Token Classification/Monolingual/tner-bionlp2004/NER Project Using tner-bionlp 2004 Dataset (BERT-Base).ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Monolingual Named Entity Recognition (Bio NLP 2004 Dataset)\n",
    "\n",
    "Dataset Source: https://huggingface.co/datasets/tner/bionlp2004"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Install Necessary Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "#%pip install transformers datasets evaluate seqeval"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Import Necessary Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Git LFS initialized.\n"
     ]
    }
   ],
   "source": [
    "import os, sys\n",
    "os.environ['TOKENIZERS_PARALLELISM']='false'\n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "import datasets\n",
    "from datasets import load_dataset\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import sklearn\n",
    "from sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay\n",
    "\n",
    "import torch\n",
    "from torch.nn.functional import cross_entropy\n",
    "\n",
    "import transformers\n",
    "from transformers import AutoTokenizer, AutoModelForTokenClassification\n",
    "from transformers import DataCollatorForTokenClassification\n",
    "from transformers import Trainer, TrainingArguments\n",
    "\n",
    "import evaluate\n",
    "\n",
    "!git lfs install"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Display Library Versions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "          Python : 3.9.12\n",
      "           NumPy : 1.24.3\n",
      "          Pandas : 2.0.1\n",
      "           Torch : 2.0.0\n",
      "    Transformers : 4.28.1\n",
      "        Datasets : 2.11.0\n",
      "        Evaluate : 0.4.0\n",
      "         Sklearn : 1.2.2\n"
     ]
    }
   ],
   "source": [
    "print(\"Python :\".rjust(18), sys.version[0:6])\n",
    "print(\"NumPy :\".rjust(18), np.__version__)\n",
    "print(\"Pandas :\".rjust(18), pd.__version__)\n",
    "print(\"Torch :\".rjust(18), torch.__version__)\n",
    "print(\"Transformers :\".rjust(18), transformers.__version__)\n",
    "print(\"Datasets :\".rjust(18), datasets.__version__)\n",
    "print(\"Evaluate :\".rjust(18), evaluate.__version__)\n",
    "print(\"Sklearn :\".rjust(18), sklearn.__version__)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Ingest English Subset of Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Found cached dataset bionlp2004 (/Users/briandunn/.cache/huggingface/datasets/tner___bionlp2004/bionlp2004/1.0.0/9f41d3f0270b773c2762dee333ae36c29331e2216114a57081f77639fdb5e904)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "40988aa8daf74c4ea243af5ba49395a6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/3 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "DatasetDict({\n",
       "    train: Dataset({\n",
       "        features: ['tokens', 'tags'],\n",
       "        num_rows: 16619\n",
       "    })\n",
       "    validation: Dataset({\n",
       "        features: ['tokens', 'tags'],\n",
       "        num_rows: 1927\n",
       "    })\n",
       "    test: Dataset({\n",
       "        features: ['tokens', 'tags'],\n",
       "        num_rows: 3856\n",
       "    })\n",
       "})"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = load_dataset(\"tner/bionlp2004\")\n",
    "\n",
    "data"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'tokens': ['The',\n",
       "  'transcription',\n",
       "  'factor',\n",
       "  'nuclear',\n",
       "  'factor-kappa',\n",
       "  'B',\n",
       "  '(',\n",
       "  'NF-kappa',\n",
       "  'B',\n",
       "  ')',\n",
       "  'is',\n",
       "  'critical',\n",
       "  'for',\n",
       "  'the',\n",
       "  'inducible',\n",
       "  'expression',\n",
       "  'of',\n",
       "  'multiple',\n",
       "  'cellular',\n",
       "  'and',\n",
       "  'viral',\n",
       "  'genes',\n",
       "  'involved',\n",
       "  'in',\n",
       "  'inflammation',\n",
       "  'and',\n",
       "  'infection',\n",
       "  'including',\n",
       "  'interleukin-1',\n",
       "  '(',\n",
       "  'IL-1',\n",
       "  ')',\n",
       "  ',',\n",
       "  'IL-6',\n",
       "  ',',\n",
       "  'and',\n",
       "  'adhesion',\n",
       "  'molecules',\n",
       "  '.'],\n",
       " 'tags': [0,\n",
       "  3,\n",
       "  4,\n",
       "  3,\n",
       "  4,\n",
       "  4,\n",
       "  0,\n",
       "  3,\n",
       "  4,\n",
       "  0,\n",
       "  0,\n",
       "  0,\n",
       "  0,\n",
       "  0,\n",
       "  0,\n",
       "  0,\n",
       "  0,\n",
       "  0,\n",
       "  1,\n",
       "  2,\n",
       "  2,\n",
       "  2,\n",
       "  0,\n",
       "  0,\n",
       "  0,\n",
       "  0,\n",
       "  0,\n",
       "  0,\n",
       "  3,\n",
       "  0,\n",
       "  3,\n",
       "  0,\n",
       "  0,\n",
       "  3,\n",
       "  0,\n",
       "  0,\n",
       "  3,\n",
       "  4,\n",
