{ "cells": [ { "cell_type": "code", "execution_count": 35, "id": "29893746-48a4-4439-ac69-a1514c048653", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T15:43:14.698999Z", "iopub.status.busy": "2024-05-23T15:43:14.698643Z", "iopub.status.idle": "2024-05-23T15:43:23.296300Z", "shell.execute_reply": "2024-05-23T15:43:23.295249Z", "shell.execute_reply.started": "2024-05-23T15:43:14.698975Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Requirement already satisfied: pynvml in /usr/local/lib/python3.11/dist-packages (11.5.0)\n", "Requirement already satisfied: numba in /usr/local/lib/python3.11/dist-packages (0.59.1)\n", "Requirement already satisfied: llvmlite<0.43,>=0.42.0dev0 in /usr/local/lib/python3.11/dist-packages (from numba) (0.42.0)\n", "Requirement already satisfied: numpy<1.27,>=1.22 in /usr/local/lib/python3.11/dist-packages (from numba) (1.26.3)\n", "\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n", "\u001b[0mCollecting evaluate\n", " Downloading evaluate-0.4.2-py3-none-any.whl.metadata (9.3 kB)\n", "Requirement already satisfied: datasets>=2.0.0 in /usr/local/lib/python3.11/dist-packages (from evaluate) (2.14.5)\n", "Requirement already satisfied: numpy>=1.17 in /usr/local/lib/python3.11/dist-packages (from evaluate) (1.26.3)\n", "Requirement already satisfied: dill in /usr/local/lib/python3.11/dist-packages (from evaluate) (0.3.7)\n", "Requirement already satisfied: pandas in /usr/local/lib/python3.11/dist-packages (from evaluate) (2.2.0)\n", "Requirement already satisfied: requests>=2.19.0 in /usr/local/lib/python3.11/dist-packages (from evaluate) (2.31.0)\n", "Requirement already satisfied: tqdm>=4.62.1 in /usr/local/lib/python3.11/dist-packages (from evaluate) (4.66.1)\n", "Requirement already satisfied: xxhash in /usr/local/lib/python3.11/dist-packages (from evaluate) (3.4.1)\n", "Requirement already satisfied: multiprocess in /usr/local/lib/python3.11/dist-packages (from evaluate) (0.70.15)\n", "Requirement already satisfied: fsspec>=2021.05.0 in /usr/local/lib/python3.11/dist-packages (from fsspec[http]>=2021.05.0->evaluate) (2023.6.0)\n", "Requirement already satisfied: huggingface-hub>=0.7.0 in /usr/local/lib/python3.11/dist-packages (from evaluate) (0.20.3)\n", "Requirement already satisfied: packaging in /usr/local/lib/python3.11/dist-packages (from evaluate) (23.2)\n", "Requirement already satisfied: pyarrow>=8.0.0 in /usr/local/lib/python3.11/dist-packages (from datasets>=2.0.0->evaluate) (15.0.0)\n", "Requirement already satisfied: aiohttp in /usr/local/lib/python3.11/dist-packages (from datasets>=2.0.0->evaluate) (3.9.1)\n", "Requirement already satisfied: pyyaml>=5.1 in /usr/lib/python3/dist-packages (from datasets>=2.0.0->evaluate) (5.4.1)\n", "Requirement already satisfied: filelock in /usr/local/lib/python3.11/dist-packages (from huggingface-hub>=0.7.0->evaluate) (3.13.1)\n", "Requirement already satisfied: typing-extensions>=3.7.4.3 in /usr/local/lib/python3.11/dist-packages (from huggingface-hub>=0.7.0->evaluate) (4.9.0)\n", "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.11/dist-packages (from requests>=2.19.0->evaluate) (3.3.2)\n", "Requirement already satisfied: idna<4,>=2.5 in /usr/lib/python3/dist-packages (from requests>=2.19.0->evaluate) (3.3)\n", "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.11/dist-packages (from requests>=2.19.0->evaluate) (2.0.7)\n", "Requirement already satisfied: certifi>=2017.4.17 in /usr/lib/python3/dist-packages (from requests>=2.19.0->evaluate) (2020.6.20)\n", "Requirement already satisfied: python-dateutil>=2.8.2 in /usr/local/lib/python3.11/dist-packages (from pandas->evaluate) (2.8.2)\n", "Requirement already satisfied: pytz>=2020.1 in /usr/lib/python3/dist-packages (from pandas->evaluate) (2022.1)\n", "Requirement already satisfied: tzdata>=2022.7 in /usr/local/lib/python3.11/dist-packages (from pandas->evaluate) (2023.4)\n", "Requirement already satisfied: attrs>=17.3.0 in /usr/local/lib/python3.11/dist-packages (from aiohttp->datasets>=2.0.0->evaluate) (23.1.0)\n", "Requirement already satisfied: multidict<7.0,>=4.5 in /usr/local/lib/python3.11/dist-packages (from aiohttp->datasets>=2.0.0->evaluate) (6.0.4)\n", "Requirement already satisfied: yarl<2.0,>=1.0 in /usr/local/lib/python3.11/dist-packages (from aiohttp->datasets>=2.0.0->evaluate) (1.9.4)\n", "Requirement already satisfied: frozenlist>=1.1.1 in /usr/local/lib/python3.11/dist-packages (from aiohttp->datasets>=2.0.0->evaluate) (1.4.1)\n", "Requirement already satisfied: aiosignal>=1.1.2 in /usr/local/lib/python3.11/dist-packages (from aiohttp->datasets>=2.0.0->evaluate) (1.3.1)\n", "Requirement already satisfied: six>=1.5 in /usr/lib/python3/dist-packages (from python-dateutil>=2.8.2->pandas->evaluate) (1.16.0)\n", "Downloading evaluate-0.4.2-py3-none-any.whl (84 kB)\n", "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m84.1/84.1 kB\u001b[0m \u001b[31m1.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0ma \u001b[36m0:00:01\u001b[0m\n", "\u001b[?25hInstalling collected packages: evaluate\n", "Successfully installed evaluate-0.4.2\n", "\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n", "\u001b[0m" ] } ], "source": [ "!pip install pynvml numba\n", "!pip install evaluate" ] }, { "cell_type": "code", "execution_count": 2, "id": "54e5daed-d4ba-4609-bc59-49b799947d95", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T15:16:55.852893Z", "iopub.status.busy": "2024-05-23T15:16:55.851997Z", "iopub.status.idle": "2024-05-23T15:16:55.858275Z", "shell.execute_reply": "2024-05-23T15:16:55.856882Z", "shell.execute_reply.started": "2024-05-23T15:16:55.852865Z" } }, "outputs": [], "source": [ "import random\n", "from collections import Counter\n", "from tqdm import tqdm\n", "\n", "import torch\n", "import matplotlib.pyplot as plt\n", "import numpy as np\n", "import pandas as pd\n", "import seaborn as sns" ] }, { "cell_type": "code", "execution_count": 3, "id": "4bf4396d-8c76-4025-b722-fa981cb4ce09", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T15:17:24.664545Z", "iopub.status.busy": "2024-05-23T15:17:24.663873Z", "iopub.status.idle": "2024-05-23T15:17:25.701955Z", "shell.execute_reply": "2024-05-23T15:17:25.701129Z", "shell.execute_reply.started": "2024-05-23T15:17:24.664519Z" } }, "outputs": [], "source": [ "import torch \n", "import torch.nn as nn\n", "from transformers import EsmTokenizer, EsmForSequenceClassification" ] }, { "cell_type": "code", "execution_count": 4, "id": "40d5d019-676b-4484-bac7-bc9f3bf8e74a", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T15:17:31.921016Z", "iopub.status.busy": "2024-05-23T15:17:31.919848Z", "iopub.status.idle": "2024-05-23T15:17:31.971089Z", "shell.execute_reply": "2024-05-23T15:17:31.970068Z", "shell.execute_reply.started": "2024-05-23T15:17:31.920952Z" } }, "outputs": [], "source": [ "device = \"cuda:0\" if torch.cuda.is_available() else \"cpu\"" ] }, { "cell_type": "code", "execution_count": 5, "id": "b11908f4-1374-46a8-992f-cf396183cf3a", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T15:17:38.440585Z", "iopub.status.busy": "2024-05-23T15:17:38.439573Z", "iopub.status.idle": "2024-05-23T15:17:38.450735Z", "shell.execute_reply": "2024-05-23T15:17:38.449625Z", "shell.execute_reply.started": "2024-05-23T15:17:38.440537Z" } }, "outputs": [ { "data": { "text/plain": [ "'cuda:0'" ] }, "execution_count": 5, "metadata": {}, "output_type": "execute_result" } ], "source": [ "device" ] }, { "cell_type": "code", "execution_count": 6, "id": "1d11319a-7704-4484-ab46-33ba2e0e1b9b", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T15:18:00.554601Z", "iopub.status.busy": "2024-05-23T15:18:00.554018Z", "iopub.status.idle": "2024-05-23T15:18:02.702640Z", "shell.execute_reply": "2024-05-23T15:18:02.701955Z", "shell.execute_reply.started": "2024-05-23T15:18:00.554575Z" } }, "outputs": [ { "data": { "text/plain": [ "(True, 1, 0, , 'Quadro P6000')" ] }, "execution_count": 6, "metadata": {}, "output_type": "execute_result" } ], "source": [ "import torch\n", "\n", "torch.cuda.is_available(), torch.cuda.device_count(), torch.cuda.current_device(), torch.cuda.device(0), torch.cuda.get_device_name(0)" ] }, { "cell_type": "code", "execution_count": 9, "id": "6690d2ff-cce2-4c1f-af2f-3a37dc18c983", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T15:19:07.447117Z", "iopub.status.busy": "2024-05-23T15:19:07.446269Z", "iopub.status.idle": "2024-05-23T15:19:10.121042Z", "shell.execute_reply": "2024-05-23T15:19:10.119780Z", "shell.execute_reply.started": "2024-05-23T15:19:07.447068Z" } }, "outputs": [], "source": [ "from pynvml import *\n", "\n", "\n", "def print_gpu_utilization():\n", " nvmlInit()\n", " handle = nvmlDeviceGetHandleByIndex(0)\n", " info = nvmlDeviceGetMemoryInfo(handle)\n", " print(f\"GPU memory occupied: {info.used//1024**2} MB.\")\n", "\n", "\n", "def print_summary(result):\n", " print(f\"Time: {result.metrics['train_runtime']:.2f}\")\n", " print(f\"Samples/second: {result.metrics['train_samples_per_second']:.2f}\")\n", " print_gpu_utilization()\n", "\n", "from numba import cuda \n", "device = cuda.get_current_device()\n", "device.reset()\n" ] }, { "cell_type": "code", "execution_count": 11, "id": "8681037d-256c-4aed-ba2b-4856beebb34d", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T15:19:32.734284Z", "iopub.status.busy": "2024-05-23T15:19:32.733624Z", "iopub.status.idle": "2024-05-23T15:19:32.739946Z", "shell.execute_reply": "2024-05-23T15:19:32.738443Z", "shell.execute_reply.started": "2024-05-23T15:19:32.734258Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "GPU memory occupied: 136 MB.\n" ] } ], "source": [ "print_gpu_utilization()" ] }, { "cell_type": "code", "execution_count": 12, "id": "e76a2e25-3a04-4af3-93da-c09d7f8de1a1", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T15:20:24.245833Z", "iopub.status.busy": "2024-05-23T15:20:24.244962Z", "iopub.status.idle": "2024-05-23T15:20:29.320493Z", "shell.execute_reply": "2024-05-23T15:20:29.319605Z", "shell.execute_reply.started": "2024-05-23T15:20:24.245791Z" } }, "outputs": [ { "data": { "application/vnd.jupyter.widget-view+json": { "model_id": "968a6df1a47b41409ab7bf11a3713ec1", "version_major": 2, "version_minor": 0 }, "text/plain": [ "tokenizer_config.json: 0%| | 0.00/40.0 [00:00" ] }, "execution_count": 29, "metadata": {}, "output_type": "execute_result" } ], "source": [ "test_loader" ] }, { "cell_type": "code", "execution_count": 30, "id": "cea06b94-3557-44e6-92b5-5f492a13d2b5", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T15:37:00.362739Z", "iopub.status.busy": "2024-05-23T15:37:00.362358Z", "iopub.status.idle": "2024-05-23T15:37:00.369008Z", "shell.execute_reply": "2024-05-23T15:37:00.368203Z", "shell.execute_reply.started": "2024-05-23T15:37:00.362713Z" } }, "outputs": [], "source": [ "from transformers import pipeline\n", "pipeline = pipeline(task=\"text-classification\", model=model, tokenizer=tokenizer, device=device)" ] }, { "cell_type": "code", "execution_count": 32, "id": "1f5cbdd2-70de-4ce1-bf7b-54fa7855f8a9", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T15:38:27.052751Z", "iopub.status.busy": "2024-05-23T15:38:27.052063Z", "iopub.status.idle": "2024-05-23T15:38:50.488290Z", "shell.execute_reply": "2024-05-23T15:38:50.487534Z", "shell.execute_reply.started": "2024-05-23T15:38:27.052725Z" } }, "outputs": [], "source": [ "predictions = pipeline(X_test)" ] }, { "cell_type": "code", "execution_count": 33, "id": "034bb9a9-d9f1-4beb-a8a3-3beb8117abb7", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T15:38:56.411699Z", "iopub.status.busy": "2024-05-23T15:38:56.411333Z", "iopub.status.idle": "2024-05-23T15:38:56.457662Z", "shell.execute_reply": "2024-05-23T15:38:56.456624Z", "shell.execute_reply.started": "2024-05-23T15:38:56.411665Z" } }, "outputs": [ { "data": { "text/plain": [ "[{'label': 'LABEL_1', 'score': 0.9920824766159058},\n", " {'label': 'LABEL_0', 'score': 0.6344404816627502},\n", " {'label': 'LABEL_1', 'score': 0.9963781237602234},\n", " {'label': 'LABEL_1', 'score': 0.8957202434539795},\n", " {'label': 'LABEL_0', 'score': 0.9899581074714661},\n", " {'label': 'LABEL_0', 'score': 0.992432713508606},\n", " {'label': 'LABEL_1', 'score': 0.9936670660972595},\n", " {'label': 'LABEL_0', 'score': 0.9957142472267151},\n", " {'label': 'LABEL_1', 'score': 0.7678724527359009},\n", " {'label': 'LABEL_0', 'score': 0.9950073957443237},\n", " {'label': 'LABEL_0', 'score': 0.9934017062187195},\n", " {'label': 'LABEL_0', 'score': 0.9901747703552246},\n", " {'label': 'LABEL_0', 'score': 0.9942784309387207},\n", " {'label': 'LABEL_1', 'score': 0.9712179899215698},\n", " {'label': 'LABEL_0', 'score': 0.9931582808494568},\n", " {'label': 'LABEL_0', 'score': 0.9989894032478333},\n", " {'label': 'LABEL_0', 'score': 0.9962522387504578},\n", " {'label': 'LABEL_0', 'score': 0.8317552208900452},\n", " {'label': 'LABEL_0', 'score': 0.980193555355072},\n", " {'label': 'LABEL_0', 'score': 0.7707980275154114},\n", " {'label': 'LABEL_0', 'score': 0.8872848153114319},\n", " {'label': 'LABEL_0', 'score': 0.9598217606544495},\n", " {'label': 'LABEL_1', 'score': 0.9657743573188782},\n", " {'label': 'LABEL_1', 'score': 0.9803487658500671},\n", " {'label': 'LABEL_1', 'score': 0.9939471483230591},\n", " {'label': 'LABEL_0', 'score': 0.6964303851127625},\n", " {'label': 'LABEL_1', 'score': 0.9915476441383362},\n", " {'label': 'LABEL_1', 'score': 0.993975818157196},\n", " {'label': 'LABEL_1', 'score': 0.9934781789779663},\n", " {'label': 'LABEL_0', 'score': 0.7910520434379578},\n", " {'label': 'LABEL_1', 'score': 0.9921974539756775},\n", " {'label': 'LABEL_1', 'score': 0.9815288782119751},\n", " {'label': 'LABEL_1', 'score': 0.9945481419563293},\n", " {'label': 'LABEL_1', 'score': 0.9962074756622314},\n", " {'label': 'LABEL_1', 'score': 0.8752002716064453},\n", " {'label': 'LABEL_1', 'score': 0.9936671853065491},\n", " {'label': 'LABEL_0', 'score': 0.9653744697570801},\n", " {'label': 'LABEL_0', 'score': 0.9754136204719543},\n", " {'label': 'LABEL_1', 'score': 0.9944813251495361},\n", " {'label': 'LABEL_1', 'score': 0.9775446653366089},\n", " {'label': 'LABEL_0', 'score': 0.7639273405075073},\n", " {'label': 'LABEL_0', 'score': 0.667346715927124},\n", " {'label': 'LABEL_0', 'score': 0.7142799496650696},\n", " {'label': 'LABEL_1', 'score': 0.9928463101387024},\n", " {'label': 'LABEL_0', 'score': 0.5917879343032837},\n", " {'label': 'LABEL_1', 'score': 0.890602707862854},\n", " {'label': 'LABEL_0', 'score': 0.9238371849060059},\n", " {'label': 'LABEL_1', 'score': 0.9933255910873413},\n", " {'label': 'LABEL_1', 'score': 0.9891610741615295},\n", " {'label': 'LABEL_0', 'score': 0.9632930159568787},\n", " {'label': 'LABEL_0', 'score': 0.9220459461212158},\n", " {'label': 'LABEL_0', 'score': 0.9919925928115845},\n", " {'label': 'LABEL_1', 'score': 0.5947464108467102},\n", " {'label': 'LABEL_0', 'score': 0.9746901988983154},\n", " {'label': 'LABEL_0', 'score': 0.9779336452484131},\n", " {'label': 'LABEL_1', 'score': 0.9941219687461853},\n", " {'label': 'LABEL_1', 'score': 0.8267089128494263},\n", " {'label': 'LABEL_0', 'score': 0.996107280254364},\n", " {'label': 'LABEL_1', 'score': 0.9836999773979187},\n", " {'label': 'LABEL_1', 'score': 0.9613105058670044},\n", " {'label': 'LABEL_0', 'score': 0.9625087380409241},\n", " {'label': 'LABEL_0', 'score': 0.9945961833000183},\n", " {'label': 'LABEL_1', 'score': 0.9716067910194397},\n", " {'label': 'LABEL_0', 'score': 0.9654806852340698},\n", " {'label': 'LABEL_0', 'score': 0.9887083768844604},\n", " {'label': 'LABEL_1', 'score': 0.9224319458007812},\n", " {'label': 'LABEL_1', 'score': 0.9960983991622925},\n", " {'label': 'LABEL_0', 'score': 0.9977193474769592},\n", " {'label': 'LABEL_1', 'score': 0.9953898191452026},\n", " {'label': 'LABEL_1', 'score': 0.9955762624740601},\n", " {'label': 'LABEL_0', 'score': 0.9048295021057129},\n", " {'label': 'LABEL_1', 'score': 0.9224010109901428},\n", " {'label': 'LABEL_1', 'score': 0.9839083552360535},\n", " {'label': 'LABEL_1', 'score': 0.7599542737007141},\n", " {'label': 'LABEL_1', 'score': 0.9917185306549072},\n", " {'label': 'LABEL_0', 'score': 0.8768678307533264},\n", " {'label': 'LABEL_1', 'score': 0.9816514253616333},\n", " {'label': 'LABEL_0', 'score': 0.5276628732681274},\n", " {'label': 'LABEL_1', 'score': 0.9766970872879028},\n", " {'label': 'LABEL_0', 'score': 0.9330600500106812},\n", " {'label': 'LABEL_0', 'score': 0.9660871624946594},\n", " {'label': 'LABEL_1', 'score': 0.9845715761184692},\n", " {'label': 'LABEL_1', 'score': 0.9950500130653381},\n", " {'label': 'LABEL_1', 'score': 0.996354341506958},\n", " {'label': 'LABEL_1', 'score': 0.9961453676223755},\n", " {'label': 'LABEL_0', 'score': 0.9704436659812927},\n", " {'label': 'LABEL_0', 'score': 0.8818399906158447},\n", " {'label': 'LABEL_1', 'score': 0.9842308163642883},\n", " {'label': 'LABEL_1', 'score': 0.6171972751617432},\n", " {'label': 'LABEL_0', 'score': 0.9949376583099365},\n", " {'label': 'LABEL_0', 'score': 0.9831938743591309},\n", " {'label': 