hatakeyama-llm-team
/

bertopic_model

Model card Files Files and versions Community

kanhatakeyama commited on Mar 12, 2024

Commit

40c4ba5

verified ·

1 Parent(s): 5aa4afe

Upload train_topic_model.ipynb

Browse files

Files changed (1) hide show

train_topic_model.ipynb +361 -0

train_topic_model.ipynb ADDED Viewed

	@@ -0,0 +1,361 @@

+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "#!pip install bertopic\n",
+    "\n",
+    "# bertopicのmodelを作るscript"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/home/user/miniconda3/envs/ft/lib/python3.11/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
+      "  from .autonotebook import tqdm as notebook_tqdm\n"
+     ]
+    }
+   ],
+   "source": [
+    "from bertopic import BERTopic"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from datasets import load_dataset\n",
+    "streaming=True\n",
+    "dataset_list =[\n",
+    "    load_dataset('mc4', 'ja', split='train',streaming=streaming),\n",
+    "    load_dataset('oscar', 'unshuffled_deduplicated_ja', split='train',streaming=streaming),\n",
+    "    load_dataset('cc100', lang='ja', split='train',streaming=streaming),\n",
+    "    load_dataset(\"augmxnt/shisa-pretrain-en-ja-v1\",split=\"train\",streaming=streaming),\n",
+    "    load_dataset(\"hpprc/wikipedia-20240101\", split=\"train\",streaming=streaming),\n",
+    "]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "10000it [00:20, 482.63it/s]\n",
+      "10000it [00:19, 524.44it/s]\n",
+      "10000it [00:12, 778.96it/s]\n",
+      "10000it [00:25, 386.40it/s]\n",
+      "10000it [00:58, 171.79it/s]\n"
+     ]
+    }
+   ],
+   "source": [
+    "from tqdm import tqdm\n",
+    "docs=[]\n",
+    "#prepare data for training model\n",
+    "for dataset in dataset_list:\n",
+    "    cnt=0\n",
+    "    for record in tqdm(dataset):\n",
+    "        text=record[\"text\"]\n",
+    "        docs.append(text)\n",
+    "        cnt+=1\n",
+    "\n",
+    "        if cnt>10000:\n",
+    "            break\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "2024-03-12 08:37:19,823 - BERTopic - Embedding - Transforming documents to embeddings.\n",
+      "Batches: 100%|██████████| 1563/1563 [00:50<00:00, 30.79it/s] \n",
+      "2024-03-12 08:38:20,622 - BERTopic - Embedding - Completed ✓\n",
+      "2024-03-12 08:38:20,622 - BERTopic - Dimensionality - Fitting the dimensionality reduction algorithm\n",
+      "2024-03-12 08:38:59,566 - BERTopic - Dimensionality - Completed ✓\n",
+      "2024-03-12 08:38:59,567 - BERTopic - Cluster - Start clustering the reduced embeddings\n",
+      "huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...\n",
+      "To disable this warning, you can either:\n",
+      "\t- Avoid using `tokenizers` before the fork if possible\n",
+      "\t- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)\n",
+      "huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...\n",
+      "To disable this warning, you can either:\n",
+      "\t- Avoid using `tokenizers` before the fork if possible\n",
+      "\t- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)\n",
+      "huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...\n",
+      "To disable this warning, you can either:\n",
+      "\t- Avoid using `tokenizers` before the fork if possible\n",
+      "\t- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)\n",
+      "huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...\n",
+      "To disable this warning, you can either:\n",
+      "\t- Avoid using `tokenizers` before the fork if possible\n",
+      "\t- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)\n",
+      "huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...\n",
+      "To disable this warning, you can either:\n",
+      "\t- Avoid using `tokenizers` before the fork if possible\n",
+      "\t- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)\n",
+      "2024-03-12 08:46:25,241 - BERTopic - Cluster - Completed ✓\n",
+      "2024-03-12 08:46:25,242 - BERTopic - Representation - Extracting topics from clusters using representation models.\n",
+      "2024-03-12 08:47:25,876 - BERTopic - Representation - Completed ✓\n",
+      "2024-03-12 08:47:25,952 - BERTopic - Topic reduction - Reducing number of topics\n",
+      "2024-03-12 08:48:28,300 - BERTopic - Topic reduction - Reduced number of topics from 435 to 342\n"
+     ]
+    }
+   ],
+   "source": [
+    "\n",
+    "model_path=\"data/topic_model.bin\"\n",
+    "topic_model = BERTopic(language=\"japanese\", calculate_probabilities=True, verbose=True, nr_topics=\"20\")\n",
+    "topics, probs = topic_model.fit_transform(docs)\n",
+    "\n",
+    "\n",
+    "#topic_model=BERTopic.load(model_path)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "2024-03-12 08:48:42,599 - BERTopic - WARNING: When you use `pickle` to save/load a BERTopic model,please make sure that the environments in which you saveand load the model are **exactly** the same. The version of BERTopic,its dependencies, and python need to remain the same.\n",