       "  0]}"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "example = data['train'][12]\n",
    "\n",
    "example"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Display Feature Information About Each Feature"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tokens: \n",
      "Sequence(feature=Value(dtype='string', id=None), length=-1, id=None)\n",
      "\n",
      "tags: \n",
      "Sequence(feature=Value(dtype='int32', id=None), length=-1, id=None)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "for k, v in data[\"train\"].features.items():\n",
    "    print(f\"{k}: \\n{v}\\n\")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Define Tag Values & Conversions Between String & Integer Values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "List of tag values: \n",
      "['O', 'B-DNA', 'I-DNA', 'B-protein', 'I-protein', 'B-cell_type', 'I-cell_type', 'B-cell_line', 'I-cell_line', 'B-RNA', 'I-RNA']\n",
      "id2label: \n",
      "{0: 'O', 1: 'B-DNA', 2: 'I-DNA', 3: 'B-protein', 4: 'I-protein', 5: 'B-cell_type', 6: 'I-cell_type', 7: 'B-cell_line', 8: 'I-cell_line', 9: 'B-RNA', 10: 'I-RNA'}\n",
      "label2id: \n",
      "{'O': 0, 'B-DNA': 1, 'I-DNA': 2, 'B-protein': 3, 'I-protein': 4, 'B-cell_type': 5, 'I-cell_type': 6, 'B-cell_line': 7, 'I-cell_line': 8, 'B-RNA': 9, 'I-RNA': 10}\n",
      "Number of NER Tags: 11\n"
     ]
    }
   ],
   "source": [
    "label2id = {\n",
    "    \"O\": 0,\n",
    "    \"B-DNA\": 1,\n",
    "    \"I-DNA\": 2,\n",
    "    \"B-protein\": 3,\n",
    "    \"I-protein\": 4,\n",
    "    \"B-cell_type\": 5,\n",
    "    \"I-cell_type\": 6,\n",
    "    \"B-cell_line\": 7,\n",
    "    \"I-cell_line\": 8,\n",
    "    \"B-RNA\": 9,\n",
    "    \"I-RNA\": 10\n",
    "}\n",
    "\n",
    "id2label = {tag: idx for idx, tag in label2id.items()}\n",
    "\n",
    "tags = list(label2id.keys())\n",
    "\n",
    "NUM_OF_LABELS = len(tags)\n",
    "\n",
    "print(f\"List of tag values: \\n{tags}\")\n",
    "print(f\"id2label: \\n{id2label}\")\n",
    "print(f\"label2id: \\n{label2id}\")\n",
    "print(f\"Number of NER Tags: {NUM_OF_LABELS}\")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Basic Values/Constants"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "MODEL_CKPT = \"bert-base-cased\"\n",
    "\n",
    "MODEL_NAME = f\"{MODEL_CKPT}-finetuned-ner-bio_nlp_2004\"\n",
    "NUM_OF_EPOCHS = 3\n",
    "\n",
    "BATCH_SIZE = 16\n",
    "STRATEGY = \"epoch\"\n",
    "\n",
    "REPORTS_TO = \"tensorboard\"\n",
    "WEIGHT_DECAY = 0.01\n",
    "\n",
    "LR = 2e-5\n",
    "DEVICE = torch.device(\"cpu\")\n",
    "\n",
    "STEPS = 100"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Function to Tokenize & Align Inputs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenizer = AutoTokenizer.from_pretrained(MODEL_CKPT)\n",
    "\n",
    "def tokenize_and_align_labels(samples):\n",
    "    tokenized_inputs = tokenizer(samples[\"tokens\"], \n",
    "                                      truncation=True, \n",
    "                                      is_split_into_words=True)\n",
    "    \n",
    "    labels = []\n",
    "    \n",
    "    for idx, label in enumerate(samples[f\"tags\"]):\n",
    "        word_ids = tokenized_inputs.word_ids(batch_index=idx)\n",
    "        prev_word_idx = None\n",
    "        label_ids = []\n",
    "        for word_idx in word_ids: # set special tokens to -100\n",
    "            if word_idx is None or word_idx == prev_word_idx:\n",
    "                label_ids.append(-100)\n",
    "            else:\n",
    "                label_ids.append(label[word_idx])\n",
    "            prev_word_idx = word_idx\n",
    "        labels.append(label_ids)\n",
    "    \n",
    "    tokenized_inputs[\"labels\"] = labels\n",
    "    return tokenized_inputs"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Apply Above Function to Tokenize Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Loading cached processed dataset at /Users/briandunn/.cache/huggingface/datasets/tner___bionlp2004/bionlp2004/1.0.0/9f41d3f0270b773c2762dee333ae36c29331e2216114a57081f77639fdb5e904/cache-d1705a21fa18271d.arrow\n",
      "Loading cached processed dataset at /Users/briandunn/.cache/huggingface/datasets/tner___bionlp2004/bionlp2004/1.0.0/9f41d3f0270b773c2762dee333ae36c29331e2216114a57081f77639fdb5e904/cache-e035560a4a5185b0.arrow\n",
      "Loading cached processed dataset at /Users/briandunn/.cache/huggingface/datasets/tner___bionlp2004/bionlp2004/1.0.0/9f41d3f0270b773c2762dee333ae36c29331e2216114a57081f77639fdb5e904/cache-f0e4067f0b6b1413.arrow\n"
     ]
    }
   ],
   "source": [
    "encoded_ds = data.map(tokenize_and_align_labels, \n",
    "                       batched=True, \n",
    "                       remove_columns=\n",
    "                        [\n",
    "                            'tags', \n",
    "                            'tokens'\n",
    "                        ]\n",