'LABEL_0', 'score': 0.993530809879303},\n", " {'label': 'LABEL_1', 'score': 0.9809488654136658},\n", " {'label': 'LABEL_1', 'score': 0.9838874936103821},\n", " {'label': 'LABEL_0', 'score': 0.9927709698677063},\n", " {'label': 'LABEL_1', 'score': 0.672204315662384},\n", " {'label': 'LABEL_0', 'score': 0.9932838678359985},\n", " {'label': 'LABEL_1', 'score': 0.9767532348632812},\n", " {'label': 'LABEL_0', 'score': 0.9813488125801086},\n", " {'label': 'LABEL_1', 'score': 0.994283139705658},\n", " {'label': 'LABEL_1', 'score': 0.6814839243888855},\n", " {'label': 'LABEL_0', 'score': 0.9924652576446533},\n", " {'label': 'LABEL_1', 'score': 0.7334816455841064},\n", " {'label': 'LABEL_0', 'score': 0.9909833669662476},\n", " {'label': 'LABEL_0', 'score': 0.9856436252593994},\n", " {'label': 'LABEL_0', 'score': 0.9319406151771545},\n", " {'label': 'LABEL_0', 'score': 0.8773038983345032},\n", " {'label': 'LABEL_1', 'score': 0.9863520860671997},\n", " {'label': 'LABEL_1', 'score': 0.9160218834877014},\n", " {'label': 'LABEL_1', 'score': 0.9942693710327148},\n", " {'label': 'LABEL_1', 'score': 0.7459067106246948},\n", " {'label': 'LABEL_0', 'score': 0.9660363793373108},\n", " {'label': 'LABEL_0', 'score': 0.8596163988113403},\n", " {'label': 'LABEL_0', 'score': 0.9941113591194153},\n", " {'label': 'LABEL_0', 'score': 0.9466278553009033},\n", " {'label': 'LABEL_0', 'score': 0.873200535774231},\n", " {'label': 'LABEL_1', 'score': 0.9967073202133179},\n", " {'label': 'LABEL_1', 'score': 0.8866151571273804},\n", " {'label': 'LABEL_1', 'score': 0.9727884531021118},\n", " {'label': 'LABEL_0', 'score': 0.9961073994636536},\n", " {'label': 'LABEL_1', 'score': 0.9898661971092224},\n", " {'label': 'LABEL_1', 'score': 0.9957423806190491},\n", " {'label': 'LABEL_1', 'score': 0.9925336837768555},\n", " {'label': 'LABEL_1', 'score': 0.9566472172737122},\n", " {'label': 'LABEL_0', 'score': 0.9743568301200867},\n", " {'label': 'LABEL_1', 'score': 0.6052708029747009},\n", " {'label': 'LABEL_0', 'score': 0.963996946811676},\n", " {'label': 'LABEL_1', 'score': 0.9576812982559204},\n", " {'label': 'LABEL_1', 'score': 0.8900694251060486},\n", " {'label': 'LABEL_0', 'score': 0.9907097816467285},\n", " {'label': 'LABEL_1', 'score': 0.9793062210083008},\n", " {'label': 'LABEL_0', 'score': 0.9792136549949646},\n", " {'label': 'LABEL_0', 'score': 0.9821605682373047},\n", " {'label': 'LABEL_1', 'score': 0.990317702293396},\n", " {'label': 'LABEL_0', 'score': 0.9872243404388428},\n", " {'label': 'LABEL_1', 'score': 0.964670717716217},\n", " {'label': 'LABEL_0', 'score': 0.8030292391777039},\n", " {'label': 'LABEL_1', 'score': 0.6403388381004333},\n", " {'label': 'LABEL_0', 'score': 0.9047842621803284},\n", " {'label': 'LABEL_1', 'score': 0.9462425708770752},\n", " {'label': 'LABEL_0', 'score': 0.9106733798980713},\n", " {'label': 'LABEL_1', 'score': 0.9901905059814453},\n", " {'label': 'LABEL_1', 'score': 0.9941834807395935},\n", " {'label': 'LABEL_1', 'score': 0.9732745885848999},\n", " {'label': 'LABEL_0', 'score': 0.9958937168121338},\n", " {'label': 'LABEL_0', 'score': 0.6962946057319641},\n", " {'label': 'LABEL_1', 'score': 0.9949690699577332},\n", " {'label': 'LABEL_1', 'score': 0.9963836669921875},\n", " {'label': 'LABEL_1', 'score': 0.9950699806213379},\n", " {'label': 'LABEL_1', 'score': 0.9299766421318054},\n", " {'label': 'LABEL_1', 'score': 0.9953497648239136},\n", " {'label': 'LABEL_1', 'score': 0.9944887161254883},\n", " {'label': 'LABEL_1', 'score': 0.9949355721473694},\n", " {'label': 'LABEL_1', 'score': 0.9922817349433899},\n", " {'label': 'LABEL_0', 'score': 0.9634330868721008},\n", " {'label': 'LABEL_1', 'score': 0.9927870631217957},\n", " {'label': 'LABEL_1', 'score': 0.9881426692008972},\n", " {'label': 'LABEL_0', 'score': 0.9957276582717896},\n", " {'label': 'LABEL_0', 'score': 0.9653200507164001},\n", " {'label': 'LABEL_0', 'score': 0.9674235582351685},\n", " {'label': 'LABEL_0', 'score': 0.7343799471855164},\n", " {'label': 'LABEL_1', 'score': 0.992625892162323},\n", " {'label': 'LABEL_1', 'score': 0.9907028675079346},\n", " {'label': 'LABEL_0', 'score': 0.9948910474777222},\n", " {'label': 'LABEL_1', 'score': 0.7471431493759155},\n", " {'label': 'LABEL_1', 'score': 0.9757040739059448},\n", " {'label': 'LABEL_1', 'score': 0.9639447927474976},\n", " {'label': 'LABEL_1', 'score': 0.9619860649108887},\n", " {'label': 'LABEL_1', 'score': 0.9964457154273987},\n", " {'label': 'LABEL_1', 'score': 0.9969297051429749},\n", " {'label': 'LABEL_1', 'score': 0.9627381563186646},\n", " {'label': 'LABEL_1', 'score': 0.6666115522384644},\n", " {'label': 'LABEL_0', 'score': 0.9931562542915344},\n", " {'label': 'LABEL_1', 'score': 0.9928695559501648},\n", " {'label': 'LABEL_1', 'score': 0.9817350506782532},\n", " {'label': 'LABEL_1', 'score': 0.9804060459136963},\n", " {'label': 'LABEL_1', 'score': 0.9960583448410034},\n", " {'label': 'LABEL_1', 'score': 0.9965972304344177},\n", " {'label': 'LABEL_1', 'score': 0.9945698380470276},\n", " {'label': 'LABEL_1', 'score': 0.9966057538986206},\n", " {'label': 'LABEL_0', 'score': 0.9204991459846497},\n", " {'label': 'LABEL_1', 'score': 0.99128657579422},\n", " {'label': 'LABEL_0', 'score': 0.9790450930595398},\n", " {'label': 'LABEL_0', 'score': 0.9350587129592896},\n", " {'label': 'LABEL_0', 'score': 0.9830670952796936},\n", " {'label': 'LABEL_0', 'score': 0.9946940541267395},\n", " {'label': 'LABEL_0', 'score': 0.9659106135368347},\n", " {'label': 'LABEL_1', 'score': 0.9770876169204712},\n", " {'label': 'LABEL_0', 'score': 0.9846975803375244},\n", " {'label': 'LABEL_1', 'score': 0.5085996389389038},\n", " {'label': 'LABEL_1', 'score': 0.9591556191444397},\n", " {'label': 'LABEL_1', 'score': 0.6862562298774719},\n", " {'label': 'LABEL_0', 'score': 0.9033928513526917},\n", " {'label': 'LABEL_0', 'score': 0.9760305881500244},\n", " {'label': 'LABEL_1', 'score': 0.8040069937705994},\n", " {'label': 'LABEL_1', 'score': 0.9912001490592957},\n", " {'label': 'LABEL_0', 'score': 0.9618396759033203},\n", " {'label': 'LABEL_1', 'score': 0.6953979134559631},\n", " {'label': 'LABEL_1', 'score': 0.9941722750663757},\n", " {'label': 'LABEL_1', 'score': 0.9810013175010681},\n", " {'label': 'LABEL_0', 'score': 0.9971204996109009},\n", " {'label': 'LABEL_0', 'score': 0.770469605922699},\n", " {'label': 'LABEL_0', 'score': 0.9928150177001953},\n", " {'label': 'LABEL_1', 'score': 0.9852375388145447},\n", " {'label': 'LABEL_0', 'score': 0.9729052782058716},\n", " {'label': 'LABEL_1', 'score': 0.9801779389381409},\n", " {'label': 'LABEL_1', 'score': 0.9669545292854309},\n", " {'label': 'LABEL_0', 'score': 0.9891452193260193},\n", " {'label': 'LABEL_1', 'score': 0.9948595762252808},\n", " {'label': 'LABEL_1', 'score': 0.9631994366645813},\n", " {'label': 'LABEL_1', 'score': 0.9868159890174866},\n", " {'label': 'LABEL_0', 'score': 0.9939367175102234},\n", " {'label': 'LABEL_1', 'score': 0.9374150037765503},\n", " {'label': 'LABEL_0', 'score': 0.9290514588356018},\n", " {'label': 'LABEL_1', 'score': 0.9712733030319214},\n", " {'label': 'LABEL_1', 'score': 0.6352583169937134},\n", " {'label': 'LABEL_1', 'score': 0.9746260046958923},\n", " {'label': 'LABEL_1', 'score': 0.9918363690376282},\n", " {'label': 'LABEL_0', 'score': 0.6233887076377869},\n", " {'label': 'LABEL_0', 'score': 0.9778889417648315},\n", " {'label': 'LABEL_1', 'score': 0.6556466221809387},\n", " {'label': 'LABEL_1', 'score': 0.6855087280273438},\n", " {'label': 'LABEL_0', 'score': 0.9843446612358093},\n", " {'label': 'LABEL_0', 'score': 0.6279629468917847},\n", " {'label': 'LABEL_0', 'score': 0.9902299046516418},\n", " {'label': 'LABEL_1', 'score': 0.9677106142044067},\n", " {'label': 'LABEL_1', 'score': 0.95401930809021},\n", " {'label': 'LABEL_1', 'score': 0.8383669853210449},\n", " {'label': 'LABEL_0', 'score': 0.9565296173095703},\n", " {'label': 'LABEL_0', 'score': 0.9821329712867737},\n", " {'label': 'LABEL_1', 'score': 0.8867975473403931},\n", " {'label': 'LABEL_1', 'score': 0.9596521258354187},\n", " {'label': 'LABEL_1', 'score': 0.9953678846359253},\n", " {'label': 'LABEL_1', 'score': 0.6180222630500793},\n", " {'label': 'LABEL_1', 'score': 0.839966893196106},\n", " {'label': 'LABEL_1', 'score': 0.9966834187507629},\n", " {'label': 'LABEL_0', 'score': 0.988347589969635},\n", " {'label': 'LABEL_1', 'score': 0.9169568419456482},\n", " {'label': 'LABEL_0', 'score': 0.9921265840530396},\n", " {'label': 'LABEL_0', 'score': 0.8399364352226257},\n", " {'label': 'LABEL_0', 'score': 0.9207215905189514},\n", " {'label': 'LABEL_0', 'score': 0.9106295108795166},\n", " {'label': 'LABEL_0', 'score': 0.9921837449073792},\n", " {'label': 'LABEL_1', 'score': 0.9959458708763123},\n", " {'label': 'LABEL_0', 'score': 0.9894911050796509},\n", " {'label': 'LABEL_1', 'score': 0.9743772745132446},\n", " {'label': 'LABEL_0', 'score': 0.9118412733078003},\n", " {'label': 'LABEL_0', 'score': 0.9683268666267395},\n", " {'label': 'LABEL_0', 'score': 0.9877650141716003},\n", " {'label': 'LABEL_1', 'score': 0.8257834315299988},\n", " {'label': 'LABEL_1', 'score': 0.9933189153671265},\n", " {'label': 'LABEL_1', 'score': 0.9968757629394531},\n", " {'label': 'LABEL_1', 'score': 0.9589669108390808},\n", " {'label': 'LABEL_0', 'score': 0.9949743747711182},\n", " {'label': 'LABEL_0', 'score': 0.9604752063751221},\n", " {'label': 'LABEL_0', 'score': 0.6470544934272766},\n", " {'label': 'LABEL_0', 'score': 0.6516719460487366},\n", " {'label': 'LABEL_0', 'score': 0.8422739505767822},\n", " {'label': 'LABEL_0', 'score': 0.9935320615768433},\n", " {'label': 'LABEL_1', 'score': 0.9534354209899902},\n", " {'label': 'LABEL_1', 'score': 0.9863380789756775},\n", " {'label': 'LABEL_1', 'score': 0.9699581265449524},\n", " {'label': 'LABEL_1', 'score': 0.5838924050331116},\n", " {'label': 'LABEL_0', 'score': 0.9731244444847107},\n", " {'label': 'LABEL_0', 'score': 0.8859334588050842},\n", " {'label': 'LABEL_1', 'score': 0.9310123324394226},\n", " {'label': 'LABEL_0', 'score': 0.9858405590057373},\n", " {'label': 'LABEL_1', 'score': 0.8960093259811401},\n", " {'label': 'LABEL_0', 'score': 0.6809247136116028},\n", " {'label': 'LABEL_1', 'score': 0.9651084542274475},\n", " {'label': 'LABEL_1', 'score': 0.9708428978919983},\n", " {'label': 'LABEL_1', 'score': 0.8113129138946533},\n", " {'label': 'LABEL_1', 'score': 0.9679713249206543},\n", " {'label': 'LABEL_1', 'score': 0.9831263422966003},\n", " {'label': 'LABEL_1', 'score': 0.9862723350524902},\n", " {'label': 'LABEL_1', 'score': 0.9903726577758789},\n", " {'label': 'LABEL_0', 'score': 0.9874128103256226},\n", " {'label': 'LABEL_0', 'score': 0.9676929712295532},\n", " {'label': 'LABEL_0', 'score': 0.9208645224571228},\n", " {'label': 'LABEL_1', 'score': 0.7963953018188477},\n", " {'label': 'LABEL_1', 'score': 0.9790200591087341},\n", " {'label': 'LABEL_0', 'score': 0.9700548052787781},\n", " {'label': 'LABEL_1', 'score': 0.9285778403282166},\n", " {'label': 'LABEL_0', 'score': 0.9298763275146484},\n", " {'label': 'LABEL_0', 'score': 0.9788060784339905},\n", " {'label': 'LABEL_0', 'score': 0.5070981979370117},\n", " {'label': 'LABEL_0', 'score': 0.6101786494255066},\n", " {'label': 'LABEL_1', 'score': 0.9401609897613525},\n", " {'label': 'LABEL_0', 'score': 0.5705845952033997},\n", " {'label': 'LABEL_0', 'score': 0.715858519077301},\n", " {'label': 'LABEL_1', 'score': 0.9841576218605042},\n", " {'label': 'LABEL_0', 'score': 0.9133334755897522},\n", " {'label': 'LABEL_1', 'score': 0.8897599577903748},\n", " {'label': 'LABEL_0', 'score': 0.9941587448120117},\n", " {'label': 'LABEL_1', 'score': 0.9589704275131226},\n", " {'label': 'LABEL_1', 'score': 0.9959990978240967},\n", " {'label': 'LABEL_0', 'score': 0.9623426198959351},\n", " {'label': 'LABEL_1', 'score': 0.9778594374656677},\n", " {'label': 'LABEL_1', 'score': 0.9872311949729919},\n", " {'label': 'LABEL_1', 'score': 0.980732262134552},\n", " {'label': 'LABEL_1', 'score': 0.9541248083114624},\n", " {'label': 'LABEL_1', 'score': 0.9724259376525879},\n", " {'label': 'LABEL_1', 'score': 0.9902752637863159},\n", " {'label': 'LABEL_1', 'score': 0.9823142290115356},\n", " {'label': 'LABEL_1', 'score': 0.9673748016357422},\n", " {'label': 'LABEL_1', 'score': 0.9039739370346069},\n", " {'label': 'LABEL_0', 'score': 0.9813490509986877},\n", " {'label': 'LABEL_0', 'score': 0.9910780191421509},\n", " {'label': 'LABEL_1', 'score': 0.7190228700637817},\n", " {'label': 'LABEL_0', 'score': 0.955741822719574},\n", " {'label': 'LABEL_1', 'score': 0.9780182242393494},\n", " {'label': 'LABEL_0', 'score': 0.9955571293830872},\n", " {'label': 'LABEL_0', 'score': 0.9368971586227417},\n", " {'label': 'LABEL_0', 'score': 0.9864141345024109},\n", " {'label': 'LABEL_0', 'score': 0.9919256567955017},\n", " {'label': 'LABEL_1', 'score': 0.9556246399879456},\n", " {'label': 'LABEL_0', 'score': 0.9608051180839539},\n", " {'label': 'LABEL_0', 'score': 0.990741491317749},\n", " {'label': 'LABEL_0', 'score': 0.9546958804130554},\n", " {'label': 'LABEL_0', 'score': 0.9897760152816772},\n", " {'label': 'LABEL_0', 'score': 0.9305821657180786},\n", " {'label': 'LABEL_1', 'score': 0.9401752352714539},\n", " {'label': 'LABEL_0', 'score': 0.8841428756713867},\n", " {'label': 'LABEL_0', 'score': 0.9952380657196045},\n", " {'label': 'LABEL_0', 'score': 0.9911361932754517},\n", " {'label': 'LABEL_0', 'score': 0.9772043824195862},\n", " {'label': 'LABEL_0', 'score': 0.765044093132019},\n", " {'label': 'LABEL_0', 'score': 0.8548526763916016},\n", " {'label': 'LABEL_1', 'score': 0.8744843006134033},\n", " {'label': 'LABEL_0', 'score': 0.9837722182273865},\n", " {'label': 'LABEL_0', 'score': 0.9745046496391296},\n", " {'label': 'LABEL_0', 'score': 0.9850363731384277},\n", " {'label': 'LABEL_0', 'score': 0.9176458120346069},\n", " {'label': 'LABEL_0', 'score': 0.9215735197067261},\n", " {'label': 'LABEL_1', 'score': 0.6565162539482117},\n", " {'label': 'LABEL_0', 'score': 0.98956698179245},\n", " {'label': 'LABEL_0', 'score': 0.8922049403190613},\n", " {'label': 'LABEL_0', 'score': 0.9962350726127625},\n", " {'label': 'LABEL_0', 'score': 0.9958523511886597},\n", " {'label': 'LABEL_0', 'score': 0.976225733757019},\n", " {'label': 'LABEL_0', 'score': 0.9917077422142029},\n", " {'label': 'LABEL_0', 'score': 0.9804897904396057},\n", " {'label': 'LABEL_1', 'score': 0.8051390051841736},\n", " {'label': 'LABEL_1', 'score': 0.9766478538513184},\n", " {'label': 'LABEL_0', 'score': 0.9742063283920288},\n", " {'label': 'LABEL_0', 'score': 0.9911614656448364},\n", " {'label': 'LABEL_1', 'score': 0.6183955073356628},\n", " {'label': 'LABEL_0', 'score': 0.9954544305801392},\n", " {'label': 'LABEL_0', 'score': 0.9946867227554321},\n", " {'label': 'LABEL_0', 'score': 0.6935546398162842},\n", " {'label': 'LABEL_1', 'score': 0.8637236952781677},\n", " {'label': 'LABEL_1', 'score': 0.9793212413787842},\n", " {'label': 'LABEL_1', 'score': 0.9931596517562866},\n", " {'label': 'LABEL_0', 'score': 0.9968294501304626},\n", " {'label': 'LABEL_1', 'score': 0.9714514017105103},\n", " {'label': 'LABEL_0', 'score': 0.9883481860160828},\n", " {'label': 'LABEL_0', 'score': 0.8043408989906311},\n", " {'label': 'LABEL_0', 'score': 0.9928232431411743},\n", " {'label': 'LABEL_0', 'score': 0.9866151809692383},\n", " {'label': 'LABEL_0', 'score': 0.9845389723777771},\n", " {'label': 'LABEL_0', 'score': 0.8769729733467102},\n", " {'label': 'LABEL_0', 'score': 0.8691015243530273},\n", " {'label': 'LABEL_0', 'score': 0.8536194562911987},\n", " {'label': 'LABEL_0', 'score': 0.9304572343826294},\n", " {'label': 'LABEL_0', 'score': 0.9983558058738708},\n", " {'label': 'LABEL_0', 'score': 0.9976704716682434},\n", " {'label': 'LABEL_0', 'score': 0.9949118494987488},\n", " {'label': 'LABEL_1', 'score': 0.9864810705184937},\n", " {'label': 'LABEL_1', 'score': 0.6623121500015259},\n", " {'label': 'LABEL_0', 'score': 0.9275172352790833},\n", " {'label': 'LABEL_1', 'score': 0.9489859938621521},\n", " {'label': 'LABEL_1', 'score': 0.9946613907814026},\n", " {'label': 'LABEL_0', 'score': 0.9962383508682251},\n", " {'label': 'LABEL_0', 'score': 0.6611701250076294},\n", " {'label': 'LABEL_0', 'score': 0.9595711827278137},\n", " {'label': 'LABEL_0', 'score': 0.7789456844329834},\n", " {'label': 'LABEL_0', 'score': 0.98863285779953},\n", " {'label': 'LABEL_0', 'score': 0.8365599513053894},\n", " {'label': 'LABEL_0', 'score': 0.9642247557640076},\n", " {'label': 'LABEL_0', 'score': 0.9932066202163696},\n", " {'label': 'LABEL_0', 'score': 0.6040222644805908},\n", " {'label': 'LABEL_0', 'score': 0.9800452589988708},\n", " {'label': 'LABEL_1', 'score': 0.8107476234436035},\n", " {'label': 'LABEL_0', 'score': 0.9928464293479919},\n", " {'label': 'LABEL_0', 'score': 0.978429913520813},\n", " {'label': 'LABEL_0', 'score': 0.9875631332397461},\n", " {'label': 'LABEL_1', 'score': 0.9946447610855103},\n", " {'label': 'LABEL_0', 'score': 0.9569626450538635},\n", " {'label': 'LABEL_0', 'score': 0.5825687050819397},\n", " {'label': 'LABEL_1', 'score': 0.9850578904151917},\n", " {'label': 