CA1aNDAYVxJ+s4VNQKAK3EqFAA4WW5urhYuXKju3bvbrwWQpA4dOmjKlClKTk7WbbfdJumvU3v69u3rkrrmzZunKVOmaPfu3Tp//rx9eKNGjUo9r9DQUN1yyy1avHixJk6cKOmv06B8fHzUp08fe7uXX35Z99xzj66++mpde+216tGjhx555BG1bt26RMsZO3asunTpIm9vb4WGhqpFixby8Sn41XXpc/jll19kGIaaNm1a6Hzz7+z066+/SlKBdvm3ty1O/m1vr7322hI9l0u5okYAcCWCBQA42b///W8dOXJECxcu1MKFCwuMnz9/vj1YmFXUnvHc3FyHuxd98sknGjhwoHr16qVnn31WtWvXlre3txITE+3XLZTWAw88oLi4OG3fvl1t2rTR4sWLdcstt9ivI5Ckm266Sfv379c///lP/etf/9LMmTP11ltvKSkpSUOGDLnsMlq1aqXo6OjLtgsMDHR4nJeXJ5vNpi+//LLQuzgFBweX4BmWr4pQIwCUBsECAJxs/vz5ql27tmbMmFFg3NKlS/X5558rKSlJgYGBaty4sX788cdi51fcaTU1atQo9IfXfv31V4e92UuWLNFVV12lpUuXOszv0ouvS6NXr1567LHH7KdD7d27VwkJCQXa1axZU3FxcYqLi9OZM2d00003afz48SUKFmXVuHFjGYahRo0a6eqrry6yXcOGDSX9dfQg/45T0l93s0pNTVVkZGSR0+av37L2nytqBABX4hoLAHCiP//8U0uXLtVdd92le++9t8BffHy8Tp8+reXLl0uS+vbtqx07dujzzz8vMC/j/9/dqEqVKpJUaIBo3Lixvv/+e+Xk5NiHrVixosDtSvP3iOfPU5L+85//aOPGjWV+riEhIYqJidHixYu1cOFC+fn5qVevXg5tjh8/7vA4ODhYTZo0UXZ2dpmXWxJ9+vSRt7e3JkyY4PCcpb/WQX5d7dq1U61atZSUlOSwDufOnXvZX8quVauWbrrpJs2ePVsHDx4ssIx8RfWfK2oEAFfiiAUAONHy5ct1+vRp3X333YWOv+GGG+w/lhcbG6tnn31WS5Ys0X333adBgwYpKipKv//+u5YvX66kpCRFRkaqcePGCgkJUVJSkqpWraoqVaqoQ4cOatSokYYMGaIlS5aoR48euv/++7V//3598sknaty4scNy77rrLi1dulS9e/fWnXfeqdTUVCUlJally5Y6c+ZMmZ9vbGysHn74Yb377ruKiYkp8MvSLVu2VLdu3RQVFaWaNWtqy5YtWrJkieLj48u8zJJo3LixXnnlFSUkJCgtLU29evVS1apVlZqaqs8//1yPPvqonnnmGfn6+uqVV17RY489pptvvlmxsbFKTU3VnDlzSnT9wvTp09W5c2e1bdtWjz76qBo1aqS0tDStXLlS27dvlyRFRUVJkl588UU98MAD8vX1Vc+ePV1WIwC4jEV3owIAj9SzZ08jICDAyMrKKrLNwIEDDV9fXyMzM9MwDMM4fvy4ER8fb9SrV8/w8/Mz6tevbwwYMMA+3jAM45///KfRsmVLw8fHp8AtT6dMmWLUq1fP8Pf3N2688UZjy5YtBW43m5eXZ0yePNlo2LCh4e/vb1x33XXGihUrjAEDBhgNGzZ0qE8luN1svlOnThmBgYGGJOOTTz4pMP6VV14x2rdvb4SEhBiBgYFG8+bNjUmTJhk5OTnFzjf/drOffvppse3ybzebkZFR6PjPPvvM6Ny5s1GlShWjSpUqRvPmzY0nn3zS2LNnj0O7d99912jUqJHh7+9vtGvXzli7dm2BdVjY7WYNwzB+/PFHo3fv3kZISIgREBBgNGvWzBgzZoxDm4kTJxr16tUzvLy8Ctx61pk1AoCVbIZxyfFXAAAAACglrrEAAAAAYBrBAgAAAIBpBAsAAAAAphEsAAAAAJhGsAAAAABgGsECAAAAgGmV7gfy8vLydPjwYVWtWlU2m83qcgAAAAC3ZRiGTp8+rbp168rLq/hjEpUuWBw+fFjh4eFWlwEAAABUGIcOHVL9+vWLbVPpgkXVqlUl/bVyqlWrZnE1AAAAgPs6deqUwsPD7dvQxal0wSL/9Kdq1aoRLAAAAIASKMklBFy8DQAAAMA0ggUAAAAA0wgWAAAAAEyrdNdYAAAAwHny8vKUk5NjdRkoI19fX3l7eztlXgQLAAAAlElOTo5SU1OVl5dndSkwISQkRHXq1DH9G28ECwAAAJSaYRg6cuSIvL29FR4eftkfT4P7MQxDZ8+e1bFjxyRJV155pan5ESwAAABQahcuXNDZs2dVt25dBQUFWV0OyigwMFCSdOzYMdWuXdvUaVFESwAAAJRabm6uJMnPz8/iSmBWfjA8f/68qfkQLAAAAFBmZs/Lh/Wc1YcECwAAAACmESwAAAAAN2Cz2bRs2TKryygzLt4GAACA0wyeu9mly5s18PoyTbdx40Z17txZPXr00MqVK0s8XUREhEaMGKERI0aUabmezNIjFmvXrlXPnj1Vt27dEie0lJQUtW3bVv7+/mrSpInmzp1b7nUCAADAs8yaNUt/+9vftHbtWh0+fNjqcjyCpcEiKytLkZGRmjFjRonap6am6s4771T37t21fft2jRgxQkOGDNFXX31VzpUCAADAU5w5c0aLFi3SsGHDdOeddxbYUf3FF1/o+uuvV0BAgEJDQ9W7d29JUrdu3fTrr79q5MiRstls9ouex48frzZt2jjMY9q0aYqIiLA/3rx5s2699VaFhoaqevXq6tq1q7Zt21aeT9PlLA0Wt99+u1555RV7Z11OUlKSGjVqpClTpqhFixaKj4/Xvffeq7feequcKwUAAICnWLx4sZo3b65mzZrp4Ycf1uzZs2UYhiRp5cqV6t27t+644