    "                    )"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Define Data Collator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_collator = DataCollatorForTokenClassification(tokenizer=tokenizer)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Load Custom Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Some weights of the model checkpoint at bert-base-cased were not used when initializing BertForTokenClassification: ['cls.predictions.transform.dense.weight', 'cls.seq_relationship.bias', 'cls.predictions.bias', 'cls.predictions.transform.LayerNorm.weight', 'cls.predictions.transform.dense.bias', 'cls.seq_relationship.weight', 'cls.predictions.transform.LayerNorm.bias']\n",
      "- This IS expected if you are initializing BertForTokenClassification from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).\n",
      "- This IS NOT expected if you are initializing BertForTokenClassification from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).\n",
      "Some weights of BertForTokenClassification were not initialized from the model checkpoint at bert-base-cased and are newly initialized: ['classifier.weight', 'classifier.bias']\n",
      "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
     ]
    }
   ],
   "source": [
    "model = (AutoModelForTokenClassification.from_pretrained(\n",
    "    MODEL_CKPT,\n",
    "    num_labels=NUM_OF_LABELS,\n",
    "    id2label=id2label,\n",
    "    label2id=label2id\n",
    "    ).to(DEVICE))"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Create Compute Metrics Function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['O', 'B-DNA', 'I-DNA', 'B-protein', 'I-protein', 'B-cell_type', 'I-cell_type', 'B-cell_line', 'I-cell_line', 'B-RNA', 'I-RNA']\n"
     ]
    }
   ],
   "source": [
    "label_list = tags\n",
    "\n",
    "print(label_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "seqeval = evaluate.load(\"seqeval\")\n",
    "\n",
    "labels = [tags[i] for i in example[f\"tags\"]]\n",
    "\n",
    "def compute_metrics(eval_preds):\n",
    "    predictions, labels = eval_preds\n",
    "    predictions = np.argmax(predictions, \n",
    "                            axis=2)\n",
    "    \n",
    "    true_predictions = [\n",
    "        [tags[p] for (p, l) in zip(prediction, label) if l != -100]\n",
    "        for prediction, label in zip(predictions, labels)\n",
    "    ]\n",
    "    \n",
    "    true_labels = [\n",
    "        [tags[l] for (p, l) in zip(prediction, label) if l != -100]\n",
    "        for prediction, label in zip(predictions, labels)\n",
    "    ]\n",
    "    \n",
    "    results = seqeval.compute(predictions=true_predictions, \n",
    "                              references=true_labels)\n",
    "    \n",
    "    return results"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Define TrainingArguments"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "args = TrainingArguments(\n",
    "    MODEL_NAME,\n",
    "    log_level=\"error\",\n",
    "    logging_first_step=True,\n",
    "    learning_rate=LR,\n",
    "    num_train_epochs=NUM_OF_EPOCHS,\n",
    "    per_device_train_batch_size=BATCH_SIZE,\n",
    "    per_device_eval_batch_size=BATCH_SIZE,\n",
    "    evaluation_strategy=STRATEGY,\n",
    "    report_to=REPORTS_TO,\n",
    "    disable_tqdm=False,\n",
    "    logging_steps=STEPS,\n",
    "    weight_decay=WEIGHT_DECAY,\n",
    "    save_strategy=STRATEGY,\n",
    "    hub_private_repo=True,\n",
    "    push_to_hub=True\n",
    ")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Define Trainer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Cloning https://huggingface.co/DunnBC22/bert-base-cased-finetuned-ner-bio_nlp_2004 into local empty directory.\n"
     ]
    }
   ],
   "source": [
    "trainer = Trainer(model, \n",
    "                  args=args,\n",
    "                  data_collator=data_collator,\n",
    "                  compute_metrics=compute_metrics,\n",
    "                  tokenizer=tokenizer,\n",
    "                  train_dataset=encoded_ds[\"train\"],\n",
    "                  eval_dataset=encoded_ds[\"test\"]\n",
    "                  )"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Train Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/briandunn/Documents/deep_learning/dl/lib/python3.9/site-packages/transformers/optimization.py:391: FutureWarning: This implementation of AdamW is deprecated and will be removed in a future version. Use the PyTorch implementation torch.optim.AdamW instead, or set `no_deprecation_warning=True` to disable this warning\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "70e5329409d845c7a7b26c28916b9b66",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/3117 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'loss': 2.4176, 'learning_rate': 1.999358357394931e-05, 'epoch': 0.0}\n",
      "{'loss': 0.583, 'learning_rate': 1.9358357394931026e-05, 'epoch': 0.1}\n",
      "{'loss': 0.2819, 'learning_rate': 1.871671478986205e-05, 'epoch': 0.19}\n",
      "{'loss': 0.2311, 'learning_rate': 1.807507218479307e-05, 'epoch': 0.29}\n",