'LABEL_0', 'score': 0.9718114137649536},\n", " {'label': 'LABEL_1', 'score': 0.9784877896308899},\n", " {'label': 'LABEL_0', 'score': 0.9373846054077148},\n", " {'label': 'LABEL_1', 'score': 0.9798907041549683},\n", " {'label': 'LABEL_1', 'score': 0.8990122675895691},\n", " {'label': 'LABEL_1', 'score': 0.966254472732544},\n", " {'label': 'LABEL_0', 'score': 0.9869864583015442},\n", " {'label': 'LABEL_0', 'score': 0.9455236196517944},\n", " {'label': 'LABEL_1', 'score': 0.9763216972351074},\n", " {'label': 'LABEL_0', 'score': 0.9844340682029724},\n", " {'label': 'LABEL_0', 'score': 0.9957629442214966},\n", " {'label': 'LABEL_0', 'score': 0.9980828762054443},\n", " {'label': 'LABEL_0', 'score': 0.9765481352806091},\n", " {'label': 'LABEL_0', 'score': 0.9793199896812439},\n", " {'label': 'LABEL_0', 'score': 0.9791136384010315},\n", " {'label': 'LABEL_0', 'score': 0.9887147545814514},\n", " {'label': 'LABEL_1', 'score': 0.9902543425559998},\n", " {'label': 'LABEL_0', 'score': 0.9907160997390747},\n", " {'label': 'LABEL_0', 'score': 0.9892561435699463},\n", " {'label': 'LABEL_0', 'score': 0.99406498670578},\n", " {'label': 'LABEL_0', 'score': 0.9581699371337891},\n", " {'label': 'LABEL_1', 'score': 0.5796918869018555},\n", " {'label': 'LABEL_0', 'score': 0.9482484459877014},\n", " {'label': 'LABEL_1', 'score': 0.9893319606781006},\n", " {'label': 'LABEL_0', 'score': 0.8491203784942627},\n", " {'label': 'LABEL_0', 'score': 0.9950813055038452},\n", " {'label': 'LABEL_0', 'score': 0.9937944412231445},\n", " {'label': 'LABEL_0', 'score': 0.9934796094894409},\n", " {'label': 'LABEL_0', 'score': 0.9109570384025574},\n", " {'label': 'LABEL_1', 'score': 0.9654124975204468},\n", " {'label': 'LABEL_0', 'score': 0.9397651553153992},\n", " {'label': 'LABEL_1', 'score': 0.8101767897605896},\n", " {'label': 'LABEL_0', 'score': 0.9851425290107727},\n", " {'label': 'LABEL_0', 'score': 0.8464933633804321},\n", " {'label': 'LABEL_0', 'score': 0.7086097598075867},\n", " {'label': 'LABEL_1', 'score': 0.9942412376403809},\n", " {'label': 'LABEL_0', 'score': 0.9564889669418335},\n", " {'label': 'LABEL_0', 'score': 0.6702427268028259},\n", " {'label': 'LABEL_0', 'score': 0.9953863024711609},\n", " {'label': 'LABEL_0', 'score': 0.9961366057395935},\n", " {'label': 'LABEL_0', 'score': 0.9579523801803589},\n", " {'label': 'LABEL_0', 'score': 0.9929841756820679},\n", " {'label': 'LABEL_1', 'score': 0.8427727818489075},\n", " {'label': 'LABEL_0', 'score': 0.988649845123291},\n", " {'label': 'LABEL_0', 'score': 0.9902087450027466},\n", " {'label': 'LABEL_0', 'score': 0.992279052734375},\n", " {'label': 'LABEL_0', 'score': 0.9860447645187378},\n", " {'label': 'LABEL_0', 'score': 0.7342256307601929},\n", " {'label': 'LABEL_1', 'score': 0.627812922000885},\n", " {'label': 'LABEL_1', 'score': 0.9554335474967957},\n", " {'label': 'LABEL_0', 'score': 0.9830189347267151},\n", " {'label': 'LABEL_0', 'score': 0.861741304397583},\n", " {'label': 'LABEL_1', 'score': 0.9945268034934998},\n", " {'label': 'LABEL_0', 'score': 0.9706934094429016},\n", " {'label': 'LABEL_0', 'score': 0.9862181544303894},\n", " {'label': 'LABEL_1', 'score': 0.6970356702804565},\n", " {'label': 'LABEL_1', 'score': 0.9795743823051453},\n", " {'label': 'LABEL_1', 'score': 0.8964106440544128},\n", " {'label': 'LABEL_1', 'score': 0.990230143070221},\n", " {'label': 'LABEL_0', 'score': 0.9843607544898987},\n", " {'label': 'LABEL_1', 'score': 0.9737773537635803},\n", " {'label': 'LABEL_1', 'score': 0.9561145901679993},\n", " {'label': 'LABEL_1', 'score': 0.7726802229881287},\n", " {'label': 'LABEL_0', 'score': 0.9867532849311829},\n", " {'label': 'LABEL_1', 'score': 0.9936423301696777},\n", " {'label': 'LABEL_1', 'score': 0.8904270529747009},\n", " {'label': 'LABEL_0', 'score': 0.8102100491523743},\n", " {'label': 'LABEL_1', 'score': 0.7072275876998901},\n", " {'label': 'LABEL_1', 'score': 0.9506065845489502},\n", " {'label': 'LABEL_1', 'score': 0.6668100357055664},\n", " {'label': 'LABEL_0', 'score': 0.9742982983589172},\n", " {'label': 'LABEL_1', 'score': 0.8297302722930908},\n", " {'label': 'LABEL_1', 'score': 0.976436972618103},\n", " {'label': 'LABEL_0', 'score': 0.965576171875},\n", " {'label': 'LABEL_1', 'score': 0.6782581806182861},\n", " {'label': 'LABEL_0', 'score': 0.9925404191017151},\n", " {'label': 'LABEL_1', 'score': 0.552829921245575},\n", " {'label': 'LABEL_1', 'score': 0.9796808958053589},\n", " {'label': 'LABEL_0', 'score': 0.8962493538856506},\n", " {'label': 'LABEL_0', 'score': 0.9819111227989197},\n", " {'label': 'LABEL_1', 'score': 0.9811175465583801},\n", " {'label': 'LABEL_0', 'score': 0.9360445737838745},\n", " {'label': 'LABEL_1', 'score': 0.9926239252090454},\n", " {'label': 'LABEL_1', 'score': 0.9821844100952148},\n", " {'label': 'LABEL_0', 'score': 0.9439947009086609},\n", " {'label': 'LABEL_1', 'score': 0.9955853223800659},\n", " {'label': 'LABEL_1', 'score': 0.9958295226097107},\n", " {'label': 'LABEL_1', 'score': 0.6333702802658081},\n", " {'label': 'LABEL_0', 'score': 0.9083454012870789},\n", " {'label': 'LABEL_0', 'score': 0.8881974220275879},\n", " {'label': 'LABEL_1', 'score': 0.6101353168487549},\n", " {'label': 'LABEL_0', 'score': 0.957714855670929},\n", " {'label': 'LABEL_0', 'score': 0.9776718020439148},\n", " {'label': 'LABEL_0', 'score': 0.5999106168746948},\n", " {'label': 'LABEL_0', 'score': 0.989844560623169},\n", " {'label': 'LABEL_1', 'score': 0.9848566651344299},\n", " {'label': 'LABEL_1', 'score': 0.7018373012542725},\n", " {'label': 'LABEL_1', 'score': 0.9768227338790894},\n", " {'label': 'LABEL_1', 'score': 0.9928419589996338},\n", " {'label': 'LABEL_1', 'score': 0.9943158030509949},\n", " {'label': 'LABEL_0', 'score': 0.8971171379089355},\n", " {'label': 'LABEL_0', 'score': 0.9800693988800049},\n", " {'label': 'LABEL_1', 'score': 0.9963098168373108},\n", " {'label': 'LABEL_1', 'score': 0.9947733283042908},\n", " {'label': 'LABEL_0', 'score': 0.9846504926681519},\n", " {'label': 'LABEL_0', 'score': 0.9964327812194824},\n", " {'label': 'LABEL_0', 'score': 0.9576809406280518},\n", " {'label': 'LABEL_1', 'score': 0.9294152855873108},\n", " {'label': 'LABEL_0', 'score': 0.7247462272644043},\n", " {'label': 'LABEL_0', 'score': 0.9721958041191101},\n", " {'label': 'LABEL_1', 'score': 0.9864373803138733},\n", " {'label': 'LABEL_1', 'score': 0.9966200590133667},\n", " {'label': 'LABEL_1', 'score': 0.9962632060050964},\n", " {'label': 'LABEL_1', 'score': 0.5823416113853455},\n", " {'label': 'LABEL_1', 'score': 0.9945600628852844},\n", " {'label': 'LABEL_1', 'score': 0.9933189153671265},\n", " {'label': 'LABEL_1', 'score': 0.9757441282272339},\n", " {'label': 'LABEL_1', 'score': 0.9938511848449707},\n", " {'label': 'LABEL_1', 'score': 0.9950587749481201},\n", " {'label': 'LABEL_1', 'score': 0.9951475262641907},\n", " {'label': 'LABEL_1', 'score': 0.9965439438819885},\n", " {'label': 'LABEL_1', 'score': 0.756714940071106},\n", " {'label': 'LABEL_0', 'score': 0.9147056341171265},\n", " {'label': 'LABEL_1', 'score': 0.9834950566291809},\n", " {'label': 'LABEL_1', 'score': 0.9941120743751526},\n", " {'label': 'LABEL_1', 'score': 0.9966141581535339},\n", " {'label': 'LABEL_0', 'score': 0.9373542070388794},\n", " {'label': 'LABEL_1', 'score': 0.9953212141990662},\n", " {'label': 'LABEL_1', 'score': 0.9967867136001587},\n", " {'label': 'LABEL_1', 'score': 0.9939202070236206},\n", " {'label': 'LABEL_1', 'score': 0.7412394285202026},\n", " {'label': 'LABEL_0', 'score': 0.9623532295227051},\n", " {'label': 'LABEL_1', 'score': 0.9957146048545837},\n", " {'label': 'LABEL_1', 'score': 0.9939712882041931},\n", " {'label': 'LABEL_1', 'score': 0.9923322200775146},\n", " {'label': 'LABEL_1', 'score': 0.9660599827766418},\n", " {'label': 'LABEL_1', 'score': 0.9844948649406433},\n", " {'label': 'LABEL_1', 'score': 0.7453780174255371},\n", " {'label': 'LABEL_1', 'score': 0.9607815742492676},\n", " {'label': 'LABEL_1', 'score': 0.9240431785583496},\n", " {'label': 'LABEL_1', 'score': 0.9936596751213074},\n", " {'label': 'LABEL_1', 'score': 0.5215803384780884},\n", " {'label': 'LABEL_1', 'score': 0.9923531413078308},\n", " {'label': 'LABEL_1', 'score': 0.996246874332428},\n", " {'label': 'LABEL_1', 'score': 0.996537446975708},\n", " {'label': 'LABEL_1', 'score': 0.9740337133407593},\n", " {'label': 'LABEL_1', 'score': 0.9953761100769043},\n", " {'label': 'LABEL_1', 'score': 0.9968298077583313},\n", " {'label': 'LABEL_1', 'score': 0.9939036965370178},\n", " {'label': 'LABEL_1', 'score': 0.9916537404060364},\n", " {'label': 'LABEL_1', 'score': 0.9962269067764282},\n", " {'label': 'LABEL_1', 'score': 0.9958069324493408},\n", " {'label': 'LABEL_1', 'score': 0.9956639409065247},\n", " {'label': 'LABEL_0', 'score': 0.9886506795883179},\n", " {'label': 'LABEL_1', 'score': 0.9480882287025452},\n", " {'label': 'LABEL_1', 'score': 0.9821166396141052},\n", " {'label': 'LABEL_1', 'score': 0.9922047257423401},\n", " {'label': 'LABEL_1', 'score': 0.9872769713401794},\n", " {'label': 'LABEL_1', 'score': 0.991014301776886},\n", " {'label': 'LABEL_1', 'score': 0.9588004946708679},\n", " {'label': 'LABEL_0', 'score': 0.782096266746521},\n", " {'label': 'LABEL_1', 'score': 0.9705247282981873},\n", " {'label': 'LABEL_0', 'score': 0.9984493255615234},\n", " {'label': 'LABEL_1', 'score': 0.987313449382782},\n", " {'label': 'LABEL_1', 'score': 0.9935043454170227},\n", " {'label': 'LABEL_1', 'score': 0.9941995143890381},\n", " {'label': 'LABEL_1', 'score': 0.9954388737678528},\n", " {'label': 'LABEL_1', 'score': 0.9934248924255371},\n", " {'label': 'LABEL_1', 'score': 0.5153225660324097},\n", " {'label': 'LABEL_1', 'score': 0.9926658272743225},\n", " {'label': 'LABEL_1', 'score': 0.9900650382041931},\n", " {'label': 'LABEL_1', 'score': 0.990449845790863},\n", " {'label': 'LABEL_1', 'score': 0.9940094947814941},\n", " {'label': 'LABEL_1', 'score': 0.9649341702461243},\n", " {'label': 'LABEL_0', 'score': 0.9854132533073425},\n", " {'label': 'LABEL_0', 'score': 0.6902562975883484},\n", " {'label': 'LABEL_0', 'score': 0.9953790903091431},\n", " {'label': 'LABEL_1', 'score': 0.9951647520065308},\n", " {'label': 'LABEL_1', 'score': 0.9959852695465088},\n", " {'label': 'LABEL_1', 'score': 0.979231595993042},\n", " {'label': 'LABEL_0', 'score': 0.5359320640563965},\n", " {'label': 'LABEL_0', 'score': 0.9935789108276367},\n", " {'label': 'LABEL_1', 'score': 0.9128223061561584},\n", " {'label': 'LABEL_1', 'score': 0.9885913729667664},\n", " {'label': 'LABEL_0', 'score': 0.9898375272750854},\n", " {'label': 'LABEL_0', 'score': 0.9732145071029663},\n", " {'label': 'LABEL_1', 'score': 0.9945152401924133},\n", " {'label': 'LABEL_0', 'score': 0.9821017980575562},\n", " {'label': 'LABEL_0', 'score': 0.5703974962234497},\n", " {'label': 'LABEL_0', 'score': 0.9287216663360596},\n", " {'label': 'LABEL_1', 'score': 0.9972570538520813},\n", " {'label': 'LABEL_0', 'score': 0.8542720675468445},\n", " {'label': 'LABEL_1', 'score': 0.995238184928894},\n", " {'label': 'LABEL_1', 'score': 0.992724597454071},\n", " {'label': 'LABEL_1', 'score': 0.9938576817512512},\n", " {'label': 'LABEL_1', 'score': 0.9965469241142273},\n", " {'label': 'LABEL_0', 'score': 0.9916887879371643},\n", " {'label': 'LABEL_1', 'score': 0.9956098198890686},\n", " {'label': 'LABEL_1', 'score': 0.9948581457138062},\n", " {'label': 'LABEL_1', 'score': 0.9800918102264404},\n", " {'label': 'LABEL_0', 'score': 0.9850999712944031},\n", " {'label': 'LABEL_1', 'score': 0.9944234490394592},\n", " {'label': 'LABEL_1', 'score': 0.9922235608100891},\n", " {'label': 'LABEL_1', 'score': 0.9933009147644043},\n", " {'label': 'LABEL_1', 'score': 0.9839582443237305},\n", " {'label': 'LABEL_0', 'score': 0.8630751371383667},\n", " {'label': 'LABEL_0', 'score': 0.9115880131721497},\n", " {'label': 'LABEL_0', 'score': 0.9788351058959961},\n", " {'label': 'LABEL_1', 'score': 0.9253813624382019},\n", " {'label': 'LABEL_0', 'score': 0.9877164959907532},\n", " {'label': 'LABEL_1', 'score': 0.9453893303871155},\n", " {'label': 'LABEL_1', 'score': 0.946543276309967},\n", " {'label': 'LABEL_0', 'score': 0.8820732235908508},\n", " {'label': 'LABEL_1', 'score': 0.9947957396507263},\n", " {'label': 'LABEL_1', 'score': 0.9912304878234863},\n", " {'label': 'LABEL_1', 'score': 0.9177975654602051},\n", " {'label': 'LABEL_0', 'score': 0.9687201380729675},\n", " {'label': 'LABEL_1', 'score': 0.9947293400764465},\n", " {'label': 'LABEL_1', 'score': 0.970740556716919},\n", " {'label': 'LABEL_0', 'score': 0.8022850155830383},\n", " {'label': 'LABEL_0', 'score': 0.9579288959503174},\n", " {'label': 'LABEL_0', 'score': 0.9937421679496765},\n", " {'label': 'LABEL_1', 'score': 0.994292140007019},\n", " {'label': 'LABEL_1', 'score': 0.9968717694282532},\n", " {'label': 'LABEL_0', 'score': 0.9592112898826599},\n", " {'label': 'LABEL_0', 'score': 0.6592299342155457},\n", " {'label': 'LABEL_1', 'score': 0.9956451654434204},\n", " {'label': 'LABEL_1', 'score': 0.9911348819732666},\n", " {'label': 'LABEL_1', 'score': 0.9951386451721191},\n", " {'label': 'LABEL_0', 'score': 0.5056076049804688},\n", " {'label': 'LABEL_1', 'score': 0.8203694224357605},\n", " {'label': 'LABEL_1', 'score': 0.9595959186553955},\n", " {'label': 'LABEL_1', 'score': 0.9836644530296326},\n", " {'label': 'LABEL_1', 'score': 0.9942445158958435},\n", " {'label': 'LABEL_0', 'score': 0.991622269153595},\n", " {'label': 'LABEL_1', 'score': 0.9954179525375366},\n", " {'label': 'LABEL_0', 'score': 0.52507483959198},\n", " {'label': 'LABEL_0', 'score': 0.9515404105186462},\n", " {'label': 'LABEL_0', 'score': 0.9718201756477356},\n", " {'label': 'LABEL_0', 'score': 0.6818857192993164},\n", " {'label': 'LABEL_0', 'score': 0.9904976487159729},\n", " {'label': 'LABEL_1', 'score': 0.9965739250183105},\n", " {'label': 'LABEL_1', 'score': 0.6346263289451599},\n", " {'label': 'LABEL_1', 'score': 0.9970904588699341},\n", " {'label': 'LABEL_1', 'score': 0.9544928073883057},\n", " {'label': 'LABEL_1', 'score': 0.9378147721290588},\n", " {'label': 'LABEL_1', 'score': 0.9889683723449707},\n", " {'label': 'LABEL_1', 'score': 0.9771023988723755},\n", " {'label': 'LABEL_1', 'score': 0.8697611689567566},\n", " {'label': 'LABEL_1', 'score': 0.9960363507270813},\n", " {'label': 'LABEL_1', 'score': 0.9348815679550171},\n", " {'label': 'LABEL_1', 'score': 0.9703534841537476},\n", " {'label': 'LABEL_0', 'score': 0.9966253042221069},\n", " {'label': 'LABEL_1', 'score': 0.8299302458763123},\n", " {'label': 'LABEL_0', 'score': 0.9111083149909973},\n", " {'label': 'LABEL_1', 'score': 0.8398126363754272},\n", " {'label': 'LABEL_1', 'score': 0.9752505421638489},\n", " {'label': 'LABEL_1', 'score': 0.9928673505783081},\n", " {'label': 'LABEL_1', 'score': 0.9740899205207825},\n", " {'label': 'LABEL_1', 'score': 0.9677812457084656},\n", " {'label': 'LABEL_1', 'score': 0.96604984998703},\n", " {'label': 'LABEL_1', 'score': 0.9920864701271057},\n", " {'label': 'LABEL_1', 'score': 0.9934585690498352},\n", " {'label': 'LABEL_1', 'score': 0.991157054901123},\n", " {'label': 'LABEL_0', 'score': 0.9470826983451843},\n", " {'label': 'LABEL_1', 'score': 0.9873031973838806},\n", " {'label': 'LABEL_1', 'score': 0.9779534339904785},\n", " {'label': 'LABEL_1', 'score': 0.9912458062171936},\n", " {'label': 'LABEL_1', 'score': 0.9942110776901245},\n", " {'label': 'LABEL_1', 'score': 0.993436336517334},\n", " {'label': 'LABEL_1', 'score': 0.9949154853820801},\n", " {'label': 'LABEL_1', 'score': 0.9935457110404968},\n", " {'label': 'LABEL_1', 'score': 0.9942652583122253},\n", " {'label': 'LABEL_1', 'score': 0.9925038814544678},\n", " {'label': 'LABEL_1', 'score': 0.9942684173583984},\n", " {'label': 'LABEL_1', 'score': 0.8534121513366699},\n", " {'label': 'LABEL_1', 'score': 0.9949132204055786},\n", " {'label': 'LABEL_0', 'score': 0.9427997469902039},\n", " {'label': 'LABEL_1', 'score': 0.9734408259391785},\n", " {'label': 'LABEL_1', 'score': 0.9921699166297913},\n", " {'label': 'LABEL_1', 'score': 0.9898235201835632},\n", " {'label': 'LABEL_1', 'score': 0.7765575647354126},\n", " {'label': 'LABEL_1', 'score': 0.9793251156806946},\n", " {'label': 'LABEL_1', 'score': 0.8394061326980591},\n", " {'label': 'LABEL_1', 'score': 0.9897447228431702},\n", " {'label': 'LABEL_0', 'score': 0.9154225587844849},\n", " {'label': 'LABEL_0', 'score': 0.9888221025466919},\n", " {'label': 'LABEL_1', 'score': 0.9890047907829285},\n", " {'label': 'LABEL_1', 'score': 0.9717256426811218},\n", " {'label': 'LABEL_1', 'score': 0.9893767833709717},\n", " {'label': 'LABEL_1', 'score': 0.9924948215484619},\n", " {'label': 'LABEL_1', 'score': 0.9964051246643066},\n", " {'label': 'LABEL_1', 'score': 0.9560785293579102},\n", " {'label': 'LABEL_0', 'score': 0.7070202231407166},\n", " {'label': 'LABEL_1', 'score': 0.9044116735458374},\n", " {'label': 'LABEL_1', 'score': 0.9941132664680481},\n", " {'label': 'LABEL_1', 'score': 0.9947010278701782},\n", " {'label': 'LABEL_1', 'score': 0.9965517520904541},\n", " {'label': 'LABEL_1', 'score': 0.9773184657096863},\n", " {'label': 