w798MMPSk5OVvv27SVJS5cuVf369fXyyy/ryJEjOnLkSImXefr0aQ0YMEDfffedvv/+ezVt2lR33HGHTp8+XS7P0QoV6hqLjRs3Kjo62mFYTEwM57gBAACgxGbNmqWHH35YktSjRw+dPHlS3377rbp166ZJkybpgQce0IQJE+ztIyMjJUk1a9aUt7e3qlatqjp16pRqmTfffLPD4w8++EAhISH69ttvddddd5l8Ru6hQt0V6ujRowoLC3MYFhYWplOnTunPP/8sdJrs7GydOnXK4Q8AAACV0549e7Rp0yb169dPkuTj46PY2FjNmjVLkrR9+3bdcsstTl9uenq6hg4dqqZNm6p69eqqVq2azpw5o4MHDzp9WVapUMGiLBITE1W9enX7X3h4uNUlua3C7uJw6bDi7vSQP+7iNq6+M0RlNnjuZpesb/q0/LjTunWnWtyVq95vFbovFsRaXUGplHZdl7VvnNmnFfr1Uc7SMrOUlpll/3++WbNm6cKFC6pbt658fHzk4+Oj9957T5999plOnjypwMBAZZ7OLtWyvLy8lH0h12FZl/6K9YABA7R9+3a9/fbb2rBhg7Zv364rrrhCOTk5Zp6mW6lQwaJOnTpKT093GJaenq5q1aopMDCw0GkSEhJ08uRJ+9+hQ4dcUao5FeyDGMBFeP8CgNu6cOGCPvroI02ZMkUrv/lr43779u3asWOH6tatq3/84x9q3bq11q9LKXIefn5+ys3NdRhWq1YtZRxLt1+nIf115ONi69ev11NPPaU77rhD11xzjfz9/ZWZmenEZ2e9CnWNRceOHbVq1SqHYWvWrFHHjh2LnMbf31/+/v7lXZpzLIiVHlxkdRUAAAAeacWKFfrjjz80ePBg/XHeRxGhVezj+vbtq1mzZumNN97QLbfconHjmumBBx7QhQsXtGrVKj3//POS/vodi7Vr1+qBBx6Qv7+/QkND1a1bN/2emanXX39dN9x8h1YuWqsvv/xS1apVs8+/adOm+vjjj9WuXTudOnVKzz77bJE7xisqS49YnDlzxp4Upb9uJ7t9+3b7uWYJCQnq37+/vf3jjz+uAwcO6LnnntPu3bv17rvvavHixRo5cqQV5cNTsIcZAIBKYdasWYqOjlb16tULjOvbt6+2bNmimjVrasasj7V8+XK1adNGN998szZt2mRv9/LLLystLU2NGzdWrVq1JEktWrTQxNff0owZM3RHt47atGmTnnnmmQLL/uOPP9S2bVs98sgjeuqpp1S7du3yfcIuZukRiy1btqh79+72x6NGjZL01zloc+fO1ZEjRxwuaGnUqJFWrlypkSNH6u2331b9+vU1c+ZMxcTEuLx2AIALcCQXqHDK+kvYrvDFF18UOa59+/b2U5mq1W2sxwc+WGi7G264QTt27Cgw/KGBQ/TiM8OVlpllPxLywgsv2Mdfd9112rzZ8ZqYe++91+HxxadSVUSWBotu3boVuwIL+1Xtbt266YcffijHqgAAAACUVoW6eBsAAACAeyJYAAAAADCNYAEAAAD3dHy/1RWgFAgWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAUBx+mR0ALo+LrCGCBQAAsAKhHR5u4MCB6tWrl/1xt27dNGLECJfXkZKSIpvNphMnTpT7siz95W0Al1gQKz24yOoqAM/A+wmwhjNDY06W5Fel+DalfJ8PHDhQ8+bNkyT5+vqqQYMG6t+/v1544QX5+JTfpvHSpUvl6+tborYpKSnq3r27/vjjD4WEhJRbTc5GsAAAAECl0qNHD7385gzVDvbWqlWr9OSTT8rX11cJCQkO7XJycuTn5+eUZdasWdMp83FnnAoFAACASsXf31+1wsLUsGFDDRs2TNHR0Vq+fLn99KW/T31ddevWVbNmzSRJhw4d0v3336+QkBDVrFlT99xzj9LS0uzzy83N1StjRiskJETXXd1Azz33nAzDcFjmpadCZWdn6/nnn1d4eLj8/f3VpEkTzZo1S2lpaerevbskqUaNGrLZbBo4cKAkKS8vT4mJiWrUqJECAwMVGRmpJUuWOCxn1apVuvrqqxUYGKju3bs71FneCBYAAACo1AIDA5WTkyNJSk5O1oH9v2jNmjVasWKFzp8/r5iYGFWtWlXr1q3T+vXrFRwcrB49etinmTJlipYsnK/Zs2fr0xX/0u+//67PP/+82GX2799f//jHPzR9+nTt2rVL77//voKDgxUeHq7PPvtMkrRnzx4dOXJEb7/9tiQpMTFRH330kZKSkvTTTz9p5MiRevjhh/Xtt99K+isA9enTRz179tT27ds1ZMgQjR49urxWWwGcCgUAAIBKyTAMJScn66uvvtLf/vY3ZWRkqEqVKnr1rRm6um4NSdInn3yivLw8zZw5UzabTZI0Z84chYSEKCUlRbfddpumTZumYcOfVp8+fZSWmaWkpCR99dVXRS537969Wrx4sdasWaPo6GhJ0lVXXWUfn3/aVO3ate3XWGRnZ2vy5Mn6+uuv1bFjR/s03333nd5//3117dpV7733nho3bqwpU6ZIkpo1a6adO3fqtddec+6KKwJHLAAA7o87CAHW8rD34IoVK3RNwzAFBATo9ttvV2xsrMaPHy9JatWqlcN1FTt27NC+fftUtWpVBQcHKzg4WDVr1tS5c+e0f/9+nTx5UkeOHFGbtu3s0/j4+Khdu3aXLtZu+/bt8vb2VteuXUtc8759+3T27Fndeuut9jqCg4P10Ucfaf/+v273u2vXLnXo0MFhuvwQ4gocsbAady0B3B/vU89Ev7o/+gjlpHv37npx0hRdVSdEdevWdbgbVJUqjnehOnPmjKKiojR//vwC86lVq1aZlh8YGFjqac6cOSNJWrlyperVq+cwzt/fv0x1OBtHLFC5XbwHxsP2xgAAgMJVqVJFEVc1VoMGDS57i9m2bdvql19+Ue3atdWkSROHv+rVq6t69eq68sortX3bFvs0Fy5c0NatW4ucZ6tWrZSXl2e/NuJS+UdMcnNz7cNatmwpf39/HTx4sEAd4eHhkqQWLVpo06ZNDvP6/vvvi18ZTkSwAAAAAIrw0EMPKTQ0VPfcc4/WrVun1NRUpaSk6KmnntJvv/0mSRo+fLiSpk/VsmXLtP+XPXriiSeK/UG6iIgIDRgwQIMGDdKyZcvs81y8eLEkqWHDhrLZbFqxYoUyMjJ05swZVa1aVc8884xGjhypefPmaf/+/dq2bZveeecd++9yPP744/rll1/07LPPas+ePVqwYIHmzp1b3qvIjmABAAAAFCEoKEhr165VgwYN1KdPH7Vo0UKDBw/WuXPnVK1aNUnS008/rd73PaABAwaoz+23qGrVqurdu3ex833vvfd077336oknnlDz5s01dOhQZWVlSZLq1aunCRMmaPTo0QoLC1N8fLwkaeLEiRozZowSExPVokUL9ejRQytXrlSjRo0kSQ0aNNBnn32mZcuWKTIyUklJSZo8eXI5rh1HXGMBAAAA53HmdTHH90tXNHbe/CT7Hvy0zKwSj6tTp479qEBhfHx8NHbS65r9/gylZWYpIrTgr4WnpKQ4PA4ICNDUqVM1derUQuc5ZswYjRkzxmGYzWbT8OHDNXz48CJrueuuu3TXXXc5DIuLiyuyvTNxxAIAAACAaQQLAAAAuLfj+62uACVAsADgmbjLF+D+eJ9WLPQXLoNgAQBwT2zEAECFQrAAAAAAYBrBAgDKE3vdAXg4wzCsWTDXXThNXl6eU+bD7WYBVD4LYp17O0RUfLwmgFLz9fWVzWZTRkaGavnmynbunPMXkpMrnTv3v38LG1eE3PPZkqRz57yVez5b5855Fxh/6bCSjCtueUUty10ZhqGcnBxlZGTIy8vL/ovfZUWwAABYi416oELy9vZW/fr19dtvvynt9DHpRK7zF5KV8dd88/+9eLhU7DKPn/krWOSc8NfxM9nKOeFfYPylw0oyrrjlFbUsdxcUFKQGDRrIy8vcyUwECwCA5yK0AOUqODhYTZs21fml06Se05y/gBXTpbve+t+/Fw+XHIddYubnOyVJk3o318zPd2pS7+YFxl86rCTjilteUctyZ97e3vLx8ZHNZjM9L4IFAAAAyszb21ve5/+QAgKcP/Oc3/+ab/6/Fw+Xil3miez8JgE6kf3Xv5eOv3RYScYVt7yillVZcPE2PBcXzQKVU3m+9/lcAYAiESxQebGBAADOZeXnKp/pgOUIFh5g8NzNDn+Xa3u5/5d0XGHLK2y6wv69XJ0e7zJfgJeux6LWXVHrsbh1XpJhl76eCuvn4voejgrrt5L048VtC5u+sOmKGleS11BR/Y2Sudy6L+z/Rb3nLm17ueUU97nKZ2/pFLdui/p8LO2wopaVP6ykNdKfKlGgLGp75NL/F9e+qO/M4pZ5ue/douqpyP1KsAAAAABgGsECAEqKUy2AioX3LOBSBAtUTs76suFLCwAAQBLBwv2woVp50NcAAMCDECwAwNkIjQCASohgYRU2PACgYuLzGwAKRbAAnIENjcqN/gdQWbjy847P1gqHYAEAAComNjzdF31TKREsAAAAAJhGsKgoSP4AKqv8z79LPwf5XASsx/sQFyFYABIfjJUd/Q84jzPeT7wngQqJYAEAADwHoQSwDMHCXfHBCDjiPQFULrznPR997HEIFoAz8SEJAMD/8L1YqRAsKgLelIBr8Z5zDTPrmT4CALdDsKho+DL1PPSpdS5e9/SD56KfAcAlCBYAAABwHQK+xyJYAJfDByAAmMdnKeDxCBYAAM/DRizg3niPeiSCBQAAAADTCBZAeWBPDAAAjvhu9HgECwAACsNGEACUCsGiIuFLzr3RP65hdj3TT5UT/Q4A5Y5gURHxBQkAAAA3Q7AAilOSEEfQAwAAIFigErE6ABS3fKtr8yRm1iX9AABAmREsAABA5cDOg4qDvqqQCBbugDdPxUA/AQDgvviethzBAgAAANYjGFR4BAtUDnxYAQAAlCuCBWAlAg8AAPAQBAsAAABYgx1sHoVgAQDOwJcjAE+0IJbPN5SY5cFixowZioiIUEBAgDp06KBNmzYV237atGlq1qyZAgMDFR4erpEjR+rcuXMuqhYAAABAYSwNFosWLdKoUaM0btw4bdu2TZGRkYqJidGxY8cKbb9gwQKNHj1a48aN065duzRr1iwtWrRIL7zwgosrBwAAQIlwxKPSsDRYTJ06VUOHDlVcXJxatmyppKQkBQUFafbs2YW237Bhg2688UY9+OCDioiI0G233aZ+/fpd9igHAAAAgPJlWbDIycnR1q1bFR0d/b9ivLwUHR2tjRs3FjpNp06dtHXrVnuQOHDggFatWqU77rjDJTUDAAAAKJyPVQvOzMxUbm6uwsLCHIaHhYVp9+7dhU7z4IMPKjMzU507d5ZhGLpw4YIef/zxYk+Fys7OVnZ2tv3xqVOnnPMEAAAAANhZfvF2aaSkpGjy5Ml69913tW3bNi1dulQrV67UxIkTi5wmMTFR1atXt/+Fh4e7sGL3Mnju5kL/X9pp4TolXe9m+rO8loGyKet6Lm66wsZdOoz+LV+D5252yTp21XJg3uX6iX4sX6767suf3sw2WEViWbAIDQ2Vt7e30tPTHYanp6erTp06hU4zZswYPfLIIxoyZIhatWql3r17a/LkyUpMTFReXl6h0yQkJOjkyZP2v0OHDjn9uQBwMS4EBADA7VgWLPz8/BQVFaXk5GT7sLy8PCUnJ6tjx46FTnP27Fl5eTmW7O3tLUkyDKPQafz9/VWtWjWHPwAAUImxcwIoF5aeCjVq1Ch9+OGHmjdvnnbt2qVhw4YpKytLcXFxkqT+/fsrISHB3r5nz5567733tHDhQqWmpmrNmjUaM2aMevbsaQ8YQLH4MgEg8VkAuALvs0rHsou3JSk2NlYZGRkaO3asjh49qjZt2mj16tX2C7oPHjzocITipZdeks1m00svvaT/+7//U61atdSzZ09NmjTJqqdQ/nhTAgCAymJBrPTgIqurQBlZGiwkKT4+XvHx8YWOS0lJcXjs4+OjcePGady4cS6oDACKwZefZ6JfAaDMKtRdoQCgzDj6BwBAuSJYAAAA5yPMey76FkUgWAAAAAAwjWABABdjTxwAAGVCsLASGzAAAADwEAQLAAAAAKYRLACgrDjqCACAHcECAMoDoQMAUMkQLICSKOtGIhuXQMXD+xYAyoRgAQCo+AgDAGA5ggVQGmy8AID1yvJZnD8Nn+NAuSFYAAAAADCNYAEAcC/sUQaAColgAcCzsZEKAIBLECwAAAAAmEawAAAAAGAawQIASovTqwAAKIBgAQDwDAQ+ALAUwQIoKzZiAAAA7AgWADwDQQ8APBOf7xUGwQKA5+NLCXBfvD8rLvoOlyBYAACAyokNY8CpCBbwbHxpAAAAuATBAnAFAg5Qfnh/AYBbIFgAAAAAMI1gAZQ39qYCAIBKgGABwHOVJNQR/NwT/QJUTLx3KzWCBQAAAADTCBZAUdjrgovxegAAoFgECwAAAACmESwAVBwcNUB54HUFAE5BsAAAAJUb4RJwCoJFBTd47uZCh108vLDHpZkf3ENRfUOfea5L+9bVfc1rq2hWrBsznwH0pXlm1mFJp6WfKp78bayy9p2n9TnBAgDgHthrDGfhtQRYgmABoHIq64bHpdOxAQMAgCSCBQAAAAAnIFhUVOwlBQAAFQXbLZUCwQIAAACAaQQLAADgXOydBiolggUAAMVhIxkASoRgAbgSGyhAxcR7FwAui2ABAAAAwDSCBQAAAADTCBYAAAAATCNYAAAqDq51AAC3RbAAAACVF2EVcBqCBQAAcA9s5AMVGsHCnfEBCwAAgAqCYAEAAKzHzjSgwiNYAAAAADCNYAEAhWHvKeA+eD8CFQLBwl3woek56EsAAFAJESwAACgJdhoAQLEIFgAAwFqENsAjECwAAAAAmEawAAAAAGAawQIAYB2rT4GxevkA4EEIFgAAAABMI1gAAICKi6NOgNsgWAAAILGBCgAmESwAAAAAmEawAC7FXksAqJj4/AYsRbAAAACei7ABuAzBApUPXzKegX70LPQnULnxGeARCBYAAAAATCNYAAAAADDN8mAxY8YMRUREKCAgQB06dNCmTZuKbX/ixAk9+eSTuvLKK+Xv76+rr75aq1atclG1AAAAAArjY+XCFy1apFGjRikpKUkdOnTQtGnTFBMToz179qh27doF2ufk5OjWW29V7dq1tWTJEtWrV0+//vqrQkJCXF88AAAAADtLg8XUqVM1dOhQxcXFSZKSkpK0cuVKzZ49W6NHjy7Qfvbs2fr999+1YcMG+fr6SpIiIiJcWTIqIi4IAzwL72kAcEuWnQqVk5OjrVu3Kjo6+n/FeHkpOjpaGzduLHSa5cuXq2PHjnryyScVFhama6+9VpMnT1Zubq6rygYAAABQCMuOWGRmZio3N1dhYWEOw8PCwrR79+5Cpzlw4ID+/e9/66GHHtKqVau0b98+PfHEEzp//rzGjRtX6DTZ2dnKzs62Pz516pTzngRQHPaqVnz0IQAAJWb5xdulkZeXp9q1a+uDDz5QVFSUYmNj9eKLLyopKanIaRITE1W9enX7X3h4uAsrdn+D527W4LmbnTIfVGxF9eHFwwtrQ9+XXGnWlSvWK/1pDVet40rVl+W0E8DqdWj18lG08ugbT+hvy4JFaGiovL29lZ6e7jA8PT1dderUKXSaK6+8UldffbW8vb3tw1q0aKGjR48qJyen0GkSEhJ08uRJ+9+hQ4ec9yQAAM7FUSIUhdcG4PYsCxZ+fn6KiopScnKyfVheXp6Sk5PVsWPHQqe58cYbtW/fPuXl5dmH7d27V1deeaX8/PwKncbf31/VqlVz+AMAAICTEf4qPUtPhRo1apQ+/PBDzZs3T7t27dKwYcOUlZVlv0tU//79lZCQYG8/bNgw/f777xo+fLj27t2rlStXavLkyXryySetegoAAAAAZPHtZmNjY5WRkaGxY8fq6NGjatOmjVavXm2/oPvgwYPy8vpf9gkPD9dXX32lkSNHqnXr1qpXr56GDx+u559/3qqnAAAAAEAWBwtJio+PV3x8fKHjUlJSCgzr2LGjvv/++3KuCgAAAEBpVKi7QgEAAABwTwQLAAAAeBYuJLcEwQIAAACAaQQLAAAAAKYRLAAAQMXA6S0oCq8Nt2AqWOTk5GjPnj26cOGCs+oBgLLjiwUAAMuUKVicPXtWgwcPVlBQkK655hodPHhQkvS3v/1Nr776qlMLBMqEDUwAACoHvvPdRpmCRUJCgnbs2KGUlBQFBATYh0dHR2vRokVOKw4AAKBQbEwCbqdMP5C3bNkyLVq0SDfccINsNpt9+DXXXKP9+/c7rTgAAAAAFUOZjlhkZGSodu3aBYZnZWU5BA3AI7GXDAAAoIAyBYt27dpp5cqV9sf5YWLmzJnq2LGjcyoDAACeh50zlRv979HKdCrU5MmTdfvtt+vnn3/WhQsX9Pbbb+vnn3/Whg0b9O233zq7RgAAAABurkxHLDp37qwdO3bowoULatWqlf71r3+pdu3a2rhxo6KiopxdIwAAAAA3V+ojFufPn9djjz2mMWPG6MMPPyyPmgAAsBanawBAqZX6iIWvr68+++yz8qgFsB4bEwAAAGVSplOhevXqpWXLljm5FAAAAAAVVZku3m7atKlefvllrV+/XlFRUapSpYrD+KeeesopxQEeiaMiAADAA5UpWMyaNUshISHaunWrtm7d6jDOZrMRLFxlQaykZ6yuAgAAwHnYAVdhlSlYpKamOrsOwL3woQYAAFAqZbrG4mKGYcgwDGfUAgAAAKCCKnOw+Oijj9SqVSsFBgYqMDBQrVu31scff+zM2gAAAABUEGU6FWrq1KkaM2aM4uPjdeONN0qSvvvuOz3++OPKzMzUyJEjnVokAAAAAPdWpmDxzjvv6L333lP//v3tw+6++25dc801Gj9+PMECAAAAqGTKdCrUkSNH1KlTpwLDO3XqpCNHjpguCnB7XNwNAADgoEzBokmTJlq8eHGB4YsWLVLTpk1NFwUAADwQO2UAj1amU6EmTJig2NhYrV271n6Nxfr165WcnFxo4AAAAADg2cp0xKJv3776z3/+o9DQUC1btkzLli1TaGioNm3apN69ezu7RgAAAABurkxHLCQpKipKn3zyiTNrAQAAAFBBlemIxapVq/TVV18VGP7VV1/pyy+/NF0UAAAAgIqlTMFi9OjRys3NLTDcMAyNHj3adFEAAAAAKpYyBYtffvlFLVu2LDC8efPm2rdvn+miAAAAAFQsZQoW1atX14EDBwoM37dvn6pUqWK6KAAAAAAVS5mCxT333KMRI0Zo//799mH79u3T008/rbvvvttpxQGV0eC5m53aDshX1GuG11LZDZ672f5n1fJRedDf5Yd16xxlChavv/66qlSpoubNm6tRo0Zq1KiRmjdvriuuuEJvvvmms2sEAAAA4ObKdLvZ6tWra8OGDVqzZo127NihwMBARUZGqkuXLs6uDwAAAEAFUKojFhs3btSKFSskSTabTbfddptq166tN998U3379tWjjz6q7OzscikUAAAAgPsqVbB4+eWX9dNPP9kf79y5U0OHDtWtt96q0aNH64svvlBiYqLTiwQAAADg3koVLLZv365bbrnF/njhwoVq3769PvzwQ40aNUrTp0/X4sWLnV4kAAAAAPdWqmDxxx9/KCwszP7422+/1e23325/fP311+vQoUPOqw4AAABAhVCqYBEWFqbU1FRJUk5OjrZt26YbbrjBPv706dPy9fV1boUAAAAA3F6pgsUdd9yh0aNHa926dUpISFBQUJDDnaD++9//qnHjxk4vEgAAwKkWxFpdAeBxSnW72YkTJ6pPnz7q2rWrgoODNW/ePPn5+dnHz549W7fddpvTiwQAAADg3koVLEJDQ7V27VqdPHlSwcHB8vb2dhj/6aefKjg42KkFAgAAAHB/Zfrl7erVqxcIFZJUs2ZNhyMYAAAAgMtwipulyhQsAAAAAOBiBAsAAAAAphEsAAAAAJhGsAAAAABgGsECAAAAgGkECwAAAACmESwAAAAAmEawAAAAAGAawQJA5cEPJwEAUG4IFgAAAABMI1gAADwHR6UAwDIECwAAAACmESwAAJULRzUAoFwQLAAAKAohBABKjGABAAAAwDSCBQAAAADTCBYAAKDi4PQ0wG0RLAAAAACYRrAAAAAAYBrBAgAAAJ6D0+UsQ7AAAAAAYBrBAgAAAIBpbhEsZsyYoYiICAUEBKhDhw7atGlTiaZbuHChbDabevXqVb4FAgAAACiW5cFi0aJFGjVqlMaNG6dt27YpMjJSMTExOnbsWLHTpaWl6ZlnnlGXLl1cVCkAACh3nB9f+dDnHsPyYDF16lQNHTpUcXFxatmypZKSkhQUFKTZs2cXOU1ubq4eeughTZgwQVdddZULqwXgVvgyAgDAbVgaLHJycrR161ZFR0fbh3l5eSk6OlobN24scrqXX35ZtWvX1uDBg11RJgB35k7hwp1qAQDAxXysXHhmZqZyc3MVFhbmMDwsLEy7d+8udJrvvvtOs2bN0vbt20u0jOzsbGVnZ9sfnzp1qsz1AgAAACicpcGitE6fPq1HHnlEH374oUJDQ0s0TWJioiZMmFDOlXmmwXM3O31aM/NEybhyHdOf5gyeu1mzBl5f6PDSzqc0w+H56PvSs3qdXbz8S2uxujZPwrosX5YGi9DQUHl7eys9Pd1heHp6uurUqVOg/f79+5WWlqaePXvah+Xl5UmSfHx8tGfPHjVu3NhhmoSEBI0aNcr++NSpUwoPD3fm0wAAAAAqPUuDhZ+fn6KiopScnGy/ZWxeXp6Sk5MVHx9foH3z5s21c+dOh2EvvfSSTp8+rbfffrvQwODv7y9/f/9yqR8AAADAXyw/FWrUqFEaMGCA2rVrp/bt22vatGnKyspSXFycJKl///6qV6+eEhMTFRAQoGuvvdZh+pCQEEkqMBwAAACA61geLGJjY5WRkaGxY8fq6NGjatOmjVavXm2/oPvgwYPy8rL8rrgAAAAAimF5sJCk+Pj4Qk99kqSUlJRip507d67zCwIAAABQKhwKAABUDvzOCACUK4IFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAgPLDbX6BSoNgAcDzsCEDAIDLESwAAAAAmEawAAAAAGAawQIAAACAaQQLwF1wXQAAAKjACBYAAMB9sdMFqDAIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWACo+BbEWl0BAACVHsECJTZ47marS6hUzK7v/Okvng996L7K2jf0r3sq7P1XmmmLmq608+M1AVyemfdrUfOqrAgWAAAAAEwjWAAAAMD9cdqr2yNYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAPA93EAJcjmABAAAAwDSCBQAAAADTCBYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATCNYAAAAADCNYAEAqLgWxFpdAQDg/yNYAAAAADCNYAEAAADANIIFAAAAKgZOf3RrBAsAAAAAphEsAAAAAJhGsAAAAOWL01eASoFgAQAAAMA0ggUAAAAA09wiWMyYMUMREREKCAhQhw4dtGnTpiLbfvjhh+rSpYtq1KihGjVqKDo6utj2AAAAcDJOb0MhLA8WixYt0qhRozRu3Dht27ZNkZGRiomJ0bFjxwptn5KSon79+umbb77Rxo0bFR4erttuu03/93//5+LKAQAAAOSzPFhMnTpVQ4cOVVxcnFq2bKmkpCQFBQVp9uzZhbafP3++nnjiCbVp00bNmzfXzJkzlZeXp+TkZBdXDgAAACCfpcEiJydHW7duVXR0tH2Yl5eXoqOjtXHjxhLN4+zZszp//rxq1qxZXmUCAAAAuAwfKxeemZmp3NxchYWFOQwPCwvT7t27SzSP559/XnXr1nUIJxfLzs5Wdna2/fGpU6fKXjAAAACAQlkaLMx69dVXtXDhQqWkpCggIKDQNomJiZowYYKLK3MPg+duLrfpLm5T2vYwz93Wp7vV487MrivWdcXhzL6i313Pleu8uGVdOo7XQvmyav16Sr9aeipUaGiovL29lZ6e7jA8PT1dderUKXbaN998U6+++qr+9a9/qXXr1kW2S0hI0MmTJ+1/hw4dckrtAAAAAP7H0mDh5+enqKgohwuv8y/E7tixY5HTvf7665o4caJWr16tdu3aFbsMf39/VatWzeEPAAAAgHNZfirUqFGjNGDAALVr107t27fXtGnTlJWVpbi4OElS//79Va9ePSUmJkqSXnvtNY0dO1YLFixQRESEjh49KkkKDg5WcHCwZc8DAAAAqMwsDxaxsbHKyMjQ2LFjdfToUbVp00arV6+2X9B98OBBeXn978DKe++9p5ycHN17770O8xk3bpzGjx/vytIBAAAA/H+WBwtJio+PV3x8fKHjUlJSHB6npaWVf0EAAAAASsXyH8gDAABABbYg1uoK4CYIFgAAAHAegkalRbAAAAAAYBrBAgAAAIBpBAsAAAAAphEsAAAAAJhGsAAAAJ6Ni4kBlyBYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAFBeuHc+AKASIVgAAAAAMI1gAQAAAMA0ggUAAHAtThMEPBLBAgAAAIBpBAsAAAAAphEsAAAAAJhGsAAAAABgGsECAAAAgGkECwAAAACmESwAAAAAmEawAAAAAGAawQIAAACAaQQLAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAAphEsAMBCg+dutroEWIB+91z0rWejf4tHsAAAAABgGsECAAAAgGkECwAAAACmESwAAAAAmEawAAAAAGAawQIAAAAVx4JYqytAEQgWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDS3CBYzZsxQRESEAgIC1KFDB23atKnY9p9++qmaN2+ugIAAtWrVSqtWrXJRpQAAAAAKY3mwWLRokUaNGqVx48Zp27ZtioyMVExMjI4dO1Zo+w0bNqhfv34aPHiwfvjhB/Xq1Uu9evXSjz/+6OLKAQAAAOSzPFhMnTpVQ4cOVVxcnFq2bKmkpCQFBQVp9uzZhbZ/++231aNHDz377LNq0aKFJk6cqLZt2+rvf/+7iysHAAAAkM/SYJGTk6OtW7cqOjraPszLy0vR0dHauHFjodNs3LjRob0kxcTEFNkeAAAAQPnzsXLhmZmZys3NVVhYmMPwsLAw7d69u9Bpjh49Wmj7o0ePFto+Oztb2dnZ9scnT56UJJ06dcpM6eadPe+U2eScP+OU+VihXPvASevXCvRpEehTS9CnhaNPi0CfWoI+LRx96hz5tRiGcdm2lgYLV0hMTNSECRMKDA8PD7egmvLwudUFlNknT1hdgbuiTz0Pfep56FPPQ596HvrUmU6fPq3q1asX28bSYBEaGipvb2+lp6c7DE9PT1edOnUKnaZOnTqlap+QkKBRo0bZH+fl5en333/XFVdcIZvNZvIZlM2pU6ecFmx+/vlntWzZ0inzgnugTz0Pfep56FPPQ596Hk/q00OHDqlatWqWLNswDJ0+fVp169a9bFtLg4Wfn5+ioqKUnJysXr16Sfprwz85OVnx8fGFTtOxY0clJydrxIgR9mFr1qxRx44dC23v7+8vf39/h2EhISHOKN8tVK1a1eoS4GT0qeehTz0Pfep56FPP40l9Wq1aNcuChaTLHqnIZ/mpUKNGjdKAAQPUrl07tW/fXtOmTVNWVpbi4uIkSf3791e9evWUmJgoSRo+fLi6du2qKVOm6M4779TChQu1ZcsWffDBB1Y+DQAAAKBSszxYxMbGKiMjQ2PHjtXRo0fVpk0brV692n6B9sGDB+Xl9b+bV3Xq1EkLFizQSy+9pBdeeEFNmzbVsmXLdO2111r1FAAAAIBKz/JgIUnx8fFFnvqUkpJSYNh9992n++67r5yrKj/+/v568cUXdeHCBVPz8fHxUbVq1ZwyL7gH+tTz0Keehz71PPSp5/GkPvXx8SlwWr+7shkluXcUAAAAABTD8l/eBgAAAFDxESwAAAAAmEawAAAAAGCaW1y8XZmsXbtW999/f4Ef+QMAAADM2rRpk66//npLls0RCxdbtWqV0tPT1bp1a0my31YXAAAAuBw/Pz/ZbDb742nTpunLL7/UzTffrEaNGqldu3aW1UawcLFvvvlGTz75pHbs2CFJ6tq1q/3FcfHvdZQ3Pz8/eXl5ydvbu8g2F79oL6datWry9vYu9nZo+T8FHxwcfNlfPw8ICLjs8