      "{'loss': 0.2207, 'learning_rate': 1.7433429579724095e-05, 'epoch': 0.38}\n",
      "{'loss': 0.1981, 'learning_rate': 1.679178697465512e-05, 'epoch': 0.48}\n",
      "{'loss': 0.1998, 'learning_rate': 1.6150144369586143e-05, 'epoch': 0.58}\n",
      "{'loss': 0.1983, 'learning_rate': 1.5508501764517164e-05, 'epoch': 0.67}\n",
      "{'loss': 0.1839, 'learning_rate': 1.4866859159448189e-05, 'epoch': 0.77}\n",
      "{'loss': 0.1834, 'learning_rate': 1.4225216554379211e-05, 'epoch': 0.87}\n",
      "{'loss': 0.1701, 'learning_rate': 1.3583573949310235e-05, 'epoch': 0.96}\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "bd93f2a63b6e4aabbb1f25b14b5f01f7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/241 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'eval_loss': 0.19273489713668823, 'eval_DNA': {'precision': 0.6152610441767068, 'recall': 0.7253787878787878, 'f1': 0.6657974793568013, 'number': 1056}, 'eval_RNA': {'precision': 0.6616541353383458, 'recall': 0.7457627118644068, 'f1': 0.7011952191235059, 'number': 118}, 'eval_cell_line': {'precision': 0.46697388632872505, 'recall': 0.608, 'f1': 0.5282363162467419, 'number': 500}, 'eval_cell_type': {'precision': 0.6997455470737913, 'recall': 0.7157730348776679, 'f1': 0.7076685537828101, 'number': 1921}, 'eval_protein': {'precision': 0.6602894693062719, 'recall': 0.783303730017762, 'f1': 0.716555334897996, 'number': 5067}, 'eval_overall_precision': 0.6498750624687656, 'eval_overall_recall': 0.7506349572846918, 'eval_overall_f1': 0.6966304173139766, 'eval_overall_accuracy': 0.9352316478121876, 'eval_runtime': 382.8236, 'eval_samples_per_second': 10.073, 'eval_steps_per_second': 0.63, 'epoch': 1.0}\n",
      "{'loss': 0.161, 'learning_rate': 1.294193134424126e-05, 'epoch': 1.06}\n",
      "{'loss': 0.1474, 'learning_rate': 1.2300288739172282e-05, 'epoch': 1.15}\n",
      "{'loss': 0.1478, 'learning_rate': 1.1658646134103306e-05, 'epoch': 1.25}\n",
      "{'loss': 0.143, 'learning_rate': 1.1017003529034329e-05, 'epoch': 1.35}\n",
      "{'loss': 0.1486, 'learning_rate': 1.0375360923965351e-05, 'epoch': 1.44}\n",
      "{'loss': 0.1441, 'learning_rate': 9.733718318896375e-06, 'epoch': 1.54}\n",
      "{'loss': 0.146, 'learning_rate': 9.092075713827398e-06, 'epoch': 1.64}\n",
      "{'loss': 0.1399, 'learning_rate': 8.450433108758422e-06, 'epoch': 1.73}\n",
      "{'loss': 0.138, 'learning_rate': 7.808790503689445e-06, 'epoch': 1.83}\n",
      "{'loss': 0.145, 'learning_rate': 7.167147898620469e-06, 'epoch': 1.92}\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "74efd89f46974cbbbc92a57c7f66696b",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/241 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'eval_loss': 0.1981300711631775, 'eval_DNA': {'precision': 0.6364372469635627, 'recall': 0.7443181818181818, 'f1': 0.6861632474901789, 'number': 1056}, 'eval_RNA': {'precision': 0.6408450704225352, 'recall': 0.7711864406779662, 'f1': 0.7000000000000002, 'number': 118}, 'eval_cell_line': {'precision': 0.4606896551724138, 'recall': 0.668, 'f1': 0.5453061224489797, 'number': 500}, 'eval_cell_type': {'precision': 0.7375615090213231, 'recall': 0.7022384174908901, 'f1': 0.7194666666666666, 'number': 1921}, 'eval_protein': {'precision': 0.6758880340481257, 'recall': 0.8148805999605289, 'f1': 0.7389047959914101, 'number': 5067}, 'eval_overall_precision': 0.6662350597609562, 'eval_overall_recall': 0.7722235049642115, 'eval_overall_f1': 0.7153245642177307, 'eval_overall_accuracy': 0.9364193315320131, 'eval_runtime': 416.2901, 'eval_samples_per_second': 9.263, 'eval_steps_per_second': 0.579, 'epoch': 2.0}\n",
      "{'loss': 0.1378, 'learning_rate': 6.525505293551493e-06, 'epoch': 2.02}\n",
      "{'loss': 0.1224, 'learning_rate': 5.8838626884825156e-06, 'epoch': 2.12}\n",
      "{'loss': 0.1119, 'learning_rate': 5.24222008341354e-06, 'epoch': 2.21}\n",
      "{'loss': 0.1238, 'learning_rate': 4.600577478344562e-06, 'epoch': 2.31}\n",
      "{'loss': 0.1148, 'learning_rate': 3.958934873275586e-06, 'epoch': 2.41}\n",
      "{'loss': 0.126, 'learning_rate': 3.3172922682066094e-06, 'epoch': 2.5}\n",
      "{'loss': 0.1144, 'learning_rate': 2.675649663137633e-06, 'epoch': 2.6}\n",
      "{'loss': 0.1148, 'learning_rate': 2.034007058068656e-06, 'epoch': 2.69}\n",
      "{'loss': 0.116, 'learning_rate': 1.3923644529996793e-06, 'epoch': 2.79}\n",
      "{'loss': 0.1182, 'learning_rate': 7.507218479307027e-07, 'epoch': 2.89}\n",
