koelectra.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from transformers import ElectraModel, ElectraTokenizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "from sklearn.metrics.pairwise import cosine_similarity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = ElectraModel.from_pretrained(\"monologg/koelectra-base-v3-discriminator\")  # KoELECTRA-Base-v3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenizer = ElectraTokenizer.from_pretrained(\"monologg/koelectra-base-v3-discriminator\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_embedding(text):\n",
    "    # 텍스트를 토큰화하고 토큰의 ID로 변환\n",
    "    inputs = tokenizer(text, return_tensors=\"pt\", padding=True, truncation=True, max_length=512)\n",
    "    \n",
    "    # 모델에 입력하여 hidden state 출력 (문맥적 임베딩)\n",
    "    with torch.no_grad():\n",
    "        outputs = model(**inputs)\n",
    "    \n",
    "    # 마지막 hidden state의 첫번째 토큰 (CLS 토큰)을 문서의 임베딩으로 사용\n",
    "    embeddings = outputs.last_hidden_state[:, 0, :].numpy()\n",
    "    return embeddings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [],
   "source": [
    "similarities = []\n",
    "correlation_matrix = []\n",
    "for i in range(1,5):\n",
    "    s = []\n",
    "    c = []\n",
    "    f = open(f'text{i}.txt', 'r')\n",
    "    text = f.read()\n",
    "    embedding = get_embedding(text)\n",
    "    for i in range(1, 5):\n",
    "        f = open(f'text{i}.txt', 'r')\n",
    "        text = f.read()\n",
    "        embedding1 = get_embedding(text)\n",
    "        s.append((1 + cosine_similarity(embedding, embedding1)[0][0])*50)\n",
    "        c.append(np.corrcoef(embedding1, embedding)[0][1])\n",
    "    similarities.append(s)\n",
    "    correlation_matrix.append(c)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100.00%\n",
      "63.41%\n",
      "61.84%\n",
      "60.31%\n",
      "63.41%\n",
      "100.00%\n",
      "99.42%\n",
      "99.30%\n",
      "61.84%\n",
      "99.42%\n",
      "100.00%\n",
      "99.38%\n",
      "60.31%\n",
      "99.30%\n",
      "99.38%\n",
      "100.00%\n"
     ]
    }
   ],
   "source": [
    "for i in similarities:\n",
    "    for j in i:\n",
    "        print(\"%.2f%%\" % (j))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "94.35%\n"
     ]
    }
   ],
   "source": [
    "t1 = \"뭐해 지금 바쁜데 빨리 나와\"\n",
    "t2 = \"서둘러서 나와\"\n",
    "print(\"%.2f%%\" % ((1 + cosine_similarity(get_embedding(t1), get_embedding(t2))[0][0])*50))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[[1.0, 0.6256223022937775, 0.6184166520833969],\n",
       " [0.6256223022937775, 0.9999999105930328, 0.9942547380924225],\n",
       " [0.6184166520833969, 0.9942547380924225, 1.0]]"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "similarities"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[[0.9999999999999999,\n",
       "  0.2457131405972726,\n",
       "  0.23123484063422808,\n",
       "  0.20041062874360838],\n",
       " [0.2457131405972726, 1.0, 0.9884048251182544, 0.9858419274933422],\n",
       " [0.23123484063422806, 0.9884048251182546, 1.0, 0.9874666151248032],\n",
       " [0.2004106287436084,\n",
       "  0.9858419274933423,\n",
       "  0.9874666151248032,\n",
       "  0.9999999999999999]]"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "correlation_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "피어슨 상관계수: 0.9858419274933422\n"
     ]
    }
   ],
   "source": [
    "correlation_matrix = np.corrcoef(embedding1, embedding2)\n",
    "print(f\"피어슨 상관계수: {correlation_matrix[0][1]}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import numpy as np\n",
    "\n",
    "ax = sns.heatmap(similarities, annot=True, cmap=\"YlGnBu\", square=True, cbar_kws={\"shrink\": 1.00}, linewidths=.5)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = sns.heatmap(correlation_matrix, annot=True, cmap=\"YlGnBu\", square=True, cbar_kws={\"shrink\": .82}, linewidths=.5)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "base",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.14"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}