l25vgDgUoGBgVaXALi9KlWqWF2CRwgICLD/PycnR9L/1m2nTp10yy23aOfOnYqLiyvV9puzsWXmQjk5Odq6dauio6Ptw2w2m/Lv+JuXl1fieZndqM7JyVFeXp5yc3OLbFOaOxGfOnVKubm5ys7OLrLN4cOHJUlnzpzRiRMnip3fuXPnLrv80qwvAHC2P//80+oSALeXlZVldQkew8/Pz/5/wzDs67Zbt266/vrrdfz4ccXFxVlVniSChUtlZmYqNzfX4fSnJUuW2P/fuXPnEs/rchvVxR2JAAAAQMWRnZ1d4GyPi3cy79u3T97e3goKCnJxZY4IFhbLf1H4+PjYT48qifzgUNThruKORJSVsw+tXXrUhVObAAAACjIMQxkZGQ7D6tevL0l69NFHlZWVpYCAAH366adWlGfHlpwLhYaGytvb2+GOUOfPn5ckXbhwQadPny7xvPKDQ2GnC9lsthIdsSjtUQ1n/Ui7zWaTj49PgaMuZT21ycvLS76+vs4oDQBKzMfn8jdWZIcJKjtn75S08voBq126Hfbbb79Jkt5++21JUkREhA4ePOjyui7GJ54L+fn5KSoqSl9//bXi4+MlSQkJCfL391fdunXLdPpSYW+w/A33i/n4+BS4sNpsUCjJm7s0HwAl+QIuan6XhhJvb29OBwMqOHffKC/JZyjXgqGyc9ZOyfKaX0Vx6Q5UX19f+/UU+TeSOHr0qBo2bOjy2i5mMyprD1lk3rx5GjRokGw2m3Jzc3XFFVfoxIkT5XLqklkXX1he0ZblytoBAACcwcvL67I7JCIiIhQSEqIff/xRFy5ckCQ1aNBA586d0+7du1WjRg1XlFoofiDPxRo2bOjwgjl+/LiF1RTPlRvm7NEAAACVXUmOcqalpTk8ttlsatKkiaZPn25pqJA4YgEAAADACdz7BFYAAAAAFQLBAgAAAIBpBAsAAAAAphEsAAAAAJhGsAAAAABgGsECAAAAgGkECwAAAACmESwAAAAAmEawAAC4lM1m07Jly6wuQ+PHj1ebNm2sLgMAPAbBAgA8TEZGhoYNG6YGDRrI399fderUUUxMjNavX291aU6RlpYmm82m7du3W10KAOAiPlYXAABwrr59+yonJ0fz5s3TVVddpfT0dCUnJ+v48eNWlwYA8GAcsQAAD3LixAmtW7dOr732mrp3766GDRuqffv2SkhI0N13321vN3XqVLVq1UpVqlRReHi4nnjiCZ05c8Y+fu7cuQoJCdGKFSvUrFkzBQUF6d5779XZs2c1b948RUREqEaNGnrqqaeUm5trny4iIkITJ05Uv379VKVKFdWrV08zZswotuZDhw7p/vvvV0hIiGrWrKl77rlHaWlpJX7OKSkpstlsSk5OVrt27RQUFKROnTppz549Du1effVVhYWFqWrVqho8eLDOnTtXYF4zZ85UixYtFBAQoObNm+vdd9+1jxs0aJBat26t7OxsSVJOTo6uu+469e/fv8S1AoAnI1gAgAcJDg5WcHCwli1bZt8ALoyXl5emT5+un376SfPmzdO///1vPffccw5tzp49q+nTp2vhwoVavXq1UlJS1Lt3b61atUqrVq3Sxx9/rPfff19LlixxmO6NN95QZGSkfvjhB40ePVrDhw/XmjVrCq3j/PnziomJUdWqVbVu3TqtX79ewcHB6tGjh3Jyckr13F988UVNmTJFW7ZskY+PjwYNGmQft3jxYo0fP16TJ0/Wli1bdOWVVzqEBkmaP3++xo4dq0mTJmnXrl2aPHmyxowZo3nz5kmSpk+frqysLI0ePdq+vBMnTujvf/97qeoEAI9lAAA8ypIlS4waNWoYAQEBRqdOnYyEhARjx44dxU7z6aefGldccYX98Zw5cwxJxr59++zDHnvsMSMoKMg4ffq0fVhMTIzx2GOP2R83bNjQ6NGjh8O8Y2Njjdtvv93+WJLx+eefG4ZhGB9//LHRrFkzIy8vzz4+OzvbCAwMNL766qtCa01NTTUkGT/88INhGIbxzTffGJKMr7/+2t5m5cqVhiTjzz//NAzDMDp27Gg88cQTDvPp0KGDERkZaX/cuHFjY8GCBQ5tJk6caHTs2NH+eMOGDYavr68xZswYw8fHx1i3bl2hNQJAZcQRCwDwMH379tXhw4e1fPly9ejRQykpKWrbtq3mzp1rb/P111/rlltuUb169VS1alU98sgjOn78uM6ePWtvExQUpMaNG9sfh4WFKSIiQsHBwQ7Djh075rD8jh07Fni8a9euQmvdsWOH9u3bp6pVq9qPttSsWVPnzp3T/v37S/W8W7dubf//lVdeKUn22nbt2qUOHToUWWdWVpb279+vwYMH2+sIDg7WK6+84lBHx44d9cwzz2jixIl6+umn1blz51LVCACejIu3AcADBQQE6NZbb9Wtt96qMWPGaMiQIRo3bpwGDhyotLQ03XXXXRo2bJgmTZqkmjVr6rvvvtPgwYOVk5OjoKAgSZKvr6/DPG02W6HD8vLyylznmTNnFBUVpfnz5xcYV6tWrVLN6+LabDabJJW4tvzrSz788MMCAcTb29v+/7y8PK1fv17e3t7at29fqeoDAE/HEQsAqARatmyprKwsSdLWrVuVl5enKVOm6IYbbtDVV1+tw4cPO21Z33//fYHHLVq0KLRt27Zt9csvv6h27dpq0qSJw1/16tWdVlOLFi30n//8p8g6w8LCVLduXR04cKBAHY0aNbK3e+ONN7R79259++23Wr16tebMmeO0GgGgoiNYAIAHOX78uG6++WZ98skn+u9//6vU1FR9+umnev3113XPPfdIkpo0aaLz58/rnXfe0YEDB/Txxx8rKSnJaTWsX79er7/+uvbu3asZM2bo008/1fDhwwtt+9BDDyk0