      "{'loss': 0.1116, 'learning_rate': 1.0907924286172603e-07, 'epoch': 2.98}\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "01ba862b1ef64d479e003eb3e1a2af2e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/241 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'eval_loss': 0.20655187964439392, 'eval_DNA': {'precision': 0.6619127516778524, 'recall': 0.7471590909090909, 'f1': 0.7019572953736656, 'number': 1056}, 'eval_RNA': {'precision': 0.589041095890411, 'recall': 0.7288135593220338, 'f1': 0.6515151515151515, 'number': 118}, 'eval_cell_line': {'precision': 0.4758522727272727, 'recall': 0.67, 'f1': 0.5564784053156145, 'number': 500}, 'eval_cell_type': {'precision': 0.7294117647058823, 'recall': 0.7100468505986466, 'f1': 0.7195990503824848, 'number': 1921}, 'eval_protein': {'precision': 0.6657656225155033, 'recall': 0.8263272153147819, 'f1': 0.7374075378654457, 'number': 5067}, 'eval_overall_precision': 0.6627781590040193, 'eval_overall_recall': 0.7805356730547218, 'eval_overall_f1': 0.71685309865875, 'eval_overall_accuracy': 0.936745944554965, 'eval_runtime': 406.1352, 'eval_samples_per_second': 9.494, 'eval_steps_per_second': 0.593, 'epoch': 3.0}\n",
      "{'train_runtime': 22380.9714, 'train_samples_per_second': 2.228, 'train_steps_per_second': 0.139, 'train_loss': 0.1687640313884942, 'epoch': 3.0}\n"
     ]
    }
   ],
   "source": [
    "train_results = trainer.train()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Push Model To Hub"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Several commits (2) will be pushed upstream.\n",
      "The progress bars may be unreliable.\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "75673ba6c7b243ce8f1e547fed811758",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Upload file pytorch_model.bin:   0%|          | 1.00/411M [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4b7ea617636f4b079e7b1e03d83dccff",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Upload file runs/Jul03_23-31-49_Brians-Mac-mini.local/events.out.tfevents.1688445114.Brians-Mac-mini.local.108…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "To https://huggingface.co/DunnBC22/bert-base-cased-finetuned-ner-bio_nlp_2004\n",
      "   63275fa..2863a28  main -> main\n",
      "\n",
      "To https://huggingface.co/DunnBC22/bert-base-cased-finetuned-ner-bio_nlp_2004\n",
      "   2863a28..bc39d78  main -> main\n",
      "\n"
     ]
    }
   ],
   "source": [
    "kwargs = {\n",
    "    \"tasks\": \"token-classification\",\n",
    "    \"tags\": ['token-classification'],\n",
    "}\n",
    "\n",
    "if args.push_to_hub:\n",
    "    trainer.push_to_hub('All DUNN!!!', **kwargs)\n",
    "else:\n",
    "    trainer.create_model_card(**kwargs)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Save & Log Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "To https://huggingface.co/DunnBC22/bert-base-cased-finetuned-ner-bio_nlp_2004\n",
      "   bc39d78..385a0ed  main -> main\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "***** train metrics *****\n",
      "  epoch                    =        3.0\n",
      "  train_loss               =     0.1688\n",
      "  train_runtime            = 6:13:00.97\n",
      "  train_samples_per_second =      2.228\n",
      "  train_steps_per_second   =      0.139\n"
     ]
    }
   ],
   "source": [
    "trainer.save_model()\n",
    "trainer.log_metrics(\"train\", train_results.metrics)\n",
    "trainer.save_metrics(\"train\", train_results.metrics)\n",
    "trainer.save_state()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Error Analysis"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Define Method to Apply to Validation Dataset (& Then Apply it)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "def forward_pass_with_label(batch):\n",
    "    # Convert dict of lists to list of dicts suitable for data collator\n",
    "    features = [dict(zip(batch, t)) for t in zip(*batch.values())]\n",
    "    # Pad inputs and labels and put all tensors on device\n",
    "    batch = data_collator(features)\n",
    "    input_ids = batch[\"input_ids\"].to(DEVICE)\n",
    "    attention_mask = batch[\"attention_mask\"].to(DEVICE)\n",
    "    labels = batch[\"labels\"].to(DEVICE)\n",
    "    with torch.no_grad():\n",
    "        # Pass data through model  \n",
    "        output = trainer.model(input_ids, \n",
    "                               attention_mask\n",
    "                               )\n",
    "        # Logit.size: [batch_size, sequence_length, classes]\n",
    "        predicted_label = torch.argmax(output.logits, \n",
    "                                       axis=-1\n",
    "                                       ).cpu().numpy()\n",
    "        \n",
    "    # Calculate loss per token after flattening batch dimension with view\n",
    "    loss = cross_entropy(output.logits.view(-1, 11), \n",
    "                         labels.view(-1), \n",
    "                         reduction=\"none\")\n",
    "    # Unflatten batch dimension and convert to numpy array\n",
    "    loss = loss.view(len(input_ids), -1).cpu().numpy()\n",
    "\n",
    "    return {\"loss\":loss, \"predicted_label\": predicted_label}"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Apply Above Function to Entire Evaluation Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "46020beab7044449b754f903ffff8cb6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Map:   0%|          | 0/1927 [00:00<?, ? examples/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "eval_set = encoded_ds['validation']\n",