'LABEL_1', 'score': 0.9909886717796326},\n", " {'label': 'LABEL_1', 'score': 0.9947931170463562},\n", " {'label': 'LABEL_1', 'score': 0.9928774237632751},\n", " {'label': 'LABEL_1', 'score': 0.9927247762680054},\n", " {'label': 'LABEL_0', 'score': 0.9351725578308105},\n", " {'label': 'LABEL_1', 'score': 0.9962174296379089},\n", " {'label': 'LABEL_1', 'score': 0.9845442175865173},\n", " {'label': 'LABEL_0', 'score': 0.9229084849357605},\n", " {'label': 'LABEL_1', 'score': 0.9958617687225342},\n", " {'label': 'LABEL_0', 'score': 0.988365888595581},\n", " {'label': 'LABEL_0', 'score': 0.9961277842521667},\n", " {'label': 'LABEL_0', 'score': 0.995089590549469},\n", " {'label': 'LABEL_1', 'score': 0.9151782393455505},\n", " {'label': 'LABEL_1', 'score': 0.7796429991722107},\n", " {'label': 'LABEL_0', 'score': 0.6353229880332947},\n", " {'label': 'LABEL_1', 'score': 0.5075734853744507},\n", " {'label': 'LABEL_1', 'score': 0.9626388549804688},\n", " {'label': 'LABEL_1', 'score': 0.991780698299408},\n", " {'label': 'LABEL_1', 'score': 0.9936391711235046},\n", " {'label': 'LABEL_1', 'score': 0.9885025024414062},\n", " {'label': 'LABEL_1', 'score': 0.9949436783790588},\n", " {'label': 'LABEL_1', 'score': 0.9955767393112183},\n", " {'label': 'LABEL_1', 'score': 0.9393144845962524},\n", " {'label': 'LABEL_0', 'score': 0.8385283946990967},\n", " {'label': 'LABEL_0', 'score': 0.6774574518203735},\n", " {'label': 'LABEL_1', 'score': 0.9914259910583496},\n", " {'label': 'LABEL_1', 'score': 0.9934968948364258},\n", " {'label': 'LABEL_1', 'score': 0.8619126677513123},\n", " {'label': 'LABEL_1', 'score': 0.9947986602783203},\n", " {'label': 'LABEL_1', 'score': 0.9928877949714661},\n", " {'label': 'LABEL_1', 'score': 0.9960314631462097},\n", " {'label': 'LABEL_0', 'score': 0.8468936681747437},\n", " {'label': 'LABEL_1', 'score': 0.9907916784286499},\n", " {'label': 'LABEL_1', 'score': 0.692417562007904},\n", " {'label': 'LABEL_1', 'score': 0.9829657673835754},\n", " {'label': 'LABEL_0', 'score': 0.9849721789360046},\n", " {'label': 'LABEL_0', 'score': 0.9317151308059692},\n", " {'label': 'LABEL_1', 'score': 0.6366862654685974},\n", " {'label': 'LABEL_1', 'score': 0.9936919212341309},\n", " {'label': 'LABEL_1', 'score': 0.9850026965141296},\n", " {'label': 'LABEL_0', 'score': 0.933652400970459},\n", " {'label': 'LABEL_1', 'score': 0.9690437912940979},\n", " {'label': 'LABEL_0', 'score': 0.936476469039917},\n", " {'label': 'LABEL_1', 'score': 0.9966297745704651},\n", " {'label': 'LABEL_1', 'score': 0.9681738615036011},\n", " {'label': 'LABEL_0', 'score': 0.9943897128105164},\n", " {'label': 'LABEL_0', 'score': 0.8239411115646362},\n", " {'label': 'LABEL_0', 'score': 0.998040497303009},\n", " {'label': 'LABEL_1', 'score': 0.9954612851142883},\n", " {'label': 'LABEL_1', 'score': 0.9367863535881042},\n", " {'label': 'LABEL_1', 'score': 0.7423374056816101},\n", " {'label': 'LABEL_0', 'score': 0.5968354940414429},\n", " {'label': 'LABEL_0', 'score': 0.9948377013206482},\n", " {'label': 'LABEL_0', 'score': 0.8263901472091675},\n", " {'label': 'LABEL_0', 'score': 0.988389253616333},\n", " {'label': 'LABEL_1', 'score': 0.995772659778595},\n", " {'label': 'LABEL_0', 'score': 0.8763516545295715},\n", " {'label': 'LABEL_0', 'score': 0.9964577555656433},\n", " {'label': 'LABEL_1', 'score': 0.9791182279586792},\n", " {'label': 'LABEL_0', 'score': 0.9895098209381104},\n", " {'label': 'LABEL_0', 'score': 0.8862988948822021},\n", " {'label': 'LABEL_0', 'score': 0.9519324898719788},\n", " {'label': 'LABEL_1', 'score': 0.9921212196350098},\n", " {'label': 'LABEL_1', 'score': 0.6741978526115417},\n", " {'label': 'LABEL_0', 'score': 0.5454573631286621},\n", " {'label': 'LABEL_0', 'score': 0.5335167646408081},\n", " {'label': 'LABEL_1', 'score': 0.5119876265525818},\n", " {'label': 'LABEL_1', 'score': 0.632554829120636},\n", " {'label': 'LABEL_0', 'score': 0.9117580652236938},\n", " {'label': 'LABEL_0', 'score': 0.9932965636253357},\n", " {'label': 'LABEL_0', 'score': 0.9813398718833923},\n", " {'label': 'LABEL_0', 'score': 0.9965149164199829},\n", " {'label': 'LABEL_0', 'score': 0.9989803433418274},\n", " {'label': 'LABEL_1', 'score': 0.9901706576347351},\n", " {'label': 'LABEL_1', 'score': 0.9950527548789978},\n", " {'label': 'LABEL_1', 'score': 0.9674326181411743},\n", " {'label': 'LABEL_1', 'score': 0.9943349957466125},\n", " {'label': 'LABEL_0', 'score': 0.9982045888900757},\n", " {'label': 'LABEL_1', 'score': 0.9984146356582642},\n", " {'label': 'LABEL_0', 'score': 0.9807737469673157},\n", " {'label': 'LABEL_1', 'score': 0.998051643371582},\n", " {'label': 'LABEL_1', 'score': 0.9973867535591125},\n", " {'label': 'LABEL_1', 'score': 0.9974567294120789},\n", " {'label': 'LABEL_1', 'score': 0.9983263611793518},\n", " {'label': 'LABEL_1', 'score': 0.9982770681381226},\n", " {'label': 'LABEL_1', 'score': 0.9981163740158081},\n", " {'label': 'LABEL_1', 'score': 0.9136704802513123},\n", " {'label': 'LABEL_1', 'score': 0.8776158094406128},\n", " {'label': 'LABEL_1', 'score': 0.9977473616600037},\n", " {'label': 'LABEL_1', 'score': 0.9981617331504822},\n", " {'label': 'LABEL_1', 'score': 0.9980959296226501},\n", " {'label': 'LABEL_0', 'score': 0.5451762676239014},\n", " {'label': 'LABEL_1', 'score': 0.5463171601295471},\n", " {'label': 'LABEL_0', 'score': 0.9952754974365234},\n", " {'label': 'LABEL_1', 'score': 0.9979830980300903},\n", " {'label': 'LABEL_1', 'score': 0.9973657727241516},\n", " {'label': 'LABEL_1', 'score': 0.9979636669158936},\n", " {'label': 'LABEL_1', 'score': 0.9980011582374573},\n", " {'label': 'LABEL_0', 'score': 0.609296441078186},\n", " {'label': 'LABEL_1', 'score': 0.9980497360229492},\n", " {'label': 'LABEL_1', 'score': 0.9953547716140747},\n", " {'label': 'LABEL_1', 'score': 0.9984027743339539},\n", " {'label': 'LABEL_1', 'score': 0.9978079199790955},\n", " {'label': 'LABEL_1', 'score': 0.9980959296226501},\n", " {'label': 'LABEL_1', 'score': 0.9979496598243713},\n", " {'label': 'LABEL_1', 'score': 0.9981176853179932},\n", " {'label': 'LABEL_1', 'score': 0.9982970356941223},\n", " {'label': 'LABEL_1', 'score': 0.9970557689666748},\n", " {'label': 'LABEL_1', 'score': 0.9968823194503784},\n", " {'label': 'LABEL_1', 'score': 0.9980416297912598},\n", " {'label': 'LABEL_1', 'score': 0.9300060272216797},\n", " {'label': 'LABEL_1', 'score': 0.9972960352897644},\n", " {'label': 'LABEL_1', 'score': 0.9978334307670593},\n", " {'label': 'LABEL_1', 'score': 0.9977651834487915},\n", " {'label': 'LABEL_1', 'score': 0.9975778460502625},\n", " {'label': 'LABEL_1', 'score': 0.9978219270706177},\n", " {'label': 'LABEL_1', 'score': 0.9968376159667969},\n", " {'label': 'LABEL_1', 'score': 0.9980719089508057},\n", " {'label': 'LABEL_1', 'score': 0.997948944568634},\n", " {'label': 'LABEL_1', 'score': 0.997114896774292},\n", " {'label': 'LABEL_1', 'score': 0.9974072575569153},\n", " {'label': 'LABEL_1', 'score': 0.9976442456245422},\n", " {'label': 'LABEL_1', 'score': 0.9970853924751282},\n", " {'label': 'LABEL_1', 'score': 0.9979702830314636},\n", " {'label': 'LABEL_1', 'score': 0.9972410202026367},\n", " {'label': 'LABEL_1', 'score': 0.9980522394180298},\n", " {'label': 'LABEL_1', 'score': 0.9978965520858765},\n", " {'label': 'LABEL_1', 'score': 0.9981445074081421},\n", " {'label': 'LABEL_1', 'score': 0.9973828196525574},\n", " {'label': 'LABEL_1', 'score': 0.9975137710571289},\n", " {'label': 'LABEL_1', 'score': 0.9973871111869812},\n", " {'label': 'LABEL_1', 'score': 0.9975990653038025},\n", " {'label': 'LABEL_1', 'score': 0.9978379607200623},\n", " {'label': 'LABEL_1', 'score': 0.9977128505706787},\n", " {'label': 'LABEL_1', 'score': 0.9977173805236816},\n", " {'label': 'LABEL_1', 'score': 0.9973612427711487},\n", " {'label': 'LABEL_1', 'score': 0.9964808821678162},\n", " {'label': 'LABEL_1', 'score': 0.9978213310241699},\n", " {'label': 'LABEL_1', 'score': 0.9965618252754211},\n", " {'label': 'LABEL_1', 'score': 0.9972785115242004},\n", " {'label': 'LABEL_1', 'score': 0.9975953698158264},\n", " {'label': 'LABEL_1', 'score': 0.9975850582122803},\n", " {'label': 'LABEL_1', 'score': 0.9974740147590637},\n", " {'label': 'LABEL_1', 'score': 0.996859073638916},\n", " {'label': 'LABEL_1', 'score': 0.997346043586731},\n", " {'label': 'LABEL_1', 'score': 0.9979947805404663},\n", " {'label': 'LABEL_1', 'score': 0.9974393844604492},\n", " {'label': 'LABEL_1', 'score': 0.9974966645240784},\n", " {'label': 'LABEL_1', 'score': 0.9979947805404663},\n", " {'label': 'LABEL_1', 'score': 0.9983198046684265},\n", " {'label': 'LABEL_1', 'score': 0.9958517551422119},\n", " {'label': 'LABEL_1', 'score': 0.9918461441993713},\n", " {'label': 'LABEL_1', 'score': 0.9947928786277771},\n", " {'label': 'LABEL_1', 'score': 0.9898167848587036},\n", " {'label': 'LABEL_1', 'score': 0.9971463084220886},\n", " {'label': 'LABEL_1', 'score': 0.9970598816871643},\n", " {'label': 'LABEL_1', 'score': 0.997458279132843},\n", " {'label': 'LABEL_1', 'score': 0.9978926777839661},\n", " {'label': 'LABEL_1', 'score': 0.997327446937561},\n", " {'label': 'LABEL_1', 'score': 0.9979981780052185},\n", " {'label': 'LABEL_1', 'score': 0.9981091022491455},\n", " {'label': 'LABEL_1', 'score': 0.998021125793457},\n", " {'label': 'LABEL_1', 'score': 0.9979619979858398},\n", " {'label': 'LABEL_1', 'score': 0.9978579878807068},\n", " {'label': 'LABEL_1', 'score': 0.9977995753288269},\n", " {'label': 'LABEL_1', 'score': 0.9970647692680359},\n", " {'label': 'LABEL_1', 'score': 0.9957797527313232},\n", " {'label': 'LABEL_1', 'score': 0.9966141581535339},\n", " {'label': 'LABEL_1', 'score': 0.9957832098007202},\n", " {'label': 'LABEL_1', 'score': 0.9975501894950867},\n", " {'label': 'LABEL_1', 'score': 0.9969447255134583},\n", " {'label': 'LABEL_1', 'score': 0.9977447986602783},\n", " {'label': 'LABEL_1', 'score': 0.9958977103233337},\n", " {'label': 'LABEL_1', 'score': 0.9965307116508484},\n", " {'label': 'LABEL_1', 'score': 0.998040497303009},\n", " {'label': 'LABEL_1', 'score': 0.9974597096443176},\n", " {'label': 'LABEL_1', 'score': 0.9975069165229797},\n", " {'label': 'LABEL_1', 'score': 0.9968619346618652},\n", " {'label': 'LABEL_1', 'score': 0.7999141216278076},\n", " {'label': 'LABEL_1', 'score': 0.9966764450073242},\n", " {'label': 'LABEL_1', 'score': 0.9978440999984741},\n", " {'label': 'LABEL_1', 'score': 0.9975869655609131},\n", " {'label': 'LABEL_1', 'score': 0.9980499744415283},\n", " {'label': 'LABEL_1', 'score': 0.9977290034294128},\n", " {'label': 'LABEL_0', 'score': 0.9947840571403503},\n", " {'label': 'LABEL_1', 'score': 0.6376346349716187},\n", " {'label': 'LABEL_1', 'score': 0.9190598726272583},\n", " {'label': 'LABEL_0', 'score': 0.9695489406585693},\n", " {'label': 'LABEL_0', 'score': 0.9246589541435242},\n", " {'label': 'LABEL_1', 'score': 0.5756272673606873},\n", " {'label': 'LABEL_0', 'score': 0.8085498213768005},\n", " {'label': 'LABEL_0', 'score': 0.9974361062049866},\n", " {'label': 'LABEL_1', 'score': 0.7996165752410889},\n", " {'label': 'LABEL_0', 'score': 0.9922124743461609},\n", " {'label': 'LABEL_1', 'score': 0.9789844155311584},\n", " {'label': 'LABEL_0', 'score': 0.9910444021224976},\n", " {'label': 'LABEL_0', 'score': 0.9960230588912964},\n", " {'label': 'LABEL_0', 'score': 0.9831807017326355},\n", " {'label': 'LABEL_0', 'score': 0.9928321838378906},\n", " {'label': 'LABEL_0', 'score': 0.9855900406837463},\n", " {'label': 'LABEL_0', 'score': 0.9944981932640076},\n", " {'label': 'LABEL_0', 'score': 0.9949136972427368},\n", " {'label': 'LABEL_0', 'score': 0.9957302212715149},\n", " {'label': 'LABEL_0', 'score': 0.9968340992927551},\n", " {'label': 'LABEL_0', 'score': 0.9956672191619873},\n", " {'label': 'LABEL_1', 'score': 0.9852407574653625},\n", " {'label': 'LABEL_0', 'score': 0.9968200922012329},\n", " {'label': 'LABEL_0', 'score': 0.9808366298675537},\n", " {'label': 'LABEL_0', 'score': 0.9898589253425598},\n", " {'label': 'LABEL_0', 'score': 0.9969015121459961},\n", " {'label': 'LABEL_0', 'score': 0.9971805810928345},\n", " {'label': 'LABEL_0', 'score': 0.9982355833053589},\n", " {'label': 'LABEL_0', 'score': 0.9987756609916687},\n", " {'label': 'LABEL_0', 'score': 0.9926889538764954},\n", " {'label': 'LABEL_0', 'score': 0.9980269074440002},\n", " {'label': 'LABEL_0', 'score': 0.995913565158844},\n", " {'label': 'LABEL_0', 'score': 0.9521064162254333},\n", " {'label': 'LABEL_0', 'score': 0.9973504543304443},\n", " {'label': 'LABEL_0', 'score': 0.9951818585395813},\n", " {'label': 'LABEL_0', 'score': 0.9943795800209045},\n", " {'label': 'LABEL_0', 'score': 0.993586003780365},\n", " {'label': 'LABEL_0', 'score': 0.996479332447052},\n", " {'label': 'LABEL_0', 'score': 0.997653067111969},\n", " {'label': 'LABEL_0', 'score': 0.9808756113052368},\n", " {'label': 'LABEL_0', 'score': 0.9970345497131348},\n", " {'label': 'LABEL_0', 'score': 0.9431267380714417},\n", " {'label': 'LABEL_0', 'score': 0.9003729224205017},\n", " {'label': 'LABEL_0', 'score': 0.9930608868598938},\n", " {'label': 'LABEL_1', 'score': 0.7583749294281006},\n", " {'label': 'LABEL_0', 'score': 0.9948474168777466},\n", " {'label': 'LABEL_1', 'score': 0.5063807964324951},\n", " {'label': 'LABEL_0', 'score': 0.9589183330535889},\n", " {'label': 'LABEL_1', 'score': 0.9781742095947266},\n", " {'label': 'LABEL_0', 'score': 0.9977648258209229},\n", " {'label': 'LABEL_0', 'score': 0.9987187385559082},\n", " {'label': 'LABEL_0', 'score': 0.9974794983863831},\n", " {'label': 'LABEL_0', 'score': 0.9986817240715027},\n", " {'label': 'LABEL_0', 'score': 0.996134877204895},\n", " {'label': 'LABEL_0', 'score': 0.9976504445075989},\n", " {'label': 'LABEL_0', 'score': 0.9987467527389526},\n", " {'label': 'LABEL_0', 'score': 0.9985254406929016},\n", " {'label': 'LABEL_0', 'score': 0.9932849407196045},\n", " {'label': 'LABEL_0', 'score': 0.5109266638755798},\n", " {'label': 'LABEL_0', 'score': 0.9801786541938782},\n", " {'label': 'LABEL_0', 'score': 0.9933167695999146},\n", " {'label': 'LABEL_0', 'score': 0.9007285833358765},\n", " {'label': 'LABEL_0', 'score': 0.986695408821106},\n", " {'label': 'LABEL_0', 'score': 0.9742431640625},\n", " {'label': 'LABEL_1', 'score': 0.7220762372016907},\n", " {'label': 'LABEL_0', 'score': 0.9923557043075562},\n", " {'label': 'LABEL_0', 'score': 0.9706437587738037},\n", " {'label': 'LABEL_0', 'score': 0.9988356232643127},\n", " {'label': 'LABEL_0', 'score': 0.9987884163856506},\n", " {'label': 'LABEL_0', 'score': 0.9797398447990417},\n", " {'label': 'LABEL_0', 'score': 0.987711489200592},\n", " {'label': 'LABEL_0', 'score': 0.9961549639701843},\n", " {'label': 'LABEL_0', 'score': 0.6292486190795898},\n", " {'label': 'LABEL_0', 'score': 0.9972221851348877},\n", " {'label': 'LABEL_0', 'score': 0.9987063407897949},\n", " {'label': 'LABEL_0', 'score': 0.998217761516571},\n", " {'label': 'LABEL_0', 'score': 0.995513379573822},\n", " {'label': 'LABEL_0', 'score': 0.9983007311820984},\n", " {'label': 'LABEL_0', 'score': 0.9979398846626282},\n", " {'label': 'LABEL_0', 'score': 0.9986240863800049},\n", " {'label': 'LABEL_0', 'score': 0.9977601766586304},\n", " {'label': 'LABEL_0', 'score': 0.9974849224090576},\n", " {'label': 'LABEL_0', 'score': 0.949386715888977},\n", " {'label': 'LABEL_0', 'score': 0.969242513179779},\n", " {'label': 'LABEL_0', 'score': 0.978050708770752},\n", " {'label': 'LABEL_0', 'score': 0.9853933453559875},\n", " {'label': 'LABEL_0', 'score': 0.9960567951202393},\n", " {'label': 'LABEL_0', 'score': 0.9965749382972717},\n", " {'label': 'LABEL_0', 'score': 0.9987173080444336},\n", " {'label': 'LABEL_0', 'score': 0.9980658888816833},\n", " {'label': 'LABEL_0', 'score': 0.9988677501678467},\n", " {'label': 'LABEL_0', 'score': 0.9987745881080627},\n", " {'label': 'LABEL_0', 'score': 0.9977912902832031},\n", " {'label': 'LABEL_1', 'score': 0.8811283707618713},\n", " {'label': 'LABEL_0', 'score': 0.9986131191253662},\n", " {'label': 'LABEL_0', 'score': 0.9301124811172485},\n", " {'label': 'LABEL_0', 'score': 0.9980668425559998},\n", " {'label': 'LABEL_1', 'score': 0.8134693503379822},\n", " {'label': 'LABEL_0', 'score': 0.9951452612876892},\n", " {'label': 'LABEL_0', 'score': 0.9973406195640564},\n", " {'label': 'LABEL_0', 'score': 0.9889533519744873},\n", " {'label': 'LABEL_0', 'score': 0.9961967468261719},\n", " {'label': 'LABEL_0', 'score': 0.9389035105705261},\n", " {'label': 'LABEL_0', 'score': 0.9983087778091431},\n", " {'label': 'LABEL_0', 'score': 0.9805503487586975},\n", " {'label': 'LABEL_0', 'score': 0.9979610443115234},\n", " {'label': 'LABEL_0', 'score': 0.9981979727745056},\n", " {'label': 'LABEL_0', 'score': 0.9982859492301941},\n", " {'label': 'LABEL_0', 'score': 0.9986773133277893},\n", " {'label': 'LABEL_0', 'score': 0.9940134882926941},\n", " {'label': 'LABEL_0', 'score': 0.9958133101463318},\n", " {'label': 'LABEL_0', 'score': 0.9986673593521118},\n", " {'label': 'LABEL_0', 'score': 0.9969382286071777},\n", " {'label': 'LABEL_0', 'score': 0.9913495779037476},\n", " {'label': 'LABEL_0', 'score': 0.9988558292388916},\n", " {'label': 