NFT33HOP1q1bp9TUVKWkpOipp57Sb7/95rSahg8frtmzZ2vOnDnau3evxo0bp59++smhzYQJE5SYmKjp06dr79692rlzp+bMmaOpU6dKkn744QeNHTtWM2fO1I033qipU6dq+PDhOnDggNPqBICKjGABAB4kODhYHTp00FtvvaWbbrpJ1157rcaMGaOhQ4fa714UGRmpqVOn6rXXXtO1116r+fPnKzEx0Wk1PP3009qyZYuuu+46vfLKK5o6dapiYmIKbRsUFKS1a9eqQYMG6tOnj1q0aGG/FWy1atWcVlNsbKzGjBmj5557TlFRUfr11181bNgwhzZDhgzRzJkzNWfOHLVq1Updu3bV3Llz1ahRI507d04PP/ywBg4cqJ49e0qSHn30UXXv3l2PPPKIwy13AaCyshmGYVhdBADAM0RERGjEiBEaMWKE1aUAAFyMIxYAAAAATCNYAAAAADCNU6EAAAAAmMYRCwAAAACmESwAAAAAmEawAAAAAGAawQIAAACAaQQLAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAApv0/RSsEIBV2R7wAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 800x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.metrics import mean_squared_error, mean_absolute_error, r2_score\n",
    "\n",
    "\n",
    "# 予測値(predicted)、実際値(actual)に分割\n",
    "predicted = [x[0] for x in l_estimate_scores]\n",
    "actual = [x[1] for x in l_estimate_scores]\n",
    "\n",
    "# --- 評価指標の計算 ---\n",
    "mse = mean_squared_error(actual, predicted)\n",
    "rmse = np.sqrt(mse)\n",
    "mae = mean_absolute_error(actual, predicted)\n",
    "r2 = r2_score(actual, predicted)\n",
    "\n",
    "print(\"MSE :\", mse)\n",
    "print(\"RMSE:\", rmse)\n",
    "print(\"MAE :\", mae)\n",
    "print(\"R^2 :\", r2)\n",
    "\n",
    "# --- 散布図 (Predicted vs Actual) ---\n",
    "plt.figure(figsize=(5, 5))\n",
    "plt.scatter(actual, predicted, color='blue', label='Data Points')\n",
    "# y = x の目安線\n",
    "plt.plot([0, 1], [0, 1], 'r--', label='Ideal line (y=x)')\n",
    "plt.xlabel('Actual')\n",
    "plt.ylabel('Predicted')\n",
    "plt.title('Predicted vs Actual')\n",
    "plt.legend()\n",
    "plt.show()\n",
    "\n",
    "# --- 残差プロット (Residual plot) ---\n",
    "residuals = [p - a for p, a in zip(predicted, actual)]\n",
    "\n",
    "plt.figure(figsize=(5, 5))\n",
    "plt.scatter(actual, residuals, color='green')\n",
    "plt.axhline(0, color='red', linestyle='--')  # 残差が0となるライン\n",
    "plt.xlabel('Actual')\n",
    "plt.ylabel('Residual (Predicted - Actual)')\n",
    "plt.title('Residual Plot')\n",
    "plt.show()\n",
    "\n",
    "# --- サンプルごとのバー比較 ---\n",
    "indices = range(len(actual))\n",
    "bar_width = 0.4\n",
    "\n",
    "plt.figure(figsize=(8, 5))\n",
    "plt.bar(indices, actual, width=bar_width, label='Actual', alpha=0.7)\n",
    "plt.bar([i + bar_width for i in indices], predicted, width=bar_width, label='Predicted', alpha=0.7)\n",
    "\n",
    "plt.xlabel('Sample Index')\n",
    "plt.ylabel('Score')\n",
    "plt.title('Actual vs Predicted')\n",
    "plt.xticks([i + bar_width/2 for i in indices], indices)  # 棒の中央にインデックスを合わせる\n",
    "plt.legend()\n",
    "plt.tight_layout()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "vllmtest",
   "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": "3.12.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}