    "\n",
    "eval_set = eval_set.map(forward_pass_with_label,\n",
    "                        batched=True,\n",
    "                        batch_size=32)\n",
    "\n",
    "eval_df = eval_set.to_pandas()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Clean Up Padding Tokens"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>input_ids</th>\n",
       "      <th>token_type_ids</th>\n",
       "      <th>attention_mask</th>\n",
       "      <th>labels</th>\n",
       "      <th>loss</th>\n",
       "      <th>predicted_label</th>\n",
       "      <th>input_tokens</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>[101, 15393, 118, 123, 5565, 2838, 1105, 151, ...</td>\n",
       "      <td>[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...</td>\n",
       "      <td>[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...</td>\n",
       "      <td>[IGN, B-DNA, IGN, IGN, I-DNA, O, O, B-protein,...</td>\n",
       "      <td>[0.0, 1.0764759, 0.0, 0.0, 0.9523054, 0.000231...</td>\n",
       "      <td>[O, B-protein, O, I-DNA, O, O, O, B-protein, I...</td>\n",
       "      <td>[[CLS], IL, -, 2, gene, expression, and, N, ##...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                           input_ids   \n",
       "0  [101, 15393, 118, 123, 5565, 2838, 1105, 151, ...  \\\n",
       "\n",
       "                                      token_type_ids   \n",
       "0  [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...  \\\n",
       "\n",
       "                                      attention_mask   \n",
       "0  [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...  \\\n",
       "\n",
       "                                              labels   \n",
       "0  [IGN, B-DNA, IGN, IGN, I-DNA, O, O, B-protein,...  \\\n",
       "\n",
       "                                                loss   \n",
       "0  [0.0, 1.0764759, 0.0, 0.0, 0.9523054, 0.000231...  \\\n",
       "\n",
       "                                     predicted_label   \n",
       "0  [O, B-protein, O, I-DNA, O, O, O, B-protein, I...  \\\n",
       "\n",
       "                                        input_tokens  \n",
       "0  [[CLS], IL, -, 2, gene, expression, and, N, ##...  "
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "id2label[-100] = \"IGN\"\n",
    "eval_df[\"input_tokens\"] = eval_df[\"input_ids\"].apply(\n",
    "    lambda x: tokenizer.convert_ids_to_tokens(x))\n",
    "eval_df[\"predicted_label\"] = eval_df[\"predicted_label\"].apply(\n",
    "    lambda x: [id2label[i] for i in x])\n",
    "eval_df[\"labels\"] = eval_df[\"labels\"].apply(\n",
    "    lambda x: [id2label[i] for i in x])\n",
    "eval_df['loss'] = eval_df.apply(\n",
    "    lambda x: x['loss'][:len(x['input_ids'])], axis=1)\n",
    "eval_df['predicted_label'] = eval_df.apply(\n",
    "    lambda x: x['predicted_label'][:len(x['input_ids'])], axis=1)\n",
    "eval_df.head(1)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Unwrap Each Token Within Sample Separately"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>input_ids</th>\n",
       "      <th>token_type_ids</th>\n",
       "      <th>attention_mask</th>\n",
       "      <th>labels</th>\n",
       "      <th>loss</th>\n",
       "      <th>predicted_label</th>\n",
       "      <th>input_tokens</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>15393</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>B-DNA</td>\n",
       "      <td>1.08</td>\n",
       "      <td>B-protein</td>\n",
       "      <td>IL</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5565</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>I-DNA</td>\n",
       "      <td>0.95</td>\n",
       "      <td>O</td>\n",
       "      <td>gene</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2838</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>O</td>\n",
       "      <td>0.00</td>\n",
       "      <td>O</td>\n",
       "      <td>expression</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1105</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>O</td>\n",
       "      <td>0.00</td>\n",
       "      <td>O</td>\n",
       "      <td>and</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>151</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>B-protein</td>\n",
       "      <td>0.01</td>\n",
       "      <td>B-protein</td>\n",
       "      <td>N</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>139</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>I-protein</td>\n",
       "      <td>0.01</td>\n",
       "      <td>I-protein</td>\n",
       "      <td>B</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>14915</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>O</td>\n",
       "      <td>0.00</td>\n",
       "      <td>O</td>\n",
       "      <td>activation</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  input_ids token_type_ids attention_mask     labels  loss predicted_label   \n",