'LABEL_0', 'score': 0.9801749587059021},\n", " {'label': 'LABEL_1', 'score': 0.9942229390144348},\n", " {'label': 'LABEL_0', 'score': 0.9904804825782776},\n", " {'label': 'LABEL_0', 'score': 0.9985619187355042},\n", " {'label': 'LABEL_0', 'score': 0.9981953501701355},\n", " {'label': 'LABEL_0', 'score': 0.9989521503448486},\n", " {'label': 'LABEL_0', 'score': 0.9985705614089966},\n", " {'label': 'LABEL_0', 'score': 0.8995286822319031},\n", " {'label': 'LABEL_0', 'score': 0.998676598072052},\n", " {'label': 'LABEL_0', 'score': 0.9984286427497864},\n", " {'label': 'LABEL_0', 'score': 0.6487972140312195},\n", " {'label': 'LABEL_0', 'score': 0.9394513368606567},\n", " ...]" ] }, "execution_count": 33, "metadata": {}, "output_type": "execute_result" } ], "source": [ "predictions" ] }, { "cell_type": "code", "execution_count": 36, "id": "3f35d373-a3a4-4186-8663-d320edb22092", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T15:43:24.194303Z", "iopub.status.busy": "2024-05-23T15:43:24.193637Z", "iopub.status.idle": "2024-05-23T15:43:24.224829Z", "shell.execute_reply": "2024-05-23T15:43:24.224212Z", "shell.execute_reply.started": "2024-05-23T15:43:24.194276Z" } }, "outputs": [], "source": [ "import numpy as np\n", "import evaluate\n", "from sklearn.metrics import accuracy_score" ] }, { "cell_type": "code", "execution_count": 42, "id": "80515948-0db5-466e-8aa6-1aeb53e79cb6", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T15:46:26.164515Z", "iopub.status.busy": "2024-05-23T15:46:26.164217Z", "iopub.status.idle": "2024-05-23T15:46:26.169211Z", "shell.execute_reply": "2024-05-23T15:46:26.168571Z", "shell.execute_reply.started": "2024-05-23T15:46:26.164488Z" } }, "outputs": [], "source": [ "preds = [int(x[\"label\"].split(\"_\")[1]) for x in predictions]" ] }, { "cell_type": "code", "execution_count": 63, "id": "b97ff6d7-6460-4ab9-b163-680a47135cbc", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T16:07:07.653818Z", "iopub.status.busy": "2024-05-23T16:07:07.653404Z", "iopub.status.idle": "2024-05-23T16:07:07.659790Z", "shell.execute_reply": "2024-05-23T16:07:07.658732Z", "shell.execute_reply.started": "2024-05-23T16:07:07.653793Z" } }, "outputs": [], "source": [ "predictions_proba = np.array([x[\"score\"] if int(x[\"label\"].split(\"_\")[1]) == 1 else (1.0 - x[\"score\"]) for x in predictions])" ] }, { "cell_type": "code", "execution_count": 64, "id": "ffa8d866-1174-4564-8776-468e7a69a0dd", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T16:07:12.709461Z", "iopub.status.busy": "2024-05-23T16:07:12.709136Z", "iopub.status.idle": "2024-05-23T16:07:12.719229Z", "shell.execute_reply": "2024-05-23T16:07:12.718143Z", "shell.execute_reply.started": "2024-05-23T16:07:12.709436Z" } }, "outputs": [], "source": [ "predictions_probs = np.array([np.array([x[\"score\"], (1.0 - x[\"score\"])]) if int(x[\"label\"].split(\"_\")[1]) == 0 else np.array([(1.0 - x[\"score\"]), x[\"score\"]]) for x in predictions])" ] }, { "cell_type": "code", "execution_count": 65, "id": "f4e20d3c-241a-4c08-adf1-527a1e3fa19c", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T16:07:14.013838Z", "iopub.status.busy": "2024-05-23T16:07:14.013485Z", "iopub.status.idle": "2024-05-23T16:07:14.021152Z", "shell.execute_reply": "2024-05-23T16:07:14.019934Z", "shell.execute_reply.started": "2024-05-23T16:07:14.013813Z" } }, "outputs": [ { "data": { "text/plain": [ "(1, 1, 0.9920824766159058, array([0.00791752, 0.99208248]))" ] }, "execution_count": 65, "metadata": {}, "output_type": "execute_result" } ], "source": [ "preds[0], y_test[0], predictions_proba[0], predictions_probs[0]" ] }, { "cell_type": "code", "execution_count": 66, "id": "553f215e-cb9c-45be-8e09-33c840a1c095", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T16:07:20.087631Z", "iopub.status.busy": "2024-05-23T16:07:20.086457Z", "iopub.status.idle": "2024-05-23T16:07:20.111547Z", "shell.execute_reply": "2024-05-23T16:07:20.110831Z", "shell.execute_reply.started": "2024-05-23T16:07:20.087591Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " precision recall f1-score support\n", "\n", "non-crystallizable 0.73 0.97 0.83 1000\n", " crystallizable 0.94 0.60 0.73 898\n", "\n", " accuracy 0.79 1898\n", " macro avg 0.83 0.78 0.78 1898\n", " weighted avg 0.83 0.79 0.78 1898\n", "\n", "[[966 34]\n", " [362 536]]\n", "0.9467594654788418\n" ] } ], "source": [ "from sklearn.metrics import classification_report, confusion_matrix, roc_auc_score\n", "target_names = ['non-crystallizable', 'crystallizable']\n", "\n", "print(classification_report(y_test, preds, target_names=target_names))\n", "print(confusion_matrix(y_test, preds))\n", "print(roc_auc_score(y_test, predictions_proba))" ] }, { "cell_type": "code", "execution_count": 67, "id": "44be3b34-5e77-4814-ac8d-682785bf58b4", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T16:07:23.695122Z", "iopub.status.busy": "2024-05-23T16:07:23.694785Z", "iopub.status.idle": "2024-05-23T16:07:23.702798Z", "shell.execute_reply": "2024-05-23T16:07:23.702046Z", "shell.execute_reply.started": "2024-05-23T16:07:23.695098Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "0.9467594654788418\n" ] } ], "source": [ "print(roc_auc_score(y_test, predictions_proba))" ] }, { "cell_type": "code", "execution_count": 59, "id": "ea12fbb0-2105-473d-9774-833d9f987a12", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T16:00:26.155037Z", "iopub.status.busy": "2024-05-23T16:00:26.154548Z", "iopub.status.idle": "2024-05-23T16:00:26.166013Z", "shell.execute_reply": "2024-05-23T16:00:26.165322Z", "shell.execute_reply.started": "2024-05-23T16:00:26.155000Z" } }, "outputs": [], "source": [ "from sklearn.metrics import roc_curve, auc\n", "n_classes = 2\n", "# Compute ROC curve and ROC area for each class\n", "fpr = dict()\n", "tpr = dict()\n", "roc_auc = dict()\n", "for i in range(n_classes):\n", " fpr[i], tpr[i], _ = roc_curve(y_test, predictions_probs[:, i])\n", " roc_auc[i] = auc(fpr[i], tpr[i])\n", "\n", "# Compute micro-average ROC curve and ROC area\n", "fpr[\"micro\"], tpr[\"micro\"], _ = roc_curve(y_test, predictions_proba)\n", "roc_auc[\"micro\"] = auc(fpr[\"micro\"], tpr[\"micro\"])" ] }, { "cell_type": "code", "execution_count": 60, "id": "5775d0d0-b46a-4b77-8610-a2b9b6ed1cd7", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T16:01:15.759696Z", "iopub.status.busy": "2024-05-23T16:01:15.759387Z", "iopub.status.idle": "2024-05-23T16:01:15.996034Z", "shell.execute_reply": "2024-05-23T16:01:15.995052Z", "shell.execute_reply.started": "2024-05-23T16:01:15.759673Z" } }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkIAAAHHCAYAAABTMjf2AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/OQEPoAAAACXBIWXMAAA9hAAAPYQGoP6dpAAB840lEQVR4nO3deVxN+f8H8Nct7ZuStIiUnWxZs2QbGTP2JYPIzlhmLDP2MLbB2GYY2wxhGLvR19YMI2vDiOwKMUJFpE2L7v38/vDrcFV0U53qvp6PRw/u+55z7vvcU913n/NZFEIIASIiIiItpCN3AkRERERyYSFEREREWouFEBEREWktFkJERESktVgIERERkdZiIURERERai4UQERERaS0WQkRERKS1WAgRERGR1mIhREWWk5MTfHx85E5DK/j4+MDJyUnuNLLVsmVL1KxZU+40Cp3AwEAoFAoEBgbmyfH8/PygUChw//79PDlecVfYf27oNRZClKWMX3gZXyVKlICDgwN8fHzw6NEjudOjfPD48WPMmjULISEhcqeiVebPn48//vhD7jTUFMaciPJLCbkToMLtu+++Q4UKFZCSkoJ//vkHfn5+OH36NK5duwZDQ0NZcwsNDYWODmv5vPL48WPMnj0bTk5OqFOnjtpz69evh0qlkiexYm7+/Pno0aMHunTpkufHbtGiBZKTk6Gvr58nOXl7e6N3794wMDDIwyyJ5MVCiN7r008/Rf369QEAQ4YMgbW1NRYuXAh/f3/06tVL1tzk+GWckpICfX39IlGA5WWuenp6eZBR0Zaeng6VSqVxUSGHt699Xv7BoqurC11d3Tw7HlFhUPh/m1Oh0rx5cwDA3bt31eK3bt1Cjx49YGVlBUNDQ9SvXx/+/v6Z9n/x4gXGjRsHJycnGBgYoGzZsujfvz9iYmKkbVJTUzFz5kxUrFgRBgYGcHR0xLfffovU1FS1Y73dR+jChQtQKBTYtGlTptcMCAiAQqHAgQMHpNijR48waNAglClTBgYGBqhRowY2bNigtl9G/4rt27dj+vTpcHBwgLGxMeLj47N9f5KSkjBhwgQ4OjrCwMAAVapUwQ8//AAhhNp2CoUCo0ePxtatW1GlShUYGhrCzc0NJ0+ezHTMj831+fPnmDhxIlxdXWFqagpzc3N8+umnuHz5str+DRo0AAAMHDhQuiXq5+cHIHNfh/v370OhUOCHH37AunXr4OLiAgMDAzRo0AD//vtvpnPYtWsXqlevDkNDQ9SsWRP79u3TqP/E4cOH4eHhATMzM5ibm6NBgwbYtm1bpu1u3LiBVq1awdjYGA4ODli0aJHa82lpafD19YWbmxssLCxgYmKC5s2b4/jx42rbvX1+y5cvl87vxo0bOT4GAKhUKqxYsQKurq4wNDRE6dKl0b59e1y4cAHA6++DpKQkbNq0SXrP3+739rHXPqs+Qrdv30b37t1ha2sLQ0NDlC1bFr1790ZcXNwHc8quj1BOr8+7PnR+ycnJqFq1KqpWrYrk5GQp/vz5c9jZ2cHd3R1KpRIAcOXKFfj4+MDZ2RmGhoawtbXFoEGD8OzZM7XXnDVrFhQKBcLCwtCvXz9YWFigdOnSmDFjBoQQiIiIQOfOnWFubg5bW1ssWbIky/d7x44dmDp1KmxtbWFiYoJOnTohIiLig+esUqmwfPly1KhRA4aGhihTpgyGDx+O2NjYD+5L+YMtQqSRjF+AlpaWUuz69eto2rQpHBwcMHnyZJiYmGDnzp3o0qUL9uzZg65duwIAEhMT0bx5c9y8eRODBg1CvXr1EBMTA39/fzx8+BDW1tZQqVTo1KkTTp8+jWHDhqFatWq4evUqli1bhrCwsGz7LdSvXx/Ozs7YuXMnBgwYoPbcjh07YGlpCU9PTwBAdHQ0GjduLBUjpUuXxuHDhzF48GDEx8fj66+/Vtt/zpw50NfXx8SJE5Gamppti4AQAp06dcLx48cxePBg1KlTBwEBAfjmm2/w6NEjLFu2TG37EydOYMeOHRg7diwMDAzw888/o3379jh//rzU8Tcvcr1x4wb++OMP9OzZExUqVEB0dDTWrl0LDw8P3LhxA/b29qhWrRq+++47+Pr6YtiwYVLB6+7unvU3wv/btm0bEhISMHz4cCgUCixatAjdunVDeHi41Ip08OBBeHl5wdXVFQsWLEBsbCwGDx4MBweH9x47g5+fHwYNGoQaNWpgypQpKFmyJC5duoQjR46gT58+0naxsbFo3749unXrhl69emH37t2YNGkSXF1d8emnnwIA4uPj8csvv+CLL77A0KFDkZCQgF9//RWenp44f/58pluCGzduREpKCoYNGwYDAwNYWVlpdIzBgwfDz88Pn376KYYMGYL09HScOnUK//zzD+rXr48tW7ZgyJAhaNiwIYYNGwYAcHFxybNr/660tDR4enoiNTUVY8aMga2tLR49eoQDBw7gxYsXsLCweG9OH3N93pWT8zMyMsKmTZvQtGlTTJs2DUuXLgUAjBo1CnFxcfDz85NaqP766y+Eh4dj4MCBsLW1xfXr17Fu3Tpcv34d//zzDxQKhdrre3l5oVq1avj+++9x8OBBzJ07F1ZWVli7di1at26NhQsXYuvWrZg4cSIaNGiAFi1aqO0/b948KBQKTJo0CU+ePMHy5cvRtm1bhISEwMjIKNvzHj58OPz8/DBw4ECMHTsW9+7dw8qVK3Hp0iWcOXOGra9yEERZ2LhxowAgjh49Kp4+fSoiIiLE7t27RenSpYWBgYGIiIiQtm3Tpo1wdXUVKSkpUkylUgl3d3dRqVIlKebr6ysAiL1792Z6PZVKJYQQYsuWLUJHR0ecOnVK7fk1a9YIAOLMmTNSrHz58mLAgAHS4ylTpgg9PT3x/PlzKZaamipKliwpBg0aJMUGDx4s7OzsRExMjNpr9O7dW1hYWIiXL18KIYQ4fvy4ACCcnZ2l2Pv88ccfAoCYO3euWrxHjx5CoVCIO3fuSDEAAoC4cOGCFPvvv/+EoaGh6Nq1a57mmpKSIpRKpVrs3r17wsDAQHz33XdS7N9//xUAxMaNGzOd24ABA0T58uXV9gcgSpUqpfZ+79+/XwAQ//vf/6SYq6urKFu2rEhISJBigYGBAoDaMbPy4sULYWZmJho1aiSSk5PVnsv4nhFCCA8PDwFAbN68WYqlpqYKW1tb0b17dymWnp4uUlNT1Y4TGxsrypQpo/Y9knF+5ubm4smTJ2rb5/QYf//9twAgxo4dm+m83s7dxMRE7fs4Q15c+4znjh8/LoQQ4tKlSwKA2LVrV6bXe1t2OWX8Xrh3754QIufXJys5PT8hXv9s6+joiJMnT4pdu3YJAGL58uVq+2X1M/r7778LAOLkyZNSbObMmQKAGDZsmBRLT08XZcuWFQqFQnz//fdSPDY2VhgZGam9FxnvqYODg4iPj5fiO3fuFADEihUrpNi7PzenTp0SAMTWrVvV8jxy5EiWcSoYvDVG79W2bVuULl0ajo6O6NGjB0xMTODv74+yZcsCeN1E/ffff6NXr15ISEhATEwMYmJi8OzZM3h6euL27dvSKLM9e/agdu3aUgvR2zL+Wtu1axeqVauGqlWrSseKiYlB69atASDL2w8ZvLy88OrVK+zdu1eK/fnnn3jx4gW8vLwAvG612bNnDzp27AghhNpreHp6Ii4uDhcvXlQ77oABA977F16GQ4cOQVdXF2PHjlWLT5gwAUIIHD58WC3epEkTuLm5SY/LlSuHzp07IyAgAEqlMs9yNTAwkPoJKZVKPHv2DKampqhSpUqm/TXl5eWl1jqY0ZIUHh4O4HUH7KtXr6J///4wNTWVtvPw8ICrq+sHj//XX38hISEBkydPztTX5d2/8E1NTdGvXz/psb6+Pho2bCjlArzu45LRUqJSqfD8+XOkp6ejfv36Wb4X3bt3R+nSpdViOT3Gnj17oFAoMHPmzEzHfTf3d+XX96mFhQWA17eLX758+d5tc0KT6/M2Tc9v1qxZqFGjBgYMGIAvv/wSHh4emX7O3j73lJQUxMTEoHHjxgCQ5bUdMmSI9H9dXV3Ur18fQggMHjxYipcsWRJVqlRR+x7K0L9/f5iZmUmPe/ToATs7Oxw6dCjb8961axcsLCzwySefqJ2zm5sbTE1N3/v7jfIPb43Re61atQqVK1dGXFwcNmzYgJMnT6p1Ur5z5w6EEJgxYwZmzJiR5TGePHkCBwcH3L17F927d3/v692+fRs3b97M9OHz9rGyU7t2bVStWhU7duyQfpnt2LED1tbWUiH19OlTvHjxAuvWrcO6dety9BoVKlR4b84Z/vvvP9jb26v9cgSAatWqSc+/rVKlSpmOUblyZbx8+RJPnz6Fjo5OnuSa0U/l559/xr1796Q+FQBQqlSpHJ1bdsqVK6f2OKMoyujvkHHOFStWzLRvxYoVP1iIZfRFy8kcQWXLls304WtpaYkrV66oxTZt2oQlS5bg1q1bePXqlRTP6r3L7trn5Bh3796Fvb09rKysPpj7u/Lr+7RChQoYP348li5diq1bt6J58+bo1KmT1FdGU5pcn7dpen76+vrYsGEDGjRoAENDQ2zcuDHTtX7+/Dlmz56N7du3Z3pvMvo/ve3d710LCwsYGhrC2to6U/zdfkZA5p9fhUKBihUrvneOpdu3byMuLg42NjZZPv++32+Uf1gI0Xs1bNhQGjXWpUsXNGvWDH369EFoaChMTU2lIdUTJ06U+uC8K6sPweyoVCq4urpKfQHe5ejo+N79vby8MG/ePMTExMDMzAz+/v744osvUKJECen4ANCvX79MfYky1KpVS+1xTlqD8kNe5Tp//nzMmDEDgwYNwpw5c2BlZQUdHR18/fXXHz0kPrsRROKdzuEFISe5/Pbbb/Dx8UGXLl3wzTffwMbGBrq6uliwYEGmAQBA1u+npsfIjfz8Pl2yZAl8fHywf/9+/Pnnnxg7diwWLFiAf/75R2rpzW+5Ob+AgAAAr1t7bt++nanw69WrF86ePYtvvvkGderUkX4/tW/fPsvv86y+X/L7+1mlUsHGxgZbt27N8vns/gCk/MVCiHIs45d9q1atsHLlSkyePBnOzs4AXg+vbtu27Xv3d3FxwbVr1z64zeXLl9GmTZsP3j7IipeXF2bPno09e/agTJkyiI+PR+/evaXnS5cuDTMzMyiVyg/mq6ny5cvj6NGjSEhIUGsVunXrlvT8227fvp3pGGFhYTA2NpZ+IeZFrrt370arVq3w66+/qsVfvHih9tdvbt7vD8k45zt37mR6LqvYuzI66V67dk2jgjo7u3fvhrOzM/bu3at2vlndvvrYY7i4uCAgIADPnz9/b6tQVu97fn6fAoCrqytcXV0xffp0nD17Fk2bNsWaNWswd+7cbHPKSm6vj6bnd+XKFXz33XcYOHAgQkJCMGTIEFy9elVqxYqNjcWxY8cwe/Zs+Pr6Svtl9TOWV949thACd+7cyVTAvc3FxQVHjx5F06ZNZfsDizJjHyHSSMuWLdGwYUMsX74cKSkpsLGxQcuWLbF27VpERkZm2v7p06fS/7t3747Lly9j3759mbbL+IurV69eePToEdavX59pm+TkZCQlJb03v2rVqsHV1RU7duzAjh07YGdnpzbaQ1dXF927d8eePXuyLMrezldTHTp0gFKpxMqVK9Xiy5Ytg0KhkEYuZQgKClK7NRQREYH9+/ejXbt20nwteZGrrq5upr9od+3alWmGcBMTEwCvC6S8Ym9vj5o1a2Lz5s1ITEyU4idOnMDVq1c/uH+7du1gZmaGBQsWICUlRe253PyVnvEX/9v7njt3DkFBQXl+jO7du0MIgdmzZ2c6xtv7mpiYZHrP8+v7ND4+Hunp6WoxV1dX6OjoqE1PkVVOWcnt9dHk/F69egUfHx/Y29tjxYoV8PPzQ3R0NMaNG6d2vKxec/ny5R88h9zavHkzEhISpMe7d+9GZGRkpp/zt/Xq1QtKpRJz5szJ9Fx6enqe/uxRzrFFiDT2zTffoGfPnvDz88OIESOwatUqNGvWDK6urhg6dCicnZ0RHR2NoKAgPHz4UJqv5ptvvsHu3bvRs2dPDBo0CG5ubnj+/Dn8/f2xZs0a1K5dG97e3ti5cydGjBiB48ePo2nTplAqlbh16xZ27tyJgIAA6VZddry8vODr6wtDQ0MMHjw404SC33//PY4fP45GjRph6NChqF69Op4/f46LFy/i6NGjeP78ea7el44dO6JVq1aYNm0a7t+/j9q1a+PPP//E/v378fXXX2caglyzZk14enqqDZ8HoPbBmRe5fv7559Jf0+7u7rh69Sq2bt0qteZlcHFxQcmSJbFmzRqYmZnBxMQEjRo1ynEfqezMnz8fnTt3RtOmTTFw4EDExsZi5cqVqFmzplpxlBVzc3MsW7YMQ4YMQYMGDdCnTx9YWlri8uXLePnyZZbzRr3P559/jr1796Jr16747LPPcO/ePaxZswbVq1f/YC6aHqNVq1bw9vbGjz/+iNu3b0u3aE6dOoVWrVph9OjRAAA3NzccPXoUS5cuhb29PSpUqIBGjRrly/fp33//jdGjR6Nnz56oXLky0tPTsWXLFqkwyZBdTu/6mOuT0/ObO3cuQkJCcOzYMZiZmaFWrVrw9fXF9OnT0aNHD3To0AHm5uZo0aIFFi1ahFevXsHBwQF//vkn7t27p/F7lFNWVlZo1qwZBg4ciOjoaCxfvhwVK1bE0KFDs93Hw8MDw4cPx4IFCxASEoJ27dpBT08Pt2/fxq5du7BixQr06NEj33KmbBTcADUqSjKGyf7777+ZnlMqlcLFxUW4uLiI9PR0IYQQd+/eFf379xe2trZCT09PODg4iM8//1zs3r1bbd9nz56J0aNHCwcHB6Gvry/Kli0rBgwYoDaENi0tTSxcuFDUqFFDGBgYCEtLS+Hm5iZmz54t4uLipO3eHT6f4fbt29Lw9NOnT2d5ftHR0WLUqFHC0dFR6OnpCVtbW9GmTRuxbt06aZuMYbIfGmr8toSEBDFu3Dhhb28v9PT0RKVKlcTixYszDSUGIEaNGiV+++03UalSJWFgYCDq1q0rDXPOy1xTUlLEhAkThJ2dnTAyMhJNmzYVQUFBwsPDQ3h4eKhtu3//flG9enVRokQJtaH02Q2fX7x4cabXAyBmzpypFtu+fbuoWrWqMDAwEDVr1hT+/v6ie/fuomrVqu9/Q/+fv7+/cHd3F0ZGRsLc3Fw0bNhQ/P7779LzHh4eokaNGpn2ezdvlUol5s+fL8qXLy+95wcOHNDo/HJ6DCFeD8tevHixqFq1qtDX1xelS5cWn376qQgODpa2uXXrlmjRooUwMjISANS+pz/22r87fD48PFwMGjRIuLi4CENDQ2FlZSVatWoljh49qrZfdjm9O3w+w4euT3Y+dH7BwcGiRIkSYsyYMZne1wYNGgh7e3sRGxsrhBDi4cOHomvXrqJkyZLCwsJC9OzZUzx+/DjT92PG8PmnT5+qHXPAgAHCxMQkU47vfm9lvKe///67mDJlirCxsRFGRkbis88+E//991+mY2Y1RcS6deuEm5ubMDIyEmZmZsLV1VV8++234vHjxx98zyjvKYSQoVcjkZZTKBQYNWpUptto2qROnTooXbo0/vrrL7lTIcqxwMBAtGrVCrt27WLrTTHBPkJElK9evXqVqV9KYGAgLl++jJYtW8qTFBHR/2MfISLKV48ePULbtm3Rr18/2Nvb49atW1izZg1sbW0xYsQIudMjIi3HQoiI8pWlpSXc3Nzwyy+/4OnTpzAxMcFnn32G77///qMndCQi+ljsI0RERERai32EiIiISGuxECIiIiKtpXV9hFQqFR4/fgwzM7N8WVKAiIiI8p4QAgkJCbC3t880Ue7H0LpC6PHjxx9cuJOIiIgKp4iIiDxdIFjrCqGMxTAjIiJgbm4uczZERESUE/Hx8XB0dFRb1DovaF0hlHE7zNzcnIUQERFREZPX3VrYWZqIiIi0FgshIiIi0loshIiIiEhrsRAiIiIircVCiIiIiLQWCyEiIiLSWiyEiIiISGuxECIiIiKtxUKIiIiItBYLISIiItJashZCJ0+eRMeOHWFvbw+FQoE//vjjg/sEBgaiXr16MDAwQMWKFeHn55fveRIREVHxJGshlJSUhNq1a2PVqlU52v7evXv47LPP0KpVK4SEhODrr7/GkCFDEBAQkM+ZEhERUXEk66Krn376KT799NMcb79mzRpUqFABS5YsAQBUq1YNp0+fxrJly+Dp6ZlfaRIREVExVaT6CAUFBaFt27ZqMU9PTwQFBcmUEREREeU3lUrg+vUn+XJsWVuENBUVFYUyZcqoxcqUKYP4+HgkJyfDyMgo0z6pqalITU2VHsfHx+d7nkRERDkSugs46wukJcidSaEVGWeEgZs8cCLMKl+OX6QKodxYsGABZs+eLXcaRJQb/JCg4i7xkdwZFGr7r1XBkF2dEJNkAiAlX16jSBVCtra2iI6OVotFR0fD3Nw8y9YgAJgyZQrGjx8vPY6Pj4ejo2O+5klUaBW1woIfEqRNTB3kzqBQeZpgiL6/90BSqh4AwMYsGU/y4VdXkSqEmjRpgkOHDqnF/vrrLzRp0iTbfQwMDGBgYJDfqRHln7wsXopyYcEPCSqu9M2ApnOAyj3kzqRQKQ1gecmLGDr0f+jSpSqWLvWAs/OKPH8dWQuhxMRE3LlzR3p87949hISEwMrKCuXKlcOUKVPw6NEjbN68GQAwYsQIrFy5Et9++y0GDRqEv//+Gzt37sTBgwflOgWi/G9lya/ipagUFvyQINIKSqUK6ekqGBi8KU0GD64LR0dztGvngoSE/PkdK2shdOHCBbRq1Up6nHELa8CAAfDz80NkZCQePHggPV+hQgUcPHgQ48aNw4oVK1C2bFn88ssvHDpPBSeroqcgW1nyonhhYUFEhUxERBz69/8DNWuWxk8/dZDiCoUCnp4V8/W1FUIIka+vUMjEx8fDwsICcXFxMDc3lzsdKqyya+X5UNGTX60sLF6IqJjaufM6hg8/gBcvXneGPniwDzp0qJRpu/z6/C5SfYSI8tT7bmnlpJXn7aKHhQoRkUbi41MxduxhbNp0WYo5OprDzEy/QPNgIUTa493CJ6e3tN5t5WHRQ0T0UYKCItCv3z6Eh8dKMS+vGli9+jNYWmY9Cjy/sBAi7XHWF3h+K+vnsrqlxYKHiChPpaerMG/eScyZcxJK5eueOWZm+li1qgP69asFhUJR4DmxEKLiJ7tbXkmRr/9V6AAmdq//z2KHiKhAPHv2Eh07/o6goIdSzN3dEb/91hUVKljKlhcLISr6NL3lZVkZGHgz//MiIiJJyZKGKFHi9RKnuroK+Pp6YOrU5lJMLiyEqOjT5JZXRgsQEREVKF1dHWzZ0hXduu3EqlUd0LhxWblTAsBCiIqid1uAeMuLiKjQOXHiPoyM9NCw4Zs/SMuXL4kLF4bK0hcoOyyEqGgJ3QUc6JX1c7zlRUQku7Q0JWbOPI6FC8+gQgVLhIQMh5nZm6WuClMRBLAQoqLg7Ragd/v/ZNz64i0vIiLZhYbGoE+fvbh48XVLfXh4LFavvoBvv20qc2bZYyFEhdv7WoA67uKtLyKiQkAIgfXrL+Lrr48gOTkdAKCnp4N581pjwgR3mbN7PxZCVHhlVQSZOrD/DxFRIfL0aRKGDv0f9u8PlWJVqpTCtm3dUa+enYyZ5QwLISqcsiqC2AJERFSoBATcgY/PfkRFJUqxESPcsGSJJ4yN9WTMLOdYCFHh8KG5gFgEEREVKtHRiejSZQdSUl7fCrO2NsaGDZ3QsWMVmTPTjLyzGBGF7gI2Vnvd+vP81usCiEUQEVGhV6aMKb7/vg0AwNPTBVevjixyRRDAFiGSS0YLUFYTIb47EoxFEBGR7FQqAaVSBT09XSk2ZkwjlC1rjq5dq0FHp3ANi88pFkJUMHKyDIZVVRY+RESFUGRkAnx89qNOnTJYuPATKa6jo0D37tVlzOzjsRCigvG+ZTBYABERFVr799/C4MH+ePYsGX/9dReenhXRunUFudPKMyyEKH9ltATFhr1+zGUwiIiKhKSkNEyY8CfWrg2WYmXKmMqYUf5gIUT5J6sh8FwGg4io0AsOfow+ffYiLOyZFOvcuQp++aUTrK2NZcws77EQorz1vuUwMm6BERFRoaRUqvDDD2cxffpxpKerAADGxnpYvtwTQ4bUK3TrhOUFFkKUt7LrC8Qh8EREhVpMzEv07LkLgYH3pZibmx22beuOypVLyZdYPmMhRB/v7VagpNcL7Ul9gdgPiIioSLCwMEBiYhoAQKEAJk9uhlmzWkJfX/cDexZtLITo42S3KCr7AhERFSl6errYurUbunTZjtWrP4OHh5PcKRUIFkKUex9aFJWIiAqtoKAIGBvroXZtWylWuXIpXLv2ZZGdHDE3WAhR7p31VX/MfkBERIVeeroK8+adxJw5J1