       "0     15393              0              1      B-DNA  1.08       B-protein  \\\n",
       "0      5565              0              1      I-DNA  0.95               O   \n",
       "0      2838              0              1          O  0.00               O   \n",
       "0      1105              0              1          O  0.00               O   \n",
       "0       151              0              1  B-protein  0.01       B-protein   \n",
       "0       139              0              1  I-protein  0.01       I-protein   \n",
       "0     14915              0              1          O  0.00               O   \n",
       "\n",
       "  input_tokens  \n",
       "0           IL  \n",
       "0         gene  \n",
       "0   expression  \n",
       "0          and  \n",
       "0            N  \n",
       "0            B  \n",
       "0   activation  "
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "eval_df_tokens = eval_df.apply(pd.Series.explode)\n",
    "eval_df_tokens = eval_df_tokens.query(\"labels != 'IGN'\")\n",
    "eval_df_tokens[\"loss\"] = eval_df_tokens[\"loss\"].astype(float).round(2)\n",
    "eval_df_tokens.head(7)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### See Which Tokens Have Accumulated Most Loss in Evaluation Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>input_tokens</th>\n",
       "      <td>T</td>\n",
       "      <td>cells</td>\n",
       "      <td>B</td>\n",
       "      <td>E</td>\n",
       "      <td>P</td>\n",
       "      <td>(</td>\n",
       "      <td>)</td>\n",
       "      <td>,</td>\n",
       "      <td>and</td>\n",
       "      <td>N</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>604</td>\n",
       "      <td>605</td>\n",
       "      <td>622</td>\n",
       "      <td>252</td>\n",
       "      <td>227</td>\n",
       "      <td>736</td>\n",
       "      <td>736</td>\n",
       "      <td>1823</td>\n",
       "      <td>1321</td>\n",
       "      <td>419</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.476</td>\n",
       "      <td>0.394</td>\n",
       "      <td>0.373</td>\n",
       "      <td>0.622</td>\n",
       "      <td>0.67</td>\n",
       "      <td>0.199</td>\n",
       "      <td>0.199</td>\n",
       "      <td>0.079</td>\n",
       "      <td>0.103</td>\n",
       "      <td>0.303</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>sum</th>\n",
       "      <td>287.66</td>\n",
       "      <td>238.29</td>\n",
       "      <td>231.81</td>\n",
       "      <td>156.78</td>\n",
       "      <td>152.09</td>\n",
       "      <td>146.79</td>\n",
       "      <td>146.66</td>\n",
       "      <td>144.54</td>\n",
       "      <td>136.42</td>\n",
       "      <td>127.13</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   0       1       2       3       4       5       6       7   \n",
       "input_tokens       T   cells       B       E       P       (       )       ,  \\\n",
       "count            604     605     622     252     227     736     736    1823   \n",
       "mean           0.476   0.394   0.373   0.622    0.67   0.199   0.199   0.079   \n",
       "sum           287.66  238.29  231.81  156.78  152.09  146.79  146.66  144.54   \n",
       "\n",
       "                   8       9  \n",
       "input_tokens     and       N  \n",
       "count           1321     419  \n",
       "mean           0.103   0.303  \n",
       "sum           136.42  127.13  "
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(\n",
    "    eval_df_tokens.groupby(\"input_tokens\")[[\"loss\"]]\n",
    "    .agg([\"count\", \"mean\", \"sum\"])\n",
    "    .droplevel(level=0, axis=1)  # Get rid of multi-level columns\n",
    "    .sort_values(by=\"sum\", ascending=False)\n",
    "    .reset_index()\n",
    "    .round(3)\n",
    "    .head(10)\n",
    "    .T\n",
    ")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### See Which Label IDs Have Most Loss in Evaluation Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>labels</th>\n",
       "      <td>B-cell_type</td>\n",
       "      <td>B-DNA</td>\n",
       "      <td>B-RNA</td>\n",
       "      <td>I-cell_line</td>\n",
       "      <td>B-cell_line</td>\n",
       "      <td>I-cell_type</td>\n",
       "      <td>I-RNA</td>\n",
       "      <td>I-DNA</td>\n",
       "      <td>B-protein</td>\n",
       "      <td>I-protein</td>\n",
       "      <td>O</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>628</td>\n",
       "      <td>1260</td>\n",
       "      <td>131</td>\n",
       "      <td>940</td>\n",
       "      <td>505</td>\n",
       "      <td>696</td>\n",
       "      <td>189</td>\n",
       "      <td>2100</td>\n",
       "      <td>3029</td>\n",
       "      <td>2743</td>\n",
       "      <td>38603</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.639</td>\n",
       "      <td>0.586</td>\n",
       "      <td>0.554</td>\n",
       "      <td>0.544</td>\n",
       "      <td>0.515</td>\n",
       "      <td>0.376</td>\n",
       "      <td>0.373</td>\n",
       "      <td>0.366</td>\n",
       "      <td>0.356</td>\n",
       "      <td>0.321</td>\n",
       "      <td>0.081</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>sum</th>\n",
       "      <td>401.09</td>\n",
       "      <td>738.1</td>\n",
       "      <td>72.52</td>\n",
       "      <td>510.92</td>\n",
       "      <td>259.91</td>\n",
       "      <td>261.98</td>\n",
       "      <td>70.46</td>\n",
       "      <td>768.54</td>\n",
       "      <td>1079.32</td>\n",
       "      <td>880.98</td>\n",
       "      <td>3123.59</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 0      1      2            3            4            5    \n",
       "labels  B-cell_type  B-DNA  B-RNA  I-cell_line  B-cell_line  I-cell_type  \\\n",
       "count           628   1260    131          940          505          696   \n",
       "mean          0.639  0.586  0.554        0.544        0.515        0.376   \n",
       "sum          401.09  738.1  72.52       510.92       259.91       261.98   \n",
       "\n",
       "           6       7          8          9        10  \n",
       "labels  I-RNA   I-DNA  B-protein  I-protein        O  \n",
       "count     189    2100       3029       2743    38603  \n",
       "mean    0.373   0.366      0.356      0.321    0.081  \n",
       "sum     70.46  768.54    1079.32     880.98  3123.59  "
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(\n",
    "    eval_df_tokens.groupby(\"labels\")[[\"loss\"]] \n",
    "    .agg([\"count\", \"mean\", \"sum\"])\n",
    "    .droplevel(level=0, axis=1)\n",
    "    .sort_values(by=\"mean\", ascending=False)\n",
    "    .reset_index()\n",
    "    .round(3)\n",
    "    .fillna(0)\n",
    "    .T\n",
    ")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Create Function to Display Confusion Matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_confusion_matrix(y_preds, y_true, labels):\n",
    "    cm = confusion_matrix(y_true, y_preds, normalize=\"true\")\n",
    "    fig, ax = plt.subplots(figsize=(12, 12))\n",
    "    disp = ConfusionMatrixDisplay(confusion_matrix=cm, display_labels=labels)\n",
    "    disp.plot(cmap=\"Blues\", values_format=\".2f\", ax=ax, colorbar=False)\n",
    "    plt.title(\"Normalized confusion matrix\")\n",
    "    plt.show()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Display Confusion Matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x1200 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "eval_token_list = list(set(eval_df_tokens['labels']))\n",
    "\n",
    "plot_confusion_matrix(eval_df_tokens[\"labels\"], eval_df_tokens[\"predicted_label\"],\n",
    "                      eval_token_list)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Notes & Other Takeaways From This Project\n",
    "\n",
    "****\n",
    "- It looks like the model overpredicted B-DNA tokens in this model. Also, it looks like the model confused B-Protein tokens for B-Cell Type's.\n",
    "\n",
    "****"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Citations\n",
    "\n",
    "- Model Checkpoint\n",
    "\n",
    "    > @article{DBLP:journals/corr/abs-1810-04805, author = {Jacob Devlin and Ming{-}Wei Chang and Kenton Lee and Kristina Toutanova}, title = {{BERT:} Pre-training of Deep Bidirectional Transformers for Language Understanding}, journal = {CoRR}, volume = {abs/1810.04805}, year = {2018}, url = {http://arxiv.org/abs/1810.04805}, archivePrefix = {arXiv}, eprint = {1810.04805}, timestamp = {Tue, 30 Oct 2018 20:39:56 +0100}, biburl = {https://dblp.org/rec/journals/corr/abs-1810-04805.bib}, bibsource = {dblp computer science bibliography, https://dblp.org}}\n",
    "\n",
    "- Dataset\n",
    "\n",
    "    > @inproceedings{collier-kim-2004-introduction, title = \"Introduction to the Bio-entity Recognition Task at {JNLPBA}\", author = \"Collier, Nigel and Kim, Jin-Dong\", booktitle = \"Proceedings of the International Joint Workshop on Natural Language Processing in Biomedicine and its Applications ({NLPBA}/{B}io{NLP})\", month = aug # \" 28th and 29th\", year = \"2004\", address = \"Geneva, Switzerland\", publisher = \"COLING\", url = \"https://aclanthology.org/W04-1213\", pages = \"73--78\",}\n",
    "\n",
    "- Metric (SeqEval)\n",
    "\n",
    "    > @inproceedings{ramshaw-marcus-1995-text, title = \"Text Chunking using Transformation-Based Learning\", author = \"Ramshaw, Lance and Marcus, Mitch\", booktitle = \"Third Workshop on Very Large Corpora\", year = \"1995\", url = \"https://www.aclweb.org/anthology/W95-0107\",}\n",
    "    \n",
    "    > @misc{seqeval, title={{seqeval}: A Python framework for sequence labeling evaluation}, url={https://github.com/chakki-works/seqeval}, note={Software available from https://github.com/chakki-works/seqeval}, author={Hiroki Nakayama}, year={2018},}"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "dl",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.12"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}