G5cilcuDBMbYFUbSqCAK41RrkVuku9UzSLICKiQi88PBYtWmzErFknoFQK3LwZg59//lfutGTFFiHKnbdbg6yqsggiIirEhBDYsuUKRo8+hISE1x2idXUVmDnTA19/3Vjm7OTFQohyJ2PNMID9gYiICrHY2GSMGHEQO3del2IuLpb47bduaNy4rIyZFQ4shEgzGUPlM4bJmzqwNYiIqJAKDLwPb+99ePgwXooNHFgHK1a0h5mZgYyZFR4shCjnsholpm8mTy5ERPRekZEJ8PT8DWlpSgCApaUh1q79HD171pA5s8KFnaUpZ7IqgrhkBhFRoWVnZ4aZMz0AAK1aOeHKlZEsgrLAFiHK2tuzRQOZ1w3jKDEiokJFCAGVSkBX900bx6RJTeHoaI6+fWtp3bD4nGKLEGUtY82wxEcsgoiICrmnT5PQtesOzJ17Ui2uq6sDb+/aLILegy1ClLWMlqCMNcMArhtGRFQIBQTcgY/PfkRFJeLAgTC0a+eCJk0c5U6ryGAhRJmF7nrTCmRiBwx/KG8+RESUSUpKOqZMOYrly89JMUtLI2meIMoZFkKU2duTJXJUGBFRoXP1ajT69t2Lq1efSDFPTxf4+XWBra2pjJkVPSyESN27S2dwVBgRUaGhUgn89NM5TJp0FKmpr4fFGxjoYtGiTzB6dEP2BcoFFkKkjktnEBEVSs+evUTfvnsREHBXirm62mDbtu6oWdNGxsyKNo4aI3VcOoOIqFAyMdHHo0dvfkePG9cY588PZRH0kVgI0Rtvd5Lm0hlERIWKoWEJbNvWDRUqlERAQD8sXeoJQ0Pe2PlYfAfpzeSJb/cNYidpIiJZBQc/homJPqpWtZZirq5lEBY2BiVKsB0jr/CdpMxFEMDbYkREMlEqVVi48DQaN/4VX3yxB6mp6WrPswjKW3w3SX3yRKuqnDmaiEgmERFxaNNmMyZPPob0dBVCQqLw88//yp1WscZbY9rq7bXEkiJfx0zsgIE35c2LiEhL7dx5HcOHH8CLFykAAIUCmDy5GUaNaihzZsUbCyFtk1V/oAzsF0REVODi41MxduxhbNp0WYo5Oppjy5au8PBwki8xLcFCSNtkVQSZOrxZR4yIiApMUFAE+vXbh/DwWCnm5VUDq1d/BktLIxkz0x4shLRFRktQbNjrxwodwLIyF1ElIpLJo0fxaNlyE9LSXs8QbWamj1WrOqBfv1pQKDhDdEFhZ2ltkdESJFSvH1tWft0fiEUQEZEsHBzMMXFiEwCAu7sjLl8eAW/v2iyCChhbhIq797UEERFRgRFCAIBaoTNrVkuUK2eBwYPrcVi8TPiuF3dsCSIikl1sbDJ6996DJUuC1OJ6eroYPrw+iyAZsUWoOMpqaDxbgoiIZBEYeB/e3vvw8GE89u27iTZtKqBuXTu506L/x0KoOMpqZFhGSxARERWItDQlfH2PY9GiM/j/u2IwNdVHVFSivImRGhZCxU3orjdFkELn9SSJHBpPRFSgQkNj0KfPXly8GCnFWrVywubNXVG2rLmMmdG7WAgVF1lNlMhWICKiAiWEwLp1wRg3LgDJya/XCNPT08G8ea0xYYI7dHQ4IqywYSFUXHDhVCIiWT1/noyBA/fD3z9UilWpUgrbtnVHvXrsE1RYsZt6URe6C9hYTX14PBdOJSIqcAYGurh1K0Z6PHJkfVy8OJxFUCHHFqGiLHQXcKCXeoy3w4iIZGFioo+tW7uhc+ftWLPmM3TsWEXulCgHWAgVVVkVQVZVeTuMiKiAXL0aDRMTfTg7W0qx+vXtER4+FgYG/HgtKnhrrKg666v+uOMuTpRIRFQAVCqBFSv+QYMG69G3716kp6vUnmcRVLSwECqK3h4iD7A/EBFRAYmMTMCnn27F118HIDVViX/+eYjVq/+VOy36CLIXQqtWrYKTkxMMDQ3RqFEjnD9//r3bL1++HFWqVIGRkREcHR0xbtw4pKSkFFC2hcTbrUFWVVkEEREVgP37b8HVdTX+/POuFBs3rjGGDnWTMSv6WLK23+3YsQPjx4/HmjVr0KhRIyxfvhyenp4IDQ2FjY1Npu23bduGyZMnY8OGDXB3d0dYWBh8fHygUCiwdOlSGc5ABu+2BrFPEBFRvkpKSsOECX9i7dpgKWZnZwo/vy5o185FxswoLyhExnK4MmjUqBEaNGiAlStXAgBUKhUcHR0xZswYTJ48OdP2o0ePxs2bN3Hs2DEpNmHCBJw7dw6nT5/O0WvGx8fDwsICcXFxMDcvgrN7bqz2phCyqsoRYkRE+Sg4+DH69NmLsLBnUqxLl6pYv74jrK2NZcxM++TX57dst8bS0tIQHByMtm3bvklGRwdt27ZFUFBQlvu4u7sjODhYun0WHh6OQ4cOoUOHDtm+TmpqKuLj49W+irS0hDf/Z2sQEVG+iYiIg7v7BqkIMjbWw/r1HbF3by8WQcWIbIVQTEwMlEolypQpoxYvU6YMoqKistynT58++O6779CsWTPo6enBxcUFLVu2xNSpU7N9nQULFsDCwkL6cnR0zNPzkI2pA/sGERHlI0dHC3z5ZX0AgJubHS5dGo4hQ+pBoeAyGcWJ7J2lNREYGIj58+fj559/xsWLF7F3714cPHgQc+Zk3zIyZcoUxMXFSV8REREFmHEeC90FJD6SOwsiomLr3d4iCxa0xdKl7XD27GBUrlxKpqwoP8nWWdra2hq6urqIjo5Wi0dHR8PW1jbLfWbMmAFvb28MGTIEAODq6oqkpCQMGzYM06ZNg45O5rrOwMAABgYGeX8CBe3dCRT1zeTLhYiomImPT8XYsYfRsKEDvvyygRQ3NCyBceOayJgZ5TfZWoT09fXh5uam1vFZpVLh2LFjaNIk62+6ly9fZip2dHV1AWSu4ouVrGaRZv8gIqI8ERQUgTp11mDTpsuYMOFP3Lz5VO6UqADJOnx+/PjxGDBgAOrXr4+GDRti+fLlSEpKwsCBAwEA/fv3h4ODAxYsWAAA6NixI5YuXYq6deuiUaNGuHPnDmbMmIGOHTtKBVGxk1URxAkUiYg+Wnq6CnPnnsTcuSehVL7+Y1pPTwd378aiWrXSMmdHBUXWQsjLywtPnz6Fr68voqKiUKdOHRw5ckTqQP3gwQO1FqDp06dDoVBg+vTpePToEUqXLo2OHTti3rx5cp1C/mIRRESUL8LDY9Gv314EBT2UYu7ujvjtt66oUMHyPXtScSPrPEJyKDLzCLEIIiLKc0IIbN58GaNHH0ZiYhoAQFdXAV9fD0yd2hwlShSpMURaJb8+v7kyXGGV1aKqLIKIiHLtxYsUDB9+ADt3Xpdizs6W2Lq1Gxo3LitjZiQnFkKF1dsTJ7IIIiL6aAoFcO7cm1thPj518OOP7WFmVgxGFlOusQ2wsOPEiUREecLCwhBbtnSFtbUxdu7sgY0bO7MIIrYIERFR8RQaGgMTE32ULfumP0nz5uVx//5XMDHRlzEzKkzYIlQYcQZpIqJcE0Jg7doLqFt3Lfr33weVSn1MEIsgehsLocKGM0gTEeXa06dJ6NJlB0aMOIjk5HQcP34f69YFy50WFWK8NVbYvDtajDNIExHlSEDAHfj47EdUVKIUGzHCDf3715YxKyrsWAgVJqG7gOe33jzmaDEiog9KSUnHlClHsXz5OSlmbW2MDRs6oWPHKjJmRkUBC6HC5O3WIKuqLIKIiD7g6tVo9O27F1evPpFinp4u8PPrAltbUxkzo6KChVBh8vbcQbwlRkT0Xv/99wINGqxHaqoSAGBgoItFiz7B6NENoaOjkDk7KirYWbqweHukGOcOIiL6oPLlS0r9f1xdbXDhwjCMHduIRRBphC1ChQFHihER5cqyZZ4oX94CEya4w9CQH2mkObYIyS2rxVV5W4yISE1SUhpGjDgAP78QtbiJiT6mTWvBIohyjd85cuIK80REHxQc/Bh9++5FaOgzbN16Fc2bl4OLi5XcaVExwRYhubAIIiJ6L6VShYULT6Nx418RGvoMAKBSCVy79uQDexLlHFuE5MAiiIjovSIi4uDtvQ8nTvwnxdzc7LBtW3dUrlxKxsyouGEhJId3Z49mEUREJNm58zqGDz+AFy9SAAAKBTB5cjPMmtUS+vq6MmdHxQ0LITm8PV8QiyAiIgBAQkIqxow5jE2bLksxR0dzbNnSFR4eTvIlRsUaCyE5cb4gIiJJaqoSf/55V3rs5VUDq1d/BktLIxmzouKOnaWJiKhQsLY2xqZNXWBuboDNm7vg99+7swiifMcWoYL29gzSRERaLDw8FiYmeihT5s2aYJ984oL//vsaJUsaypgZaRO2CBWU0F3AxmqcQZqItJ4QAps2haB27TUYNMgfQgi151kEUUFiIVRQzvoCz2+pxziDNBFpmdjYZPTuvQc+PvuRmJiGQ4duY+PGELnTIi3GW2MFJWOkmEIHsKz8ughiR2ki0iKBgffh7b0PDx/GSzEfnzro2bO6jFmRtmMhVBDe7hdkYgcMvClvPkREBSgtTQlf3+NYtOgMMu6CWVoaYu3az9GzZw15kyOtx0Iov3FleSLSYrduxaBv3724eDFSirVq5YTNm7uibFlzGTMjeo2FUH57dxZp9gsiIi0RHh6LevXWIjk5HQCgp6eDefNaY8IEd+joKGTOjug1dpbOb5xFmoi0lLOzJbp1qwYAqFKlFP75Zwi++aYpiyAqVNgiVFA4izQRaaFVqzqgfHkLTJvWAsbGenKnQ5TJR7UIpaSk5FUexRMnTyQiLZGSko5x445g167ranELC0PMm9eGRRAVWhoXQiqVCnPmzIGDgwNMTU0RHh4OAJgxYwZ+/fXXPE+wSOLkiUSkRa5ejUbDhuuxfPk5DBt2ABERcXKnRJRjGhdCc+fOhZ+fHxYtWgR9fX0pXrNmTfzyyy95mlyRxckTiUgLqFQCK1b8gwYN1uPq1ScAgOTkV7hw4bHMmRHlnMaF0ObNm7Fu3Tr07dsXurq6Urx27dq4devWe/bUIm9PnmhVlZ2kiajYiYxMQIcOW/H11wFITVUCAFxdbXDhwjB07VpN5uyIck7jztKPHj1CxYoVM8VVKhVevXqVJ0kVG5w8kYiKof37b2HIkP8hJualFBs3rjHmz28DQ0OOwaGiRePv2OrVq+PUqVMoX768Wnz37t2oW7duniVGRESFS1JSGiZM+BNr1wZLMTs7U/j5dUG7di4yZkaUexoXQr6+vhgwYAAePXoElUqFvXv3IjQ0FJs3b8aBAwfyI0ciIioE4uNTsWfPm1buLl2qYv36jrC2NpYxK6KPo3Efoc6dO+N///sfjh49ChMTE/j6+uLmzZv43//+h08++SQ/ciQiokLAzs4Mv/zSEcbGeli/viP27u3FIoiKPIUQGUvgaYf4+HhYWFggLi4O5ub5sM7N22uLmToAwx/m/WsQERWAiIg4mJjow8rKSC3+5EkSbGxMZMqKtFV+fX5r3CLk7OyMZ8+eZYq/ePECzs7OeZJUkfb22mKcO4iIiqidO6+jVq01GD78AN79e5lFEBUnGhdC9+/fh1KpzBRPTU3Fo0ecRVltbTHOHURERUx8fCp8fP6Al9duvHiRgt27b2Dbtqtyp0WUb3LcWdrf31/6f0BAACwsLKTHSqUSx44dg5OTU54mV6RxbTEiKmKCgiLQt+9e3Lv3Qop5edVAhw6V5EuKKJ/luBDq0qULAEChUGDAgAFqz+np6cHJyQlLlizJ0+SKHK4tRkRFUHq6CvPmncScOSehVL6+DWZmpo9VqzqgX79aUCi4WjwVXzkuhFQqFQCgQoUK+Pfff2FtbZ1vSRVZ7B9EREVMeHgs+vXbi6CgNwM73N0d8dtvXVGhgqWMmREVDI3nEbp3715+5FE8sH8QERUhd+48R716a5GQkAYA0NVVwNfXA1OnNkeJEhp3ISUqknI1F3pSUhJOnDiBBw8eIC0tTe25sWPH5kliRRr7BxFREeDiYok2bZzxxx+34Oxsia1bu6Fx47Jyp0VUoDQuhC5duoQOHTrg5cuXSEpKgpWVFWJiYmBsbAwbGxsWQkRERYRCocD69R1RvrwF5sxpBTMzA7lTIipwGrd9jhs3Dh07dkRsbCyMjIzwzz//4L///oObmxt++OGH/MiRiIg+UlqaEpMnH8XBg2FqcWtrYyxf3p5FEGktjQuhkJAQTJgwATo6OtDV1UVqaiocHR2xaNEiTJ06NT9yJCKijxAaGoMmTX7FwoVnMGiQP6KjE+VOiajQ0LgQ0tPTg47O691sbGzw4MEDAICFhQUiIiLyNruihEPniaiQEUJg7doLqFt3LS5ejAQAxMYm48wZLf5dTfQOjfsI1a1bF//++y8qVaoEDw8P+Pr6IiYmBlu2bEHNmjXzI8fC7+31xQAOnSci2T19moQhQ/4Hf/9QKValSils29Yd9erZyZgZUeGicYvQ/PnzYWf3+odo3rx5sLS0xMiRI/H06VOsXbs2zxMsEt6ePwjg0HkiklVAwB3UqrVGrQgaObI+Ll4cziKI6B0atwjVr19f+r+NjQ2OHDmSpwkVSW/PH9RxF4fOE5EsUlLSMWXKUSxffk6KWVsbY8OGTujYsYqMmREVXnk2Y9bFixfx+eef59XhiibOH0REMnryJAkbN4ZIj9u3r4irV0eyCCJ6D40KoYCAAEycOBFTp05FeHg4AODWrVvo0qULGjRoIC3DQUREBa9cOQusXv0ZDAx08eOP7XHoUB/Y2prKnRZRoZbjW2O//vorhg4dCisrK8TGxuKXX37B0qVLMWbMGHh5eeHatWuoVq1afuZKRERviYxMgImJPszN38wB9MUXrmjWrBwcHS1kzIyo6Mhxi9CKFSuwcOFCxMTEYOfOnYiJicHPP/+Mq1evYs2aNdpbBHHYPBHJYP/+W6hVaw3Gjj2c6TkWQUQ5l+NC6O7du+jZsycAoFu3bihRogQWL16MsmW1eF0aDpsnogKWlJSGESMOoEuXHYiJeYlNmy5jz54bcqdFVGTl+NZYcnIyjI2NAbxen8bAwEAaRq+1OGyeiApQcPBj9OmzF2Fhz6RYly5V4eHhJF9SREWcRsPnf/nlF5iavu54l56eDj8/P1hbW6tto1WLrnLYPBEVAKVShR9+OIvp048jPf31oBRjYz2sWNEegwfXhUKhkDlDoqJLIYQQOdnQycnpgz9sCoVCGk2WU6tWrcLixYsRFRWF2rVr46effkLDhg2z3f7FixeYNm0a9u7di+fPn6N8+fJYvnw5OnTokKPXi4+Ph4WFBeLi4mBubq5RrpmsLfu6f5CpAzD84ccdi4goCxERcfD23ocTJ/6TYm5udti2rTsqVy4lY2ZEBStPP7/fkuMWofv37+fZi2bYsWMHxo8fjzVr1qBRo0ZYvnw5PD09ERoaChsbm0zbp6Wl4ZNPPoGNjQ12794NBwcH/PfffyhZsmSe50ZEJLewsGdo1OgXvHiRAgBQKIDJk5th1qyW0NfXlTk7ouJB45ml89LSpUsxdOhQDBw4EACwZs0aHDx4EBs2bMDkyZMzbb9hwwY8f/4cZ8+ehZ6eHoDXLVVERMVRxYpWaNTIAQEBd+HoaI4tW7qyPxBRHsuzmaU1lZaWhuDgYLRt2/ZNMjo6aNu2LYKCgrLcx9/fH02aNMGoUaNQpkwZ1KxZE/Pnz4dSqSyotN/gsHkiymc6Ogps3NgZw4bVw+XLI1gEEeUD2VqEYmJioFQqUaZMGbV4mTJlcOvWrSz3CQ8Px99//42+ffvi0KFDuHPnDr788ku8evUKM2fOzHKf1NRUpKamSo/j4+M/PnkOmyeiPJaersK8eSfRvHl5tG5dQYrb2Zlh7dqOMmZGVLzJemtMUyqVCjY2Nli3bh10dXXh5uaGR48eYfHixdkWQgsWLMDs2bPzNhEOmyeiPBQeHot+/fYiKOghHBzMcOXKSFhZGcmdFpFWkO3WmLW1NXR1dREdHa0Wj46Ohq2tbZb72NnZoXLlytDVfdNJsFq1aoiKikJaWlqW+0yZMgVxcXHSV0RExMcnz2HzRJQHhBDYvPky6tRZg6Cg1yNPo6IScfz4PZkzI9IeuSqE7t69i+nTp+OLL77AkydPAACHDx/G9evXc3wMfX19uLm54dixY1JMpVLh2LFjaNKkSZb7NG3aFHfu3FFb3DUsLAx2dnbQ19fPch8DAwOYm5urfX2Ut/sGcbV5Isql2Nhk9O69BwMG/IGEhNd/yDk7W+L06UHo3r26zNkRaQ+NC6ETJ07A1dUV586dw969e5GYmAgAuHz5cra3p7Izfvx4rF+/Hps2bcLNmzcxcuRIJCUlSaPI+vfvjylTpkjbjxw5Es+fP8dXX32FsLAwHDx4EPPnz8eoUaM0PY3cYd8gIsoDgYH3UavWGuzc+eaPRx+fOggJGY7GjbV42SIiGWjcR2jy5MmYO3cuxo8fDzOzN4VA69atsXLlSo2O5eXlhadPn8LX1xdRUVGoU6cOjhw5InWgfvDgAXR03tRqjo6OCAgIwLhx41CrVi04ODjgq6++wqRJkzQ9jdxh3yAi+ghpaUrMnHkcCxeeQcZUtiVLGmLdus/Rs2cNeZMj0lI5nlk6g6mpKa5evYoKFSrAzMwMly9fhrOzM+7fv4+qVasiJSUlv3LNEx81M2XGTNIA+wYRkcbCw2NRq9ZqJCW9AgC0bOmEzZu7cLV4ohzIr5mlNb41VrJkSURGRmaKX7p0CQ4ODnmSVKHEvkFE9JGcnS2xYkV76OnpYNGitjh2rD+LICKZaXxrrHfv3pg0aRJ27doFhUIBlUqFM2fOYOLEiejfv39+5Cg/9g0iolyIiXkJY2M9GBvrSbFBg+rCw8MJFStayZgZEWXQuEVo/vz5qFq1KhwdHZGYmIjq1aujRYsWcHd3x/Tp0/MjR/mxbxARaSgg4A5cXVfjm2/+VIsrFAoWQUSFiMZ9hDI8ePAA165dQ2JiIurWrYtKlSrldW75Ilf3GNk3iIhyKCUlHVOmHMXy5eek2IEDX+CzzyrLmBVR0Sf76vMZTp8+jWbNmqFcuXIoV65cniVSJLBvEBG9x9Wr0ejbdy+uXn0ixdq3rwg3N3sZsyKi99H41ljr1q1RoUIFTJ06FTdu3MiPnAoXLq5KRB+gUgmsWPEPGjRYLxVBBga6+PHH9jh0qA9sbU1lzpCIsqNxIfT48WNMmDABJ06cQM2aNVGnTh0sXrwYDx8+zI/85Pd2/yB2kiaid0RGJqBDh634+usApKYqAQCurja4cGEYxoxpBIVCIXOGRPQ+GhdC1tbWGD16NM6cOYO7d++iZ8+e2LRpE5ycnNC6dev8yFFeb68rxk7SRPSW0NAY1Kq1BgEBd6XYuHGNcf78UNSsaSNjZkSUUx+16GqFChUwefJkfP/993B1dcWJEyfyKq/Ch/2DiOgdFStaoXr10gAAOztTBAT0w9KlnjA01Lj7JRHJJNeF0JkzZ/Dll1/Czs4Offr0Qc2aNXHw4MG8zI2IqFDT1dXBli1d4e1dC1eujES7di5yp0REGtL4z5YpU6Zg+/btePz4MT755BOsWLECnTt3hrGxcX7kJy92lCai/6dUqvDDD2fRvHl5uLs7SvFy5SyweXNXGTMjoo+hcSF08uRJfPPNN+jVqxesra3zI6fCgx2liQhAREQcvL334cSJ/1ChQkmEhIyAubmB3GkRUR7QuBA6c+ZMfuRR+ITuAp7fevOYHaWJtNLOndcxfPgBvHjxekHp+/df4M8/76JHj+oyZ0ZEeSFHhZC/vz8+/fRT6Onpwd/f/73bdurUKU8Sk93brUFWVdlRmkjLxMenYuzYw9i06bIUc3Q0x5YtXeHh4SRfYkSUp3JUCHXp0gVRUVGwsbFBly5dst1OoVBAqVTmVW7yYWsQkVYLCopAv377EB4eK8W8vGpg9erPYGlpJGNmRJTXclQIqVSqLP9fbLE1iEgrpaerMG/eScyZcxJK5etlGM3M9LFqVQf061eLkyMSFUMaD5/fvHkzUlNTM8XT0tKwefPmPElKdpxEkUgr3b37HAsWnJaKIHd3R1y+PALe3rVZBBEVUxoXQgMHDkRcXFymeEJCAgYOHJgnSRUanESRSKtUqWKNRYs+ga6uArNnt8SJEz6oUMFS7rSIKB9pPGpMCJHlX0YPHz6EhYVFniRFRFQQYmOTYWysBwODN78Kx4xpiNatK3CJDCItkeNCqG7dulAoFFAoFGjTpg1KlHizq1KpxL1799C+fft8SZKIKK8FBt6Ht/c+9O5dA4sXt5PiCoWCRRCRFslxIZQxWiwkJASenp4wNTWVntPX14eTkxO6d++e5wkSEeWltDQlZs48joULz0AI4IcfgtC+fUW0aeMsd2pEJIMcF0IzZ84EADg5OcHLywuGhob5lhQRUX4IDY1Bnz57cfFipBRr1coJVaoU81nyiShbGvcRGjBgQH7kUXhwfTGiYkcIgXXrgjFuXACSk9MBAHp6Opg3rzUmTHCHjg5HhBFpqxwVQlZWVggLC4O1tTUsLS3fO4z0+fPneZacLLi+GFGx8vRpEoYM+R/8/UOlWJUqpbBtW3fUq2cnY2ZEVBjkqBBatmwZzMzMpP8X6/k0OIcQUbERGhqDli03ISoqUYqNHFkfP/zQDsbGejJmRkSFRY4Kobdvh/n4+ORXLoUL5xAiKvKcnS3h6GiOqKhEWFsbY8OGTujYsYrcaRFRIaLxhIoXL17E1atXpcf79+9Hly5dMHXqVKSlpeVpckREH0NPTxdbt3ZDt27VcPXqSBZBRJSJxoXQ8OHDERYWBgAIDw+Hl5cXjI2NsWvXLnz77bd5niARUU6oVAI//ngOly5FqsUrVSqFPXt6wdbWNJs9iUibaVwIhYWFoU6dOgCAXbt2wcPDA9u2bYOfnx/27NmT1/kREX1QZGQCOnTYiq++OoI+ffbi5ctXcqdEREWExoWQEEJagf7o0aPo0KEDAMDR0RExMTF5m11B49B5oiJn//5bqFVrDQIC7gIAbt2KweHDt2XOioiKCo3nEapfvz7mzp2Ltm3b4sSJE1i9ejUA4N69eyhTpkyeJ1igOHSeqMhISkrDhAl/Yu3aYClmZ2cKP78uaNfORcbMiKgo0bgQWr58Ofr27Ys//vgD06ZNQ8WKFQEAu3fvhru7e54nWKA4dJ6oSAgOfow+ffYiLOyZFOvSpSrWr+8Ia2tjGTMjoqJGIYQQeXGglJQU6OrqQk+vcM/NER8fDwsLC8TFxcHc3Fz9ybVlX98aM3UAhj+UJ0EiypZSqcLixWcxY8ZxpKe/vkVvbKyH5cs9MWRIveI9xxmRlnvv5/dH0LhFKENwcDBu3rwJAKhevTrq1auXZ0kREWXl1q0YtSLIzc0O27Z1R+XKpWTOjIiKKo0LoSdPnsDLywsnTpxAyZIlAQAvXrxAq1atsH37dpQuXTqvcywY7ChNVOjVqGGDOXNaYerUY5g8uRlmzWoJfX1dudMioiJM41FjY8aMQWJiIq5fv47nz5/j+fPnuHbtGuLj4zF27Nj8yLFgsKM0UaGTkJAqtf5k+OYbd5w/PxTz57dhEUREH03jQujIkSP4+eefUa1aNSlWvXp1rFq1CocPH87T5AoUO0oTFSpBQRGoU2ct5s49qRbX1dVB/fr2MmVFRMWNxoWQSqXKskO0np6eNL9QkcY1xohklZ6uwuzZgWjefCPCw2MxZ85JnD0bIXdaRFRMaVwItW7dGl999RUeP34sxR49eoRx48ahTZs2eZocEWmX8PBYtGixEbNmnYBS+XpAa+PGZWFnx+UxiCh/aFwIrVy5EvHx8XBycoKLiwtcXFxQoUIFxMfH46effsqPHPMfO0oTyUoIgc2bL6NOnTUICno9dYWurgKzZ7fEiRM+qFDBUt4EiajY0njUmKOjIy5evIhjx45Jw+erVauGtm3b5nlyBYYdpYlkExubjJEjD2LHjutSzNnZElu3dkPjxmVlzIyItIFGhdCOHTvg7++PtLQ0tGnTBmPGjMmvvAoWO0oTySI0NAaffLIFERHxUszHpw5+/LE9zMwMZMyMiLRFjguh1atXY9SoUahUqRKMjIywd+9e3L17F4sXL87P/AoWO0oTFajy5UuiZElDRETEw9LSEGvXfo6ePWvInRYRaZEc9xFauXIlZs6cidDQUISEhGDTpk34+eef8zM3IirmDA1LYNu27ujQoRKuXBnJIoiIClyOC6Hw8HAMGDBAetynTx+kp6cjMjIyXxIjouJFCIF164Jx48ZTtXjNmjY4eLAPypbNu7WDiIhyKseFUGpqKkxMTN7sqKMDfX19JCcn50tiRFR8PH2ahC5ddmD48APo02cPUlPT5U6JiAiAhp2lZ8yYAWNjY+lxWloa5s2bBwsLCym2dOnSvMuOiIq8gIA78PHZj6ioRADA5cvROHAgDN27V5c5MyIiDQqhFi1aIDQ0VC3m7u6O8PBw6bFCoci7zIioSEtJScfkyUexYsU5KWZtbYwNGzqhY8cqMmZGRPRGjguhwMDAfEyDiIqTq1ej0afPXly79kSKeXq6wM+vC2xtOUs0ERUeGk+oSESUHZVK4KefzmHSpKNITVUCAAwMdLFo0ScYPbohdHTYakxEhQsLISLKM1evRmP8+D+hUr1eJ8zV1QbbtnVHzZo2MmdGRJQ1jdcaIyLKTu3atpg6tRkAYNy4xjh/fiiLICIq1NgiRES59vLlKxgallC75eXr64F27VzQvHl5GTMjIsoZtggRUa4EBz9G3bprsWTJWbW4np4uiyAiKjJyVQidOnUK/fr1Q5MmTfDo0SMAwJYtW3D69Ok8TY6ICh+lUoWFC0+jceNfERb2DNOm/Y2LFznDPBEVTRoXQnv27IGnpyeMjIxw6dIlpKamAgDi4uIwf/78PE+QiAqPiIg4tGmzGZMnH0N6ugoAUKtWGZia6sucGRFR7mhcCM2dOxdr1qzB+vXroaenJ8WbNm2Kixcv5mlyRFR47Nx5HbVqrcGJE/8BABQKYMqUZjh7djAqVy4lc3ZERLmjcWfp0NBQtGjRIlPcwsICL168yIuciKgQiY9Pxdixh7Fp02Up5uhoji1busLDw0m+xIiI8oDGhZCtrS3u3LkDJycntfjp06fh7OycV3kRUSEQGhqDDh22ITw8Vop5edXAmjWfo2RJQxkzIyLKGxrfGhs6dCi++uornDt3DgqFAo8fP8bWrVsxceJEjBw5Mj9yzF+hu4DER3JnQVQolS1rjhIlXv+aMDPTx+bNXfD7791ZBBFRsaFxITR58mT06dMHbdq0QWJiIlq0aIEhQ4Zg+PDhGDNmTK6SWLVqFZycnGBoaIhGjRrh/PnzOdpv+/btUCgU6NKlS65eFwBw1vfN//XNcn8comLIxEQf27Z1Q8uWTrh8eQS8vWtzcWUiKlYUQgiRmx3T0tJw584dJCYmonr16jA1zd1Cijt27ED//v2xZs0aNGrUCMuXL8euXbsQGhoKG5vsZ6S9f/8+mjVrBmdnZ1hZWeGPP/7I0evFx8fDwsICcXFxMDc3B9aWfdMi1HEXULlHrs6DqKgTQmDLlito2tQRLi5WmZ5jAUREcsr0+Z1Hcj2hor6+PqpXr46GDRvmuggCgKVLl2Lo0KEYOHAgqlevjjVr1sDY2BgbNmzIdh+lUom+ffti9uzZedcvydSBRRBprdjYZPTuvQcDBvyBvn334tUrpdrzLIKIqLjSuLN0q1at3vtL8e+//87xsdLS0hAcHIwpU6ZIMR0dHbRt2xZBQUHZ7vfdd9/BxsYGgwcPxqlTp977GqmpqdJcR8DripKI3ggMvA9v7314+PD1z8a5c49w4EAYunatJnNmRET5T+NCqE6dOmqPX716hZCQEFy7dg0DBgzQ6FgxMTFQKpUoU6aMWrxMmTK4detWlvucPn0av/76K0JCQnL0GgsWLMDs2bM1yotIG6SlKeHrexyLFp1Bxg1yS0tDrFvXkUUQEWkNjQuhZcuWZRmfNWsWEhMTPzqh90lISIC3tzfWr18Pa2vrHO0zZcoUjB8/XnocHx8PR0fH/EqRqEgIDY1Bnz571ZbGaNXKCZs3d0XZsnl3752IqLDLs9Xn+/Xrh4YNG+KHH37I8T7W1tbQ1dVFdHS0Wjw6Ohq2traZtr979y7u37+Pjh07SjGV6vU0/yVKlEBoaChcXFzU9jEwMICBgYEmp0JUbAkhsG5dMMaNC0BycjoAQE9PB/PmtcaECe5qq8gTEWmDPCuEgoKCYGio2dwi+vr6cHNzw7Fjx6Qh8CqVCseOHcPo0aMzbV+1alVcvXpVLTZ9+nQkJCRgxYoVbOkh+oBLl6IwYsRB6XGVKqWwbVt31KtnJ2NWRETy0bgQ6tatm9pjIQQiIyNx4cIFzJgxQ+MExo8fjwEDBqB+/fpo2LAhli9fjqSkJAwcOBAA0L9/fzg4OGDBggUwNDREzZo11fYvWbIkAGSKE1Fm9erZYfz4xli69B+MHFkfP/zQDsbGeh/ekYiomNK4ELKwsFB7rKOjgypVquC7775Du3btNE7Ay8sLT58+ha+vL6KiolCnTh0cOXJE6kD94MED6OjkepQ/kVZLTU2Hvr6u2kjP+fPboH37ivjkE5f37ElEpB00mlBRqVTizJkzcHV1haWlZX7mlW/UJmSKDAAO9Hr9hKkDMPyhvMkR5aGrV6PRp89ejBxZH19+2UDudIiIPkqhmFBRV1cX7dq1Kz6rzHN5DSqGVCqBFSv+QYMG63Ht2hNMmPAnbtx4KndaRESFksa3xmrWrInw8HBUqFAhP/IpWGkJb/7fdI58eRDlkcjIBAwcuB8BAXelWKVKVu/Zg4hIu2nc+Wbu3LmYOHEiDhw4gMjISMTHx6t9FUlcXoOKgf37b6FWrTVqRdC4cY1x/vxQVK9eWsbMiIgKrxy3CH333XeYMGECOnToAADo1KmTWgfMjEUZlUpldocgonyQlJSGCRP+xNq1wVLMzs4Ufn5d0K4dO0QTEb1Pjguh2bNnY8SIETh+/Hh+5kNEGggLe4aOHX9HWNgzKdalS1WsX98R1tbGMmZGRFQ05LgQyhhc5uHhkW/JEJFmypQxQVra61ZYY2M9rFjRHoMH1+Vq8UREOaRRHyH+ciUqXCwsDPHbb13RqJEDLl0ajiFD6vHnlIhIAxqNGqtcufIHf8k+f/78oxIiouzt2nUdjRuXhaPjm4lNmzYth6CgwSyAiIhyQaNCaPbs2Zlmliai/Bcfn4qxYw9j06bLaNnSCUePekNX902DLosgIqLc0agQ6t27N2xsbPIrl4J1ex+Q+EjuLIg+KCgoAv367UN4eCwAIDDwPg4cCEPnzlVlzoyIqOjLcR+hYvcX57l5b/7PWaWpEEpPV2H27EA0b75RKoLMzPSxeXMXdOpURebsiIiKB41HjRUbaYlv/s9ZpamQCQ+PRb9+exEU9Gb9O3d3R/z2W1dUqFA01/kjIiqMclwIqVSq/MxDPpxVmgoRIQS2bLmC0aMPISEhDQCgq6uAr68Hpk5tjhIlNJ4MnoiI3kPjtcaIKP9cuPAYAwb8IT12drbE1q3d0LhxWfmSIiIqxvjnJVEh0qCBA4YPdwMA+PjUQUjIcBZBRET5iC1CRDJ69UqJEiV01AYjLFnSDh06VGKHaCKiAsAWISKZhIbGoHHjX7Fp02W1uImJPosgIqICwkKIqIAJIbB27QXUrbsWFy9GYsyYw7hzhzOyExHJQXtvjSVFAoZyJ0Ha5unTJAwZ8j/4+4dKMQcHMyQnv5IxKyIi7aW9hVAGTqZIBSQg4A58fPYjKurNHFYjRrhhyRJPGBvryZgZEZH2YiHEyRQpn6WkpGPKlKNYvvycFLO2NsaGDZ3QsSP7AhERyUm7CyFOpkj57M6d5+jWbQeuXn0ixdq3r4iNGzvD1tZUxsyIiAjQ9kKIKJ9ZWhri2bNkAICBgS4WL/4Eo0c3LH5r9xERFVEcNUaUj0qVMoafX2fUrl0GFy4Mw5gxjVgEEREVImwRIspD//tfKBo0cFC77fXJJy4IDq4AXV3+3UFEVNjwNzNRHkhKSsOIEQfQqdN2DBq0H0IItedZBBERFU787Uz0kYKDH6NevXVYuzYYAHD48B0cOBAmc1ZERJQTLISIckmpVGHhwtNo3PhXhIU9AwAYG+th/fqO+PzzyjJnR0REOcE+QkS5EBERB2/vfThx4j8p5uZmh23buqNy5VIyZkZERJpgIUSkoR07rmHEiIN48SIFAKBQAJMnN8OsWS2hr68rc3ZERKQJFkJEGvjnn4fo3XuP9NjR0RxbtnSFh4eTfEkREVGusY8QkQYaNy4Lb+9aAAAvrxq4fHkEiyAioiKMLUJE76FSCejoqE+AuHJlB3z2WSX06lWDkyMSERVxbBEiykZ4eCyaNduAnTuvq8XNzQ3g5VWTRRARUTHAFiGidwghsGXLFYwefQgJCWm4efMAmjQpC0dHC7lTIyKiPMYWIaK3xMYmo3fvPRgw4A8kJKQBAKysjKSFU4mIqHhhixDR/wsMvA9v7314+DBeivn41MGPP7aHmZmBjJkREVF+YSFEWi8tTQlf3+NYtOgMMpYIK1nSEOvWfY6ePWvImxwREeUrFkKk1cLDY9Gz5y5cvBgpxVq2dMLmzV3YJ4iISAuwjxBpNSOjEnjwIA4AoKeng0WL2uLYsf4sgoiItAQLIdJqdnZm+PXXTqha1Rr//DME33zTNNO8QUREVHzx1hhplaNHw1G3ri1KlTKWYp06VcGnn1aEnh7XCSMi0jZsESKtkJKSjnHjjuCTT7Zg+PADEBm9ov8fiyAiIu3EQoiKvatXo9Gw4XosX34OALBnz00cOXJH5qyIiKgwYCFExZZKJbBixT9o0GA9rl59AgAwMNDFjz+2R/v2FWXOjoiICgP2EaJiKTIyAQMH7kdAwF0p5upqg23buqNmTRsZMyMiosKEhRAVO/7+oRg82B8xMS+l2LhxjTF/fhsYGvJbnoiI3uCnAhUrZ848QOfO26XHtram2LSpC9q1c5ExKyIiKqzYR4iKFXd3R3TtWhUA0LlzFVy9OpJFEBERZYstQlSkCSGgULyZAFGhUGD9+o7o1KkKBgyorfYcERHRu9giREVWREQcWrfejAMHwtTipUoZw8enDosgIiL6ILYIUZG0c+d1DB9+AC9epOD69Se4cmUkbG1N5U6LiIiKGLYIUZESH58KH58/4OW1Gy9epAAADA1L4PHjBJkzIyKioogtQlRkBAVFoG/fvbh374UU8/KqgdWrP4OlpZF8iRERUZHFQogKvfR0FebOPYm5c09CqXy9RpiZmT5WreqAfv1qsS8QERHlGgshKtTu33+BPn32ICjooRRzd3fEb791RYUKljJmRkRExQH7CFGhpqOjwI0bTwEAuroKzJ7dEidO+LAIIiKiPMFCiAq1cuUssGbN53B2tsTp04Pg6+uBEiX4bUtERHlDuz9R9M3kzoDecerUf4iPT1WL9e5dE9evf4nGjcvKlBURERVXhaIQWrVqFZycnGBoaIhGjRrh/Pnz2W67fv16NG/eHJaWlrC0tETbtm3fu/17NZ2Ty4wpr6WlKTF58lF4ePhhzJjDmZ7nYqlERJQfZC+EduzYgfHjx2PmzJm4ePEiateuDU9PTzx58iTL7QMDA/HFF1/g+PHjCAoKgqOjI9q1a4dHjx5p9sImdkDlHnlwBvSxQkNj0KTJr1i48AyEADZvvow//7wrd1pERKQFFEIIIWcCjRo1QoMGDbBy5UoAgEqlgqOjI8aMGYPJkyd/cH+lUglLS0usXLkS/fv3/+D28fHxsLCwQNwyO5h//fij86fcE0Jg3bpgjBsXgOTkdACAnp4O5s1rjQkT3KGjw2HxRET0mvT5HRcHc3PzPDuurPcb0tLSEBwcjClTpkgxHR0dtG3bFkFBQTk6xsuXL/Hq1StYWVll+XxqaipSU9/0OYmPj/+4pClPPH2ahCFD/gd//1ApVqVKKWzb1h316tnJmBkREWkTWW+NxcTEQKlUokyZMmrxMmXKICoqKkfHmDRpEuzt7dG2bdssn1+wYAEsLCykL0dHx4/Omz5OQMAd1Kq1Rq0IGjmyPi5eHM4iiIiICpTsfYQ+xvfff4/t27dj3759MDQ0zHKbKVOmIC4uTvqKiIgo4CzpbadO/Yf27bciKioRAGBtbQx//974+efPYGysJ3N2RESkbWS9NWZtbQ1dXV1ER0erxaOjo2Fra/vefX/44Qd8//33OHr0KGrVqpXtdgYGBjAwMMiTfOnjNWtWDu3bV8SRI3fQvn1FbNzYmavGExGRbGRtEdLX14ebmxuOHTsmxVQqFY4dO4YmTZpku9+iRYswZ84cHDlyBPXr1y+IVCmPKBQKbNzYGT//3AGHDvVhEURERLKS/dbY+PHjsX79emzatAk3b97EyJEjkZSUhIEDBwIA+vfvr9aZeuHChZgxYwY2bNgAJycnREVFISoqComJiXKdAmUjKioRn322DceOhavFbW1NMXJkAy6WSkREspN9ljovLy88ffoUvr6+iIqKQp06dXDkyBGpA/WDBw+go/OmXlu9ejXS0tLQo4f6HEAzZ87ErFmzCjJ1eg9//1AMHuyPmJiXuHw5Cpcvj0CpUsZyp0VERKRG9nmEChrnEcpfSUlpmDDhT6xdGyzF7OxM8b//fQE3N3sZMyMioqKsWM4jRMVLcPBj9O27F6Ghz6RYly5VsX59R1hbszWIiIgKHxZC9NGUShV++OEspk8/jvR0FQDA2FgPK1a0x+DBddkXiIiICi0WQvRRHj6Mh7f3PgQG3pdibm522LatOypXLiVfYkRERDkg+6gxKtqSk1/h339fL3irUABTpjTD2bODWQQREVGRwEKIPkqlSqXw44+fwtHRHMePD8D8+W2gr68rd1pEREQ5wkKINHL+/CO8fPlKLTZwYB3cuDEKHh5O8iRFRESUSyyEKEfS01WYPTsQ7u6/YuLEP9WeUygUMDXVlykzIiKi3GMhRB8UHh6LFi02YtasE1AqBVavvoDjx+/JnRYREdFH46gxypYQAlu2XMHo0YeQkJAGANDVVcDX1wPNm5eXOTsiIqKPx0KIshQbm4yRIw9ix47rUszZ2RJbt3ZD48ZlZcyMiIgo77AQokxOnLgPb+99iIiIl2I+PnXw44/tYWZmIGNmREREeYuFEKk5ceI+WrXahIwV6CwtDbF27efo2bOGvIkRERHlA3aWJjXNmpVDixav+/+0auWEK1dGsggiIqJiiy1CpEZXVwdbtnTFrl038PXXjaGjw3XCiIio+GKLkBZ7+jQJ3bvvxJkzD9Tijo4WGD++CYsgIiIq9tgipKUCAu7Ax2c/oqIScfFiJC5fHgFzc3aEJiIi7cIWIS2TkpKOr78+gvbttyIqKhEAkJiYhrCwZzJnRkREVPDYIqRFrl6NRp8+e3Ht2hMp1r59RWzc2Bm2tqYyZkZERCQPFkJaQKUS+Omnc5g06ShSU5UAAAMDXSxe/AlGj24IhYJ9gYiISDuxECrmIiMTMHDgfgQE3JVirq422LatO2rWtJExMyIiIvmxj1Ax9/x5MgID70uPx41rjPPnh7IIIiIiAguhYq9GDRssXvwJbG1NERDQD0uXesLQkA2BREREAAuhYufy5SikpqarxUaPbogbN75Eu3YuMmVFRERUOLEQKiaUShUWLjyN+vXXY9q0v9WeUygUsLQ0kikzIiKiwouFUDEQERGHNm02Y/LkY0hPV2HJkiCcPv3gwzsSERFpOXYWKeJ27ryO4cMP4MWLFACAQgFMntwMDRs6yJwZERFR4cdCqIiKj0/F2LGHsWnTZSnm6GiOLVu6wsPDSb7EiIiIihAWQkVQUFAE+vXbh/DwWCnm5VUDq1d/xr5AREREGmAhVMQEBt5H27aboVQKAICZmT5WreqAfv1qcYZoIiIiDbGzdBHTtKkj3NzsAQDu7o64fHkEvL1rswgiIiLKBbYIFTF6errYurUbduy4hkmTmqFECdayREREucVCqBCLjU3G6NGHMX58Y6kVCAAqVrTCtGktZMyMSLsIIZCeng6lUil3KkTFmp6eHnR1dQv0NVkIFVKBgffh7b0PDx/GIzj4MS5eHA5jYz250yLSOmlpaYiMjMTLly/lToWo2FMoFChbtixMTU0L7DVZCBUyaWlK+Poex6JFZyBe94fGkydJuH79CRo04NxARAVJpVLh3r170NXVhb29PfT19dkfjyifCCHw9OlTPHz4EJUqVSqwliEWQoVIaGgM+vTZi4sXI6VYq1ZO2Ly5K8qWNZcxMyLtlJaWBpVKBUdHRxgbG8udDlGxV7p0ady/fx+vXr1iIaRNhBBYty4Y48YFIDn59YKpeno6mDevNSZMcIeODv8CJZKTjg4HJRAVBDlaXFkIyezp0yQMGfI/+PuHSrEqVUph27buqFfPTsbMiIiIij8WQjKLiIjHoUO3pccjR9bHDz+0Y8doIiKiAsD2XpnVq2eHuXNbwdraGP7+vfHzz5+xCCIiklFoaChsbW2RkJAgdyrFSlpaGpycnHDhwgW5U1HDQqiA3boVg1ev1OcimTjRHdevf4mOHavIlBURFTc+Pj5QKBRQKBTQ09NDhQoV8O233yIlJSXTtgcOHICHhwfMzMxgbGyMBg0awM/PL8vj7tmzBy1btoSFhQVMTU1Rq1YtfPfdd3j+/Hk+n1HBmTJlCsaMGQMzMzO5U8k3q1atgpOTEwwNDdGoUSOcP3/+vdu/evUK3333HVxcXGBoaIjatWvjyJEjatvMmjVL+p7L+Kpatar0vL6+PiZOnIhJkyblyznlFguhAqJSCaxY8Q/q1FmDuXNPqj2nq6sDGxsTmTIjouKqffv2iIyMRHh4OJYtW4a1a9di5syZatv89NNP6Ny5M5o2bYpz587hypUr6N27N0aMGIGJEyeqbTtt2jR4eXmhQYMGOHz4MK5du4YlS5bg8uXL2LJlS4GdV1paWr4d+8GDBzhw4AB8fHw+6jj5mePH2rFjB8aPH4+ZM2fi4sWLqF27Njw9PfHkyZNs95k+fTrWrl2Ln376CTdu3MCIESPQtWtXXLp0SW27GjVqIDIyUvo6ffq02vN9+/bF6dOncf369Xw5t1wRWiYuLk4AEHHL7ArsNR8/jheenlsEMEsAs4SOzmxx7tzDAnt9Isqd5ORkcePGDZGcnCx3KhobMGCA6Ny5s1qsW7duom7dutLjBw8eCD09PTF+/PhM+//4448CgPjnn3+EEEKcO3dOABDLly/P8vViY2OzzSUiIkL07t1bWFpaCmNjY+Hm5iYdN6s8v/rqK+Hh4SE99vDwEKNGjRJfffWVKFWqlGjZsqX44osvRK9evdT2S0tLE6VKlRKbNm0SQgihVCrF/PnzhZOTkzA0NBS1atUSu3btyjZPIYRYvHixqF+/vlosJiZG9O7dW9jb2wsjIyNRs2ZNsW3bNrVtsspRCCGuXr0q2rdvL0xMTISNjY3o16+fePr0qbTf4cOHRdOmTYWFhYWwsrISn332mbhz5857c/xYDRs2FKNGjZIeK5VKYW9vLxYsWJDtPnZ2dmLlypVqsW7duom+fftKj2fOnClq1679wddv1aqVmD59epbPve9nTvr8jov74Gtogp2l89n+/bcwZMj/EBPzZlbasWMbolatMjJmRUQf5bf6QFJUwb+uiS3QL3f9K65du4azZ8+ifPnyUmz37t149epVppYfABg+fDimTp2K33//HY0aNcLWrVthamqKL7/8MsvjlyxZMst4YmIiPDw84ODgAH9/f9ja2uLixYtQqVQa5b9p0yaMHDkSZ86cAQDcuXMHPXv2RGJiojQLcUBAAF6+fImuXbsCABYsWIDffvsNa9asQaVKlXDy5En069cPpUuXhoeHR5avc+rUKdSvX18tlpKSAjc3N0yaNAnm5uY4ePAgvL294eLigoYNG2ab44sXL9C6dWsMGTIEy5YtQ3JyMiZNmoRevXrh77//BgAkJSVh/PjxqFWrFhITE+Hr64uuXbsiJCQk22kb5s+fj/nz57/3/bpx4wbKlSuXKZ6Wlobg4GBMmTJFiuno6KBt27YICgrK9nipqakwNDRUixkZGWVq8bl9+zbs7e1haGiIJk2aYMGCBZnyaNiwIU6dOvXe/AsSC6F8kpSUhgkT/sTatcFSzNbWFJs2dUG7di4yZkZEHy0pCkh8JHcWH3TgwAGYmpoiPT0dqamp0NHRwcqVK6Xnw8LCYGFhATu7zFN16Ovrw9nZGWFhYQBef8A5OztDT0+zwRzbtm3D06dP8e+//8LKygoAULFiRY3PpVKlSli0aJH02MXFBSYmJti3bx+8vb2l1+rUqRPMzMyQmpqK+fPn4+jRo2jSpAkAwNnZGadPn8batWuzLYT++++/TIWQg4ODWrE4ZswYBAQEYOfOnWqF0Ls5zp07F3Xr1lUrWjZs2ABHR0eEhYWhcuXK6N69u9prbdiwAaVLl8aNGzdQs2bNLHMcMWIEevXq9d73y97ePst4TEwMlEolypRR/2O8TJkyuHXrVrbH8/T0xNKlS9GiRQu4uLjg2LFj2Lt3r9r6e40aNYKfnx+qVKmCyMhIzJ49G82bN8e1a9fU+lvZ29vjv//+e2/+BYmFUD4IDn6MPn32IizsmRTr3LkKfvmlE6ytOTstUZFnYlskXrdVq1ZYvXo1kpKSsGzZMpQoUSLTB29OiYw1fzQUEhKCunXrSkVQbrm5uak9LlGiBHr16oWtW7fC29sbSUlJ2L9/P7Zv3w7gdYvRy5cv8cknn6jtl5aWhrp162b7OsnJyZlaPpRKJebPn4+dO3fi0aNHSEtLQ2pqaqbZxt/N8fLlyzh+/HiW62bdvXsXlStXxu3bt+Hr64tz584hJiZGail78OBBtoWQlZXVR7+fmlqxYgWGDh2KqlWrQqFQwMXFBQMHDsSGDRukbT799FPp/7Vq1UKjRo1Qvnx57Ny5E4MHD5aeMzIyKlRr97EQymN//30Pnp6/IT399TezsbEeli/3xJAh9bhGEVFxkcvbUwXNxMREan3ZsGEDateujV9//VX6UKpcuTLi4uLw+PHjTC0IaWlpuHv3Llq1aiVte/r0abx69UqjViEjI6P3Pq+jo5OpyHr16lWW5/Kuvn37wsPDA0+ePMFff/0FIyMjtG/fHsDrW3IAcPDgQTg4qK/TaGBgkG0+1tbWiI2NVYstXrwYK1aswPLly+Hq6goTExN8/fXXmTpEv5tjYmIiOnbsiIULF2Z6nYxWuI4dO6J8+fJYv3497O3toVKpULNmzfd2tv6YW2PW1tbQ1dVFdHS0Wjw6Ohq2ttkX2qVLl8Yff/yBlJQUPHv2DPb29pg8eTKcnZ2z3adkyZKoXLky7ty5oxZ//vw5Spcu/d78CxJHjeWxpk0dUb366wvs5maHS5eGY+hQNxZBRCQrHR0dTJ06FdOnT0dycjIAoHv37tDT08OSJUsybb9mzRokJSXhiy++AAD06dMHiYmJ+Pnnn7M8/osXL7KM16pVCyEhIdkOry9dujQiIyPVYiEhITk6J3d3dzg6OmLHjh3YunUrevbsKRVp1atXh4GBAR48eICKFSuqfTk6OmZ7zLp16+LGjRtqsTNnzqBz587o168fateurXbL8H3q1auH69evw8nJKVMOJiYmePbsGUJDQzF9+nS0adMG1apVy1SEZWXEiBEICQl571d2t8b09fXh5uaGY8eOSTGVSoVjx45JtxDfx9DQEA4ODkhPT8eePXvQuXPnbLdNTEzE3bt3M916vXbt2ntb5Qpcnna9LgIKYtTYtWvRYtq0YyI1NT3fXoOI8l9xGzX26tUr4eDgIBYvXizFli1bJnR0dMTUqVPFzZs3xZ07d8SSJUuEgYGBmDBhgtr+3377rdDV1RXffPONOHv2rLh//744evSo6NGjR7ajyVJTU0XlypVF8+bNxenTp8Xdu3fF7t27xdmzZ4UQQhw5ckQoFAqxadMmERYWJnx9fYW5uXmmUWNfffVVlsefNm2aqF69uihRooQ4depUpudKlSol/Pz8xJ07d0RwcLD48ccfhZ+fX7bvm7+/v7CxsRHp6W9+f48bN044OjqKM2fOiBs3boghQ4YIc3Nztfc3qxwfPXokSpcuLXr06CHOnz8v7ty5I44cOSJ8fHxEenq6UCqVolSpUqJfv37i9u3b4tixY6JBgwYCgNi3b1+2OX6s7du3CwMDA+Hn5ydu3Lghhg0bJkqWLCmioqKkbby9vcXkyZOlx//884/Ys2ePuHv3rjh58qRo3bq1qFChgtpowQkTJojAwEBx7949cebMGdG2bVthbW0tnjx5ovb65cuXF5s3b84yNzlGjbEQ+qhjpYghQ/aLa9ei8yAzIipsilshJIQQCxYsEKVLlxaJiYlSbP/+/aJ58+bCxMREGBoaCjc3N7Fhw4Ysj7tjxw7RokULYWZmJkxMTEStWrXEd999997h8/fv3xfdu3cX5ubmwtjYWNSvX1+cO3dOet7X11eUKVNGWFhYiHHjxonRo0fnuBC6ceOGACDKly8vVCqV2nMqlUosX75cVKlSRejp6YnSpUsLT09PceLEiWxzffXqlbC3txdHjhyRYs+ePROdO3cWpqamwsbGRkyfPl3079//g4WQEEKEhYWJrl27ipIlSwojIyNRtWpV8fXXX0u5/vXXX6JatWrCwMBA1KpVSwQGBuZ7ISSEED/99JMoV66c0NfXFw0bNpSmM3j7fAYMGCA9DgwMlPIsVaqU8Pb2Fo8ePVLbx8vLS9jZ2Ql9fX3h4OAgvLy8Mk0FcPbsWVGyZEnx8uXLLPOSoxBSCJHLHnBFVHx8PCwsLBC3zA7mXz/O9XGCgiLQr98+hIfHolatMjh/fggMDNjliqg4SUlJwb1791ChQoVMHWip+Fq1ahX8/f0REBAgdyrFjpeXF2rXro2pU6dm+fz7fuakz++4OJibm+dZTuwjpKH0dBVmzw5E8+YbER7++l7uvXuxuHIl+gN7EhFRUTB8+HC0aNGCa43lsbS0NLi6umLcuHFyp6KGTRgaCA+PRb9+exEU9FCKubs74rffuqJCBUsZMyMiorxSokQJTJs2Te40ih19fX1Mnz5d7jQyYSGUA0IIbNlyBaNHH0JCwushjbq6Cvj6emDq1OYoUYINa0REREURC6EPiI1NxsiRB7Fjx5sF4pydLbF1azc0blxWxsyIiIjoY7EQ+oCbN2Owa9ebOSV8fOrgxx/bw8ws+wm5iKh40bIxJUSykeNnjfd0PsDd3RHTpjVHyZKG2LmzBzZu7MwiiEhLZEzOV5iWAyAqzjJm1NbV1S2w12SL0Dvu3YtFuXIW0NV9UyPOmNECw4e7wcEh74brEVHhp6uri5IlS+LJkycAAGNjY84ST5RPVCoVnj59CmNjY5QoUXDlCQuh/yeEwLp1wRg3LgAzZ3pg0qRm0nN6erosgoi0VMb6SxnFEBHlHx0dHZQrV65A/+BgIQTg6dMkDBnyP/j7hwIApk8/jnbtXFC3rt0H9iSi4k6hUMDOzg42NjZZLgZKRHlHX18fOjoF22unUBRCq1atwuLFixEVFYXatWvjp59+QsOGDbPdfteuXZgxYwbu37+PSpUqYeHChejQoUOuXjsg4A58fPYjKipRig0ZUhdVqljn6nhEVDzp6uoWaL8FIioYsneW3rFjB8aPH4+ZM2fi4sWLqF27Njw9PbNthj579iy++OILDB48GJcuXUKXLl3QpUsXXLt2TaPXTXmli6+/PoL27bdKRZC1tTH8/Xtj9erPYWys99HnRkRERIWb7GuNNWrUCA0aNMDKlSsBvO4s5ejoiDFjxmDy5MmZtvfy8kJSUhIOHDggxRo3bow6depgzZo1H3y9jLVKqtkOx82oN7e+2reviI0bO8PW1jQPzoqIiIjyUrFcaywtLQ3BwcFo27atFNPR0UHbtm0RFBSU5T5BQUFq2wOAp6dntttn52bU6yUxDAx08eOP7XHoUB8WQURERFpG1j5CMTExUCqVKFOmjFq8TJkyuHXrVpb7REVFZbl9VFRUltunpqYiNTVVehwXF5fxDKpXL41ff+2M6tVLc3E9IiKiQiw+Ph5A3k+6WCg6S+enBQsWYPbs2Vk8sww3bgBNmkwo8JyIiIgod549ewYLC4s8O56shZC1tTV0dXURHR2tFo+Ojpbm7niXra2tRttPmTIF48ePlx6/ePEC5cuXx4MHD/L0jSTNxcfHw9HREREREXl6v5dyh9ej8OC1KDx4LQqPuLg4lCtXDlZWVnl6XFkLIX19fbi5ueHYsWPo0qULgNedpY8dO4bRo0dnuU+TJk1w7NgxfP3111Lsr7/+QpMmTbLc3sDAAAYGmZfEsLCw4Dd1IWFubs5rUYjwehQevBaFB69F4ZHX8wzJfmts/PjxGDBgAOrXr4+GDRti+fLlSEpKwsCBAwEA/fv3h4ODAxYsWAAA+Oqrr+Dh4YElS5bgs88+w/bt23HhwgWsW7dOztMgIiKiIkj2QsjLywtPnz6Fr68voqKiUKdOHRw5ckTqEP3gwQO16s/d3R3btm3D9OnTMXXqVFSqVAl//PEHatasKdcpEBERUREleyEEAKNHj872VlhgYGCmWM+ePdGzZ89cvZaBgQFmzpyZ5e0yKli8FoULr0fhwWtRePBaFB75dS1kn1CRiIiISC6yL7FBREREJBcWQkRERKS1WAgRERGR1mIhRERERFqrWBZCq1atgpOTEwwNDdGoUSOcP3/+vdvv2rULVatWhaGhIVxdXXHo0KECyrT40+RarF+/Hs2bN4elpSUsLS3Rtm3bD1470oymPxsZtm/fDoVCIU18Sh9P02vx4sULjBo1CnZ2djAwMEDlypX5uyqPaHotli9fjipVqsDIyAiOjo4YN24cUlJSCijb4uvkyZPo2LEj7O3toVAo8Mcff3xwn8DAQNSrVw8GBgaoWLEi/Pz8NH9hUcxs375d6Ovriw0bNojr16+LoUOHipIlS4ro6Ogstz9z5ozQ1dUVixYtEjdu3BDTp08Xenp64urVqwWcefGj6bXo06ePWLVqlbh06ZK4efOm8PHxERYWFuLhw4cFnHnxpOn1yHDv3j3h4OAgmjdvLjp37lwwyRZzml6L1NRUUb9+fdGhQwdx+vRpce/ePREYGChCQkIKOPPiR9NrsXXrVmFgYCC2bt0q7t27JwICAoSdnZ0YN25cAWde/Bw6dEhMmzZN7N27VwAQ+/bte+/24eHhwtjYWIwfP17cuHFD/PTTT0JXV1ccOXJEo9ctdoVQw4YNxahRo6THSqVS2NvbiwULFmS5fa9evcRnn32mFmvUqJEYPnx4vuapDTS9Fu9KT08XZmZmYtOmTfmVolbJzfVIT08X7u7u4pdffhEDBgxgIZRHNL0Wq1evFs7OziItLa2gUtQaml6LUaNGidatW6vFxo8fL5o2bZqveWqbnBRC3377rahRo4ZazMvLS3h6emr0WsXq1lhaWhqCg4PRtm1bKaajo4O2bdsiKCgoy32CgoLUtgcAT0/PbLennMnNtXjXy5cv8erVqzxfYE8b5fZ6fPfdd7CxscHgwYMLIk2tkJtr4e/vjyZNmmDUqFEoU6YMatasifnz50OpVBZU2sVSbq6Fu7s7goODpdtn4eHhOHToEDp06FAgOdMbefX5XShmls4rMTExUCqV0vIcGcqUKYNbt25luU9UVFSW20dFReVbntogN9fiXZMmTYK9vX2mb3TSXG6ux+nTp/Hrr78iJCSkADLUHrm5FuHh4fj777/Rt29fHDp0CHfu3MGXX36JV69eYebMmQWRdrGUm2vRp08fxMTEoFmzZhBCID09HSNGjMDUqVMLImV6S3af3/Hx8UhOToaRkVGOjlOsWoSo+Pj++++xfft27Nu3D4aGhnKno3USEhLg7e2N9evXw9raWu50tJ5KpYKNjQ3WrVsHNzc3eHl5Ydq0aVizZo3cqWmdwMBAzJ8/Hz///DMuXryIvXv34uDBg5gzZ47cqVEuFasWIWtra+jq6iI6OlotHh0dDVtb2yz3sbW11Wh7ypncXIsMP/zwA77//nscPXoUtWrVys80tYam1+Pu3bu4f/8+OnbsKMVUKhUAoESJEggNDYWLi0v+Jl1M5eZnw87ODnp6etDV1ZVi1apVQ1RUFNLS0qCvr5+vORdXubkWM2bMgLe3N4YMGQIAcHV1RVJSEoYNG4Zp06apLRJO+Su7z29zc/MctwYBxaxFSF9fH25ubjh27JgUU6lUOHbsGJo0aZLlPk2aNFHbHgD++uuvbLennMnNtQCARYsWYc6cOThy5Ajq169fEKlqBU2vR9WqVXH16lWEhIRIX506dUKrVq0QEhICR0fHgky/WMnNz0bTpk1x584dqRgFgLCwMNjZ2bEI+gi5uRYvX77MVOxkFKiCS3cWqDz7/NasH3fht337dmFgYCD8/PzEjRs3xLBhw0TJkiVFVFSUEEIIb29vMXnyZGn7M2fOiBIlSogffvhB3Lx5U8ycOZPD5/OIptfi+++/F/r6+mL37t0iMjJS+kpISJDrFIoVTa/HuzhqLO9oei0ePHggzMzMxOjRo0VoaKg4cOCAsLGxEXPnzpXrFIoNTa/FzJkzhZmZmfj9999FeHi4+PPPP4WLi4vo1auXXKdQbCQkJIhLly6JS5cuCQBi6dKl4tKlS+K///4TQggxefJk4e3tLW2fMXz+m2++ETdv3hSrVq3i8PkMP/30kyhXrpzQ19cXDRs2FP/884/0nIeHhxgwYIDa9jt37hSVK1cW+vr6okaNGuLgwYMFnHHxpcm1KF++vACQ6WvmzJkFn3gxpenPxttYCOUtTa/F2bNnRaNGjYSBgYFwdnYW8+bNE+np6QWcdfGkybV49eqVmDVrlnBxcRGGhobC0dFRfPnllyI2NrbgEy9mjh8/nuVnQMb7P2DAAOHh4ZFpnzp16gh9fX3h7OwsNm7cqPHrKoRgWx4RERFpp2LVR4iIiIhIEyyEiIiISGuxECIiIiKtxUKIiIiItBYLISIiItJaLISIiIhIa7EQIiIiIq3FQoiI1Pj5+aFkyZJyp5FrCoUCf/zxx3u38fHxQZcuXQokHyIq3FgIERVDPj4+UCgUmb7u3Lkjd2rw8/OT8tHR0UHZsmUxcOBAPHnyJE+OHxkZiU8//RQAcP/+fSgUCoSEhKhts2LFCvj5+eXJ62Vn1qxZ0nnq6urC0dERw4YNw/PnzzU6Dos2ovxVrFafJ6I32rdvj40bN6rFSpcuLVM26szNzREaGgqVSoXLly9j4MCBePz4MQICAj762NmtGv42CwuLj36dnKhRowaOHj0KpVKJmzdvYtCgQYiLi8OOHTsK5PWJ6MPYIkRUTBkYGMDW1lbtS1dXF0uXLoWrqytMTEzg6OiIL7/8EomJidke5/Lly2jVqhXMzMxgbm4ONzc3XLhwQXr+9OnTaN68OYyMjODo6IixY8ciKSnpvbkpFArY2trC3t4en376KcaOHYujR48iOTkZKpUK3333HcqWLQsDAwPUqVMHR44ckfZNS0vD6NGjYWdnB0NDQ5QvXx4LFixQO3bGrbEKFSoAAOrWrQuFQoGWLVsCUG9lWbduHezt7dVWdgeAzp07Y9CgQdLj/fv3o169ejA0NISzszNmz56N9PT0955niRIlYGtrCwcHB7Rt2xY9e/bEX3/9JT2vVCoxePBgVKhQAUZGRqhSpQpWrFghPT9r1ixs2rQJ+/fvl1qXAgMDAQARERHo1asXSpYsCSsrK3Tu3Bn3799/bz5ElBkLISIto6Ojgx9//BHXr1/Hpk2b8Pfff+Pbb7/Ndvu+ffuibNmy+PfffxEcHIzJkydDT08PAHD37l20b98e3bt3x5UrV7Bjxw6cPn0ao0eP1ignIyMjqFQqpKenY8WKFViyZAl++OEHXLlyBZ6enujUqRNu374NAPjxxx/h7++PnTt3IjQ0FFu3boWTk1OWxz1//jwA4OjRo4iMjMTevXszbdOzZ088e/YMx48fl2LPnz/HkSNH0LdvXwDAqVOn0L9/f3z11Ve4ceMG1q5dCz8/P8ybNy/H53j//n0EBARAX19fiqlUKpQtWxa7du3CjRs34Ovri6lTp2Lnzp0AgIkTJ6JXr15o3749IiMjERkZCXd3d7x69Qqenp4wMzPDqVOncObMGZiamqJ9+/ZIS0vLcU5EBBTL1eeJtN2AAQOErq6uMDExkb569OiR5ba7du0SpUqVkh5v3LhRWFhYSI/NzMyEn59flvsOHjxYDBs2TC126tQpoaOjI5KTk7Pc593jh4WFicqVK4v69esLIYSwt7cX8+bNU9unQYMG4ssvvxRCCDFmzBjRunVroVKpsjw+ALFv3z4hhBD37t0TAMSlS5fUthkwYIDo3Lmz9Lhz585i0KBB0uO1a9cKe3t7oVQqhRBCtGnTRsyfP1/tGFu2bBF2dnZZ5iCEEDNnzhQ6OjrCxMREGBoaSitpL126NNt9hBBi1KhRonv37tnmmvHaVapUUXsPUlNThZGRkQgICHjv8YlIHfsIERVTrVq1wurVq6XHJiYmAF63jixYsAC3bt1CfHw80tPTkZKSgpcvX8LY2DjTccaPH48hQ4Zgy5Yt0u0dFxcXAK9vm125cgVbt26VthdCQKVS4d69e6hWrVqWucXFxcHU1BQqlQopKSlo1qwZfvnlF8THx+Px48do2rSp2vZNmzbF5cuXAby+rfXJJ5+gSpUqaN++PT7//HO0a9fuo96rvn37YujQofj5559hYGCArVu3onfv3tDR0ZHO88yZM2otQEql8r3vGwBUqVIF/v7+SElJwW+//YaQkBCMGTNGbZtVq1Zhw4YNePDgAZKTk5GWloY6deq8N9/Lly/jzp07MDMzU4unpKTg7t27uXgHiLQXCyGiYsrExAQVK1ZUi92/fx+ff/45Ro4ciXnz5sHKygqnT5/G4MGDkZaWluUH+qxZs9CnTx8cPHgQhw8fxsyZM7F9+3Z07doViYmJGD58OMaOHZtpv3LlymWbm5mZGS5evAgdHR3Y2dnByMgIABAfH//B86pXrx7u3buHw4cP4+jRo+jVqxfatm2L3bt3f3Df7HTs2BFCCBw8eBANGjTAqVOnsGzZMun5xMREzJ49G926dcu0r6GhYbbH1dfXl67B999/j88++wyzZ8/GnDlzAADbt2/HxIkTsWTJEjRp0gRmZmZYvHgxzp079958ExMT4ebmplaAZigsHeKJigoWQkRaJDg4GCqVCkuWLJFaOzL6o7xP5cqVUblyZYwbNw5ffPEFNm7ciK5du6JevXq4ceNGpoLrQ3R0dLLcx9zcHPb29jhz5gw8PDyk+JkzZ9CwYUO17by8vODl5YUePXqgffv2eP78OaysrNSOl9EfR6lUvjcfQ0NDdOvWDVu3bsWdO3dQpUoV1KtXT3q+Xr16CA0N1fg83zV9+nS0bt0aI0eOlM7T3d0dX375pbTNuy06+vr6mfKvV68eduzYARsbG5ibm39UTkTajp2libRIxYoV8erVK/z0008IDw/Hli1bsGbNmmy3T05OxujRoxEYGIj//vsPZ86cwb///ivd8po0aRLOnj2L0aNHIyQkBLdv38b+/fs17iz9tm+++QYLFy7Ejh07EBoaismTJyMkJARfffUVAGDp0qX4/fffcevWLYSFhWHXrl2wtbXNchJIGxsbGBkZ4ciRI4iOjkZcXFy2r9u3b18cPHgQGzZskDpJZ/D19cXmzZsxe/ZsXL9+HTdv3sT27dsxffp0jc6tSZMmqFWrFubPnw8AqFSpEi5cuICAgACEhYVhxowZ+Pfff9X2cXJywpUrVxAaGoqYmBi8evUKffv2hbW1NTp37oxTp07h3r17CAwMxNixY/Hw4UONciLSenJ3UiKivJdVB9sMS5cuFXZ2dsLIyEh4enqKzZs3CwAiNjZWCKHemTk1NVX07t1bODo6Cn19fWFvby9Gjx6t1hH6/Pnz4pNPPhGmpqbCxMRE1KpVK1Nn57e921n6XUqlUsyaNUs4ODgIPT09Ubt2bXH48GHp+XXr1ok6deoIExMTYW5uLtq0aSMuXrwoPY+3OksLIcT69euFo6Oj0NHRER4eHtm+P0qlUtjZ2QkA4u7du5nyOnLkiHB3dxdGRkbC3NxcNGzYUKxbty7b85g5c6aoXbt2pvjvv/8uDAwMxIMHD0RKSorw8fERFhYWomTJkmLkyJFi8uTJavs9efJEen8BiOPHjwshhIiMjBT9+/cX1tbWwsDAQDg7O4uhQ4eKuLi4bHMioswUQgghbylGREREJA/eGiMiIiKtxUKIiIiItBYLISIiItJaLISIiIhIa7EQIiIiIq3FQoiIiIi0FgshIiIi0loshIiIiEhrsRAiIiIircVCiIiIiLQWCyEiIiLSWiyEiIiISGv9H9CnLyIgDxcAAAAAAElFTkSuQmCC", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "import matplotlib.pyplot as plt\n", "plt.figure()\n", "lw = 2\n", "plt.plot(\n", " fpr[1],\n", " tpr[1],\n", " color=\"darkorange\",\n", " lw=lw,\n", " label=\"ROC curve (area = %0.2f)\" % roc_auc[1],\n", ")\n", "plt.plot([0, 1], [0, 1], color=\"navy\", lw=lw, linestyle=\"--\")\n", "plt.xlim([0.0, 1.0])\n", "plt.ylim([0.0, 1.05])\n", "plt.xlabel(\"False Positive Rate\")\n", "plt.ylabel(\"True Positive Rate\")\n", "plt.title(\"Receiver operating characteristic example\")\n", "plt.legend(loc=\"lower right\")\n", "plt.show()" ] }, { "cell_type": "code", "execution_count": 74, "id": "54717414-d116-4dd3-9f70-2e823986dc46", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T16:14:49.010576Z", "iopub.status.busy": "2024-05-23T16:14:49.009575Z", "iopub.status.idle": "2024-05-23T16:14:49.211862Z", "shell.execute_reply": "2024-05-23T16:14:49.211110Z", "shell.execute_reply.started": "2024-05-23T16:14:49.010549Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "from sklearn.metrics import PrecisionRecallDisplay\n", "\n", "import matplotlib.pyplot as plt\n", "fig, ax = plt.subplots()\n", "display = PrecisionRecallDisplay.from_predictions(y_test, predictions_proba, name=\"ESMCrystal\", ax=ax)\n", "_ = display.ax_.set_title(\"2-class Precision-Recall curve\")\n", "plt.show()\n" ] }, { "cell_type": "code", "execution_count": 88, "id": "1fd70e82-cbd6-4399-b9ea-8bbace1f34c8", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T16:32:02.066439Z", "iopub.status.busy": "2024-05-23T16:32:02.065781Z", "iopub.status.idle": "2024-05-23T16:32:02.080057Z", "shell.execute_reply": "2024-05-23T16:32:02.079358Z", "shell.execute_reply.started": "2024-05-23T16:32:02.066411Z" } }, "outputs": [], "source": [ "import csv\n", "\n", "testdatacsvfilepath = \"Datasets/BCrystal_Balanced_Test_set/test.fasta\"\n", "testcsvfilepath = \"Datasets/BCrystal_Balanced_Test_set/y_test.csv\"\n", "\n", "X_test_B = []\n", "y_test_B = []\n", "\n", "with open(testdatacsvfilepath) as testcsvfile:\n", " csvreader = csv.reader(testcsvfile)\n", " for row in csvreader:\n", " #print(row)\n", " if '>' not in row[0]:\n", " X_test_B.append(row[0])\n", " else:\n", " pass\n", " \n", "with open(testcsvfilepath) as testcsvfile:\n", " csvreader = csv.reader(testcsvfile)\n", " for row in csvreader:\n", " #print(row)\n", " if '1' in row:\n", " y_test_B.append(1)\n", " elif '0' in row:\n", " y_test_B.append(0)\n", " else:\n", " pass" ] }, { "cell_type": "code", "execution_count": 89, "id": "f90b2e57-0b59-476f-9d52-6682f14d26f4", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T16:32:04.442180Z", "iopub.status.busy": "2024-05-23T16:32:04.441861Z", "iopub.status.idle": "2024-05-23T16:32:04.449920Z", "shell.execute_reply": "2024-05-23T16:32:04.449064Z", "shell.execute_reply.started": "2024-05-23T16:32:04.442155Z" } }, "outputs": [ { "data": { "text/plain": [ "('MRVLFIGDVFGQPGRRVLQNHLPTIRPQFDFVIVNMENSAGGFGMHRDAARGALEAGAGCLTLGNHAWHHKDIYPMLSEDTYPIVRPLNYADPGTPGVGWRTFDVNGEKLTVVNLLGRVFMEAVDNPFRTMDALLERDDLGTVFVDFHAEATSEKEAMGWHLAGRVAAVIGTHTHVPTADTRILKGGTAYQTDAGFTGPHDSIIGSAIEGPLQRFLTERPHRYGVAEGRAELNGVALHFEGGKATAAERYRFIED',\n", " 255,\n", " 1787,\n", " 1,\n", " 1787)" ] }, "execution_count": 89, "metadata": {}, "output_type": "execute_result" } ], "source": [ "X_test_B[0], len(X_test_B[0]), len(X_test_B), y_test_B[0], len(y_test_B)" ] }, { "cell_type": "code", "execution_count": 91, "id": "fa5ad501-5962-41b1-90ff-75797f683cc7", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T16:34:49.007157Z", "iopub.status.busy": "2024-05-23T16:34:49.006603Z", "iopub.status.idle": "2024-05-23T16:34:49.014441Z", "shell.execute_reply": "2024-05-23T16:34:49.013695Z", "shell.execute_reply.started": "2024-05-23T16:34:49.007131Z" } }, "outputs": [], "source": [ "testdatacsvfilepath = \"Datasets/SP_Final_set/FULL_SP.fasta\"\n", "testcsvfilepath = \"Datasets/SP_Final_set/SP_True_Label.csv\"\n", "\n", "X_test_S = []\n", "y_test_S = []\n", "\n", "with open(testdatacsvfilepath) as testcsvfile:\n", " csvreader = csv.reader(testcsvfile)\n", " for row in csvreader:\n", " #print(row)\n", " if '>' not in row[0]:\n", " X_test_S.append(row[0])\n", " else:\n", " pass\n", " \n", "with open(testcsvfilepath) as testcsvfile:\n", " csvreader = csv.reader(testcsvfile)\n", " for row in csvreader:\n", " #print(row)\n", " if '1' in row:\n", " y_test_S.append(1)\n", " elif '0' in row:\n", " y_test_S.append(0)\n", " else:\n", " pass" ] }, { "cell_type": "code", "execution_count": 92, "id": "b4827864-c65d-4b98-8e15-8c4582436251", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T16:34:53.261113Z", "iopub.status.busy": "2024-05-23T16:34:53.260231Z", "iopub.status.idle": "2024-05-23T16:34:53.267569Z", "shell.execute_reply": "2024-05-23T16:34:53.266331Z", "shell.execute_reply.started": "2024-05-23T16:34:53.261084Z" } }, "outputs": [ { "data": { "text/plain": [ "('MVDMQSLDEEDFSVSKSSDADAEFDIVIGNIEDIIMEDEFQHLQQSFMEKYYLEFDDSEENKLSYTPIFNEYIEILEKHLEQQLVERIPGFNMDAFTHSLKQHKDEVSGDILDMLLTFTDFMAFKEMFTDYRAEKEGRGLDLSTGLVVKSLNSSSASPLTPSMASQSI',\n", " 168,\n", " 237,\n", " 1,\n", " 237)" ] }, "execution_count": 92, "metadata": {}, "output_type": "execute_result" } ], "source": [ "X_test_S[0], len(X_test_S[0]), len(X_test_S), y_test_S[0], len(y_test_S)" ] }, { "cell_type": "code", "execution_count": 93, "id": "c2d8ff10-751e-4751-a8f4-0e2974b2c647", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T16:35:30.446881Z", "iopub.status.busy": "2024-05-23T16:35:30.445742Z", "iopub.status.idle": "2024-05-23T16:35:30.458738Z", "shell.execute_reply": "2024-05-23T16:35:30.457975Z", "shell.execute_reply.started": "2024-05-23T16:35:30.446841Z" } }, "outputs": [], "source": [ "testdatacsvfilepath = \"Datasets/TR_Final_set/FULL_TR.fasta\"\n", "testcsvfilepath = \"Datasets/TR_Final_set/TR_True_Label.csv\"\n", "\n", "X_test_T = []\n", "y_test_T = []\n", "\n", "with open(testdatacsvfilepath) as testcsvfile:\n", " csvreader = csv.reader(testcsvfile)\n", " for row in csvreader:\n", " #print(row)\n", " if '>' not in row[0]:\n", " X_test_T.append(row[0])\n", " else:\n", " pass\n", " \n", "with open(testcsvfilepath) as testcsvfile:\n", " csvreader = csv.reader(testcsvfile)\n", " for row in csvreader:\n", " #print(row)\n", " if '1' in row:\n", " y_test_T.append(1)\n", " elif '0' in row:\n", " y_test_T.append(0)\n", " else:\n", " pass" ] }, { "cell_type": "code", "execution_count": 94, "id": "1b96e5c6-26cc-4659-8f2b-3cff26550262", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T16:35:32.595636Z", "iopub.status.busy": "2024-05-23T16:35:32.595055Z", "iopub.status.idle": "2024-05-23T16:35:32.601035Z", "shell.execute_reply": "2024-05-23T16:35:32.600211Z", "shell.execute_reply.started": "2024-05-23T16:35:32.595610Z" } }, "outputs": [ { "data": { "text/plain": [ "('MRVLFIGDVFGQPGRRVLQNHLPTIRPQFDFVIVNMENSAGGFGMHRDAARGALEAGAGCLTLGNHAWHHKDIYPMLSEDTYPIVRPLNYADPGTPGVGWRTFDVNGEKLTVVNLLGRVFMEAVDNPFRTMDALLERDDLGTVFVDFHAEATSEKEAMGWHLAGRVAAVIGTHTHVPTADTRILKGGTAYQTDAGFTGPHDSIIGSAIEGPLQRFLTERPHRYGVAEGRAELNGVALHFEGGKATAAERYRFIED',\n", " 255,\n", " 1012,\n", " 1,\n", " 1012)" ] }, "execution_count": 94, "metadata": {}, "output_type": "execute_result" } ], "source": [ "X_test_T[0], len(X_test_T[0]), len(X_test_T), y_test_T[0], len(y_test_T)" ] }, { "cell_type": "code", "execution_count": 95, "id": "ca0e0eb8-6593-4e3f-9175-c557be423fbc", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T16:41:50.836321Z", "iopub.status.busy": "2024-05-23T16:41:50.835798Z", "iopub.status.idle": "2024-05-23T16:42:48.961559Z", "shell.execute_reply": "2024-05-23T16:42:48.960734Z", "shell.execute_reply.started": "2024-05-23T16:41:50.836248Z" } }, "outputs": [], "source": [ "predictions_D = pipeline(X_test)\n", "predictions_B = pipeline(X_test_B)\n", "predictions_S = pipeline(X_test_S)\n", "predictions_T = pipeline(X_test_T)\n", "\n", "preds_D = [int(x[\"label\"].split(\"_\")[1]) for x in predictions_D]\n", "predictions_D_proba = np.array([x[\"score\"] if int(x[\"label\"].split(\"_\")[1]) == 1 else (1.0 - x[\"score\"]) for x in predictions_D])\n", "predictions_D_probs = np.array([np.array([x[\"score\"], (1.0 - x[\"score\"])]) if int(x[\"label\"].split(\"_\")[1]) == 0 else np.array([(1.0 - x[\"score\"]), x[\"score\"]]) for x in predictions_D])\n", "\n", "preds_B = [int(x[\"label\"].split(\"_\")[1]) for x in predictions_B]\n", "predictions_B_proba = np.array([x[\"score\"] if int(x[\"label\"].split(\"_\")[1]) == 1 else (1.0 - x[\"score\"]) for x in predictions_B])\n", "predictions_B_probs = np.array([np.array([x[\"score\"], (1.0 - x[\"score\"])]) if int(x[\"label\"].split(\"_\")[1]) == 0 else np.array([(1.0 - x[\"score\"]), x[\"score\"]]) for x in predictions_B])\n", "\n", "preds_S = [int(x[\"label\"].split(\"_\")[1]) for x in predictions_S]\n", "predictions_S_proba = np.array([x[\"score\"] if int(x[\"label\"].split(\"_\")[1]) == 1 else (1.0 - x[\"score\"]) for x in predictions_S])\n", "predictions_S_probs = np.array([np.array([x[\"score\"], (1.0 - x[\"score\"])]) if int(x[\"label\"].split(\"_\")[1]) == 0 else np.array([(1.0 - x[\"score\"]), x[\"score\"]]) for x in predictions_S])\n", "\n", "preds_T = [int(x[\"label\"].split(\"_\")[1]) for x in predictions_T]\n", "predictions_T_proba = np.array([x[\"score\"] if int(x[\"label\"].split(\"_\")[1]) == 1 else (1.0 - x[\"score\"]) for x in predictions_T])\n", "predictions_T_probs = np.array([np.array([x[\"score\"], (1.0 - x[\"score\"])]) if int(x[\"label\"].split(\"_\")[1]) == 0 else np.array([(1.0 - x[\"score\"]), x[\"score\"]]) for x in predictions_T])" ] }, { "cell_type": "code", "execution_count": 96, "id": "29bbcc1a-5200-493c-9332-759e7828c1a4", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T16:42:48.963024Z", "iopub.status.busy": "2024-05-23T16:42:48.962793Z", "iopub.status.idle": "2024-05-23T16:42:48.982858Z", "shell.execute_reply": "2024-05-23T16:42:48.982175Z", "shell.execute_reply.started": "2024-05-23T16:42:48.963000Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " precision recall f1-score support\n", "\n", "non-crystallizable 0.73 0.97 0.83 1000\n", " crystallizable 0.94 0.60 0.73 898\n", "\n", " accuracy 0.79 1898\n", " macro avg 0.83 0.78 0.78 1898\n", " weighted avg 0.83 0.79 0.78 1898\n", "\n", "[[966 34]\n", " [362 536]]\n", "0.9467594654788418\n" ] } ], "source": [ "print(classification_report(y_test, preds_D, target_names=target_names))\n", "print(confusion_matrix(y_test, preds_D))\n", "print(roc_auc_score(y_test, predictions_D_proba))" ] }, { "cell_type": "code", "execution_count": 97, "id": "5dc3d485-773f-416a-a50b-97e862dece61", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T16:43:03.684115Z", "iopub.status.busy": "2024-05-23T16:43:03.683465Z", "iopub.status.idle": "2024-05-23T16:43:03.702174Z", "shell.execute_reply": "2024-05-23T16:43:03.701390Z", "shell.execute_reply.started": "2024-05-23T16:43:03.684072Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " precision recall f1-score support\n", "\n", "non-crystallizable 0.71 0.97 0.82 896\n", " crystallizable 0.94 0.60 0.73 891\n", "\n", " accuracy 0.78 1787\n", " macro avg 0.83 0.78 0.77 1787\n", " weighted avg 0.83 0.78 0.77 1787\n", "\n", "[[865 31]\n", " [360 531]]\n", "0.9465463163379829\n" ] } ], "source": [ "print(classification_report(y_test_B, preds_B, target_names=target_names))\n", "print(confusion_matrix(y_test_B, preds_B))\n", "print(roc_auc_score(y_test_B, predictions_B_proba))" ] }, { "cell_type": "code", "execution_count": 98, "id": "30b6d080-31d3-42ab-a77b-08dd624e610d", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T16:43:26.062153Z", "iopub.status.busy": "2024-05-23T16:43:26.061551Z", "iopub.status.idle": "2024-05-23T16:43:26.079462Z", "shell.execute_reply": "2024-05-23T16:43:26.078773Z", "shell.execute_reply.started": "2024-05-23T16:43:26.062128Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " precision recall f1-score support\n", "\n", "non-crystallizable 0.56 0.96 0.70 89\n", " crystallizable 0.95 0.54 0.69 148\n", "\n", " accuracy 0.70 237\n", " macro avg 0.75 0.75 0.70 237\n", " weighted avg 0.80 0.70 0.69 237\n", "\n", "[[85 4]\n", " [68 80]]\n", "0.9328120255086547\n" ] } ], "source": [ "print(classification_report(y_test_S, preds_S, target_names=target_names))\n", "print(confusion_matrix(y_test_S, preds_S))\n", "print(roc_auc_score(y_test_S, predictions_S_proba))" ] }, { "cell_type": "code", "execution_count": 99, "id": "814222c8-242d-420f-a314-e210fb7a5103", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T16:43:50.482502Z", "iopub.status.busy": "2024-05-23T16:43:50.481734Z", "iopub.status.idle": "2024-05-23T16:43:50.498927Z", "shell.execute_reply": "2024-05-23T16:43:50.498016Z", "shell.execute_reply.started": "2024-05-23T16:43:50.482475Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " precision recall f1-score support\n", "\n", "non-crystallizable 0.79 0.97 0.87 638\n", " crystallizable 0.93 0.55 0.69 374\n", "\n", " accuracy 0.82 1012\n", " macro avg 0.86 0.76 0.78 1012\n", " weighted avg 0.84 0.82 0.81 1012\n", "\n", "[[622 16]\n", " [167 207]]\n", "0.9562804888270497\n" ] } ], "source": [ "print(classification_report(y_test_T, preds_T, target_names=target_names))\n", "print(confusion_matrix(y_test_T, preds_T))\n", "print(roc_auc_score(y_test_T, predictions_T_proba))" ] }, { "cell_type": "code", "execution_count": 100, "id": "9ee5ca85-2247-40c6-90a6-4eaad2289469", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T16:50:49.394652Z", "iopub.status.busy": "2024-05-23T16:50:49.393793Z", "iopub.status.idle": "2024-05-23T16:50:49.421009Z", "shell.execute_reply": "2024-05-23T16:50:49.419986Z", "shell.execute_reply.started": "2024-05-23T16:50:49.394609Z" } }, "outputs": [], "source": [ "from sklearn.metrics import roc_curve, auc\n", "n_classes = 2\n", "\n", "# Compute ROC curve and ROC area for each class\n", "fpr_D = dict()\n", "tpr_D = dict()\n", "roc_auc_D = dict()\n", "for i in range(n_classes):\n", " fpr_D[i], tpr_D[i], _ = roc_curve(y_test, predictions_D_probs[:, i])\n", " roc_auc_D[i] = auc(fpr_D[i], tpr_D[i])\n", "\n", "# Compute micro-average ROC curve and ROC area\n", "fpr_D[\"micro\"], tpr_D[\"micro\"], _ = roc_curve(y_test, predictions_D_proba)\n", "roc_auc_D[\"micro\"] = auc(fpr_D[\"micro\"], tpr_D[\"micro\"])\n", "\n", "# Compute ROC curve and ROC area for each class\n", "fpr_B = dict()\n", "tpr_B = dict()\n", "roc_auc_B = dict()\n", "for i in range(n_classes):\n", " fpr_B[i], tpr_B[i], _ = roc_curve(y_test_B, predictions_B_probs[:, i])\n", " roc_auc_B[i] = auc(fpr_B[i], tpr_B[i])\n", "\n", "# Compute micro-average ROC curve and ROC area\n", "fpr_B[\"micro\"], tpr_B[\"micro\"], _ = roc_curve(y_test_B, predictions_B_proba)\n", "roc_auc_B[\"micro\"] = auc(fpr_B[\"micro\"], tpr_B[\"micro\"])\n", "\n", "# Compute ROC curve and ROC area for each class\n", "fpr_S = dict()\n", "tpr_S = dict()\n", "roc_auc_S = dict()\n", "for i in range(n_classes):\n", " fpr_S[i], tpr_S[i], _ = roc_curve(y_test_S, predictions_S_probs[:, i])\n", " roc_auc_S[i] = auc(fpr_S[i], tpr_S[i])\n", "\n", "# Compute micro-average ROC curve and ROC area\n", "fpr_S[\"micro\"], tpr_S[\"micro\"], _ = roc_curve(y_test_S, predictions_S_proba)\n", "roc_auc_S[\"micro\"] = auc(fpr_S[\"micro\"], tpr_S[\"micro\"])\n", "\n", "# Compute ROC curve and ROC area for each class\n", "fpr_T = dict()\n", "tpr_T = dict()\n", "roc_auc_T = dict()\n", "for i in range(n_classes):\n", " fpr_T[i], tpr_T[i], _ = roc_curve(y_test_T, predictions_T_probs[:, i])\n", " roc_auc_T[i] = auc(fpr_T[i], tpr_T[i])\n", "\n", "# Compute micro-average ROC curve and ROC area\n", "fpr_T[\"micro\"], tpr_T[\"micro\"], _ = roc_curve(y_test_T, predictions_T_proba)\n", "roc_auc_T[\"micro\"] = auc(fpr_T[\"micro\"], tpr_T[\"micro\"])" ] }, { "cell_type": "code", "execution_count": 114, "id": "4d9e97a8-38cc-4b7c-abfc-aa0b49d22ddd", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T17:21:05.329932Z", "iopub.status.busy": "2024-05-23T17:21:05.329083Z", "iopub.status.idle": "2024-05-23T17:21:05.547845Z", "shell.execute_reply": "2024-05-23T17:21:05.546666Z", "shell.execute_reply.started": "2024-05-23T17:21:05.329889Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "import matplotlib.pyplot as plt\n", "plt.figure()\n", "lw = 2\n", "\n", "plt.plot(\n", " fpr_D[1],\n", " tpr_D[1],\n", " color=\"navy\",\n", " lw=lw,\n", " label=\"DeepCrystal Test ROC curve (area = %0.2f)\" % roc_auc_D[1],\n", ")\n", "\n", "plt.plot(\n", " fpr_B[1],\n", " tpr_B[1],\n", " color=\"orange\",\n", " lw=lw,\n", " label=\"Balanced Test ROC curve (area = %0.2f)\" % roc_auc_B[1],\n", ")\n", "\n", "plt.plot(\n", " fpr_S[1],\n", " tpr_S[1],\n", " color=\"lightgreen\",\n", " lw=lw,\n", " label=\"SP ROC curve (area = %0.2f)\" % roc_auc_S[1],\n", ")\n", "\n", "plt.plot(\n", " fpr_T[1],\n", " tpr_T[1],\n", " color=\"red\",\n", " lw=lw,\n", " label=\"TR ROC curve (area = %0.2f)\" % roc_auc_T[1],\n", ")\n", "\n", "plt.plot([0, 1], [0, 1], color=\"grey\", lw=lw, linestyle=\"--\")\n", "plt.xlim([0.0, 1.0])\n", "plt.ylim([0.0, 1.05])\n", "plt.xlabel(\"False Positive Rate\")\n", "plt.ylabel(\"True Positive Rate\")\n", "plt.title(\"Receiver operating characteristic - ESMCrystal_t6_8M_v1\")\n", "plt.legend(loc=\"lower right\")\n", "plt.show()" ] }, { "cell_type": "code", "execution_count": 108, "id": "3d217ed3-a4ca-4140-aeb3-67ae71d7b157", "metadata": { "execution": { "iopub.execute_input": "2024-05-23T17:19:05.198574Z", "iopub.status.busy": "2024-05-23T17:19:05.197828Z", "iopub.status.idle": "2024-05-23T17:19:05.432540Z", "shell.execute_reply": "2024-05-23T17:19:05.431766Z", "shell.execute_reply.started": "2024-05-23T17:19:05.198548Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "from sklearn.metrics import PrecisionRecallDisplay\n", "\n", "import matplotlib.pyplot as plt\n", "fig, ax = plt.subplots()\n", "\n", "display = PrecisionRecallDisplay.from_predictions(y_test, predictions_D_proba, name=\"DeepCrystal Test\", ax=ax)\n", "display = PrecisionRecallDisplay.from_predictions(y_test_B, predictions_B_proba, name=\"Balanced Test\", ax=ax)\n", "display = PrecisionRecallDisplay.from_predictions(y_test_S, predictions_S_proba, name=\"SP Test\", ax=ax)\n", "display = PrecisionRecallDisplay.from_predictions(y_test_T, predictions_T_proba, name=\"TR Test\", ax=ax)\n", "\n", "_ = display.ax_.set_title(\"Precision-Recall curve - ESMCrystal_t6_8M_v1\")\n", "plt.show()" ] }, { "cell_type": "code", "execution_count": null, "id": "d90cc358-7f51-4c2e-962f-7f4069ffc694", "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "kernelspec": { "display_name": "Python 3 (ipykernel)", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.11.7" } }, "nbformat": 4, "nbformat_minor": 5 }