+      "/home/user/miniconda3/envs/ft/lib/python3.11/site-packages/scipy/sparse/_index.py:143: SparseEfficiencyWarning: Changing the sparsity structure of a csr_matrix is expensive. lil_matrix is more efficient.\n",
+      "  self._set_arrayXarray(i, j, x)\n"
+     ]
+    }
+   ],
+   "source": [
+    "topic_model.save(model_path)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "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>Topic</th>\n",
+       "      <th>Count</th>\n",
+       "      <th>Name</th>\n",
+       "      <th>Representation</th>\n",
+       "      <th>Representative_Docs</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>-1</td>\n",
+       "      <td>22559</td>\n",
+       "      <td>-1_the_and_to_of</td>\n",
+       "      <td>[the, and, to, of, 送料無料, in, 12, 11, 10, また]</td>\n",
+       "      <td>[Створення сайту - Сторінка 419 - Форум\\nЧетве...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>0</td>\n",
+       "      <td>1585</td>\n",
+       "      <td>0_送料無料_サマータイヤ_代引不可_中古</td>\n",
+       "      <td>[送料無料, サマータイヤ, 代引不可, 中古, ブラック, diy, レディース, 工具,...</td>\n",
+       "      <td>[上品なスタイル 【5/1(土)クーポン&amp;ワンダフルデー 4本1台分!!】 215/45R1...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>1</td>\n",
+       "      <td>1209</td>\n",
+       "      <td>1_としあき_無念_name_投稿日</td>\n",
+       "      <td>[としあき, 無念, name, 投稿日, id, 16, 名前, 柳宗理, no, 11]</td>\n",
+       "      <td>[ハニーセレクト日曜昼の部テンプレセット髪型全然使ってなかったけど - ふたろぐばこ−二次元...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>2</td>\n",
+       "      <td>801</td>\n",
+       "      <td>2_ワンピース_5cm_レディース_着丈</td>\n",
+       "      <td>[ワンピース, 5cm, レディース, 着丈, 肩幅, 素材, 格安通販, シューズ, 袖丈...</td>\n",
+       "      <td>[非売品 入学式 セレモニー 秋冬 秋 他と被らない 冬 小さいサイズ スカート セット 卒...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>3</td>\n",
+       "      <td>799</td>\n",
+       "      <td>3_ベンジャミン_フランクリン_passion_thee</td>\n",
+       "      <td>[ベンジャミン, フランクリン, passion, thee, nベンジャミン, 全業種, ...</td>\n",
+       "      <td>[it's ok with me 意味\\t9\\n英語で「It's okay.（イッツオーケー...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>337</th>\n",
+       "      <td>336</td>\n",
+       "      <td>11</td>\n",
+       "      <td>336_abuse_you_counselling_emotional</td>\n",
+       "      <td>[abuse, you, counselling, emotional, addiction...</td>\n",
+       "      <td>[スピリチュアルカウンセリングは、魂の向上を目的とした、至高神からのヒーリングで魂を整えて頂...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>338</th>\n",
+       "      <td>337</td>\n",
+       "      <td>10</td>\n",
+       "      <td>337_京都の道_snorkeling_その１_中の池</td>\n",
+       "      <td>[京都の道, snorkeling, その１, 中の池, k7, silfra, 今だけ特別...</td>\n",
+       "      <td>[オアフ島（ホノルル） 福岡発 ◎今だけ無料で海の見える部屋へアップグレード！◎シェラトン・...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>339</th>\n",
+       "      <td>338</td>\n",
+       "      <td>10</td>\n",
+       "      <td>338_実印_いつ使う_件のレビュー例えば_いつ使うは</td>\n",
+       "      <td>[実印, いつ使う, 件のレビュー例えば, いつ使うは, しっかりした会社, 印鑑, 実印の...</td>\n",
+       "      <td>[冊子の「契約内容のお知らせ」ページをめくると、登録情報の変更シートがあります。\\n, 今回...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>340</th>\n",
+       "      <td>339</td>\n",
+       "      <td>10</td>\n",
+       "      <td>339_galaxy_s7_samsung_edge</td>\n",
+       "      <td>[galaxy, s7, samsung, edge, i9195i, s8, 3i9200...</td>\n",
+       "      <td>[ S8 PlusとS9 Plus - bajatyoutube.com\\n2019/0...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>341</th>\n",
+       "      <td>340</td>\n",
+       "      <td>10</td>\n",
+       "      <td>340_ﾟдﾟ_対価_労働_産業別組合</td>\n",
+       "      <td>[ﾟдﾟ, 対価, 労働, 産業別組合, 工会, 約款, union, 契約書, 規約, 労...</td>\n",
+       "      <td>[ただし、中小企業の事業主等、労働者以外でも業務の実態や災害の発生状況からみて、労働者に準じ...</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>342 rows × 5 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "     Topic  Count                                 Name  \\\n",
+       "0       -1  22559                     -1_the_and_to_of   \n",
+       "1        0   1585                0_送料無料_サマータイヤ_代引不可_中古   \n",
+       "2        1   1209                   1_としあき_無念_name_投稿日   \n",
+       "3        2    801                 2_ワンピース_5cm_レディース_着丈   \n",
+       "4        3    799         3_ベンジャミン_フランクリン_passion_thee   \n",
+       "..     ...    ...                                  ...   \n",
+       "337    336     11  336_abuse_you_counselling_emotional   \n",
+       "338    337     10          337_京都の道_snorkeling_その１_中の池   \n",
+       "339    338     10          338_実印_いつ使う_件のレビュー例えば_いつ使うは   \n",
+       "340    339     10           339_galaxy_s7_samsung_edge   \n",
+       "341    340     10                  340_ﾟдﾟ_対価_労働_産業別組合   \n",
+       "\n",
+       "                                        Representation  \\\n",
+       "0         [the, and, to, of, 送料無料, in, 12, 11, 10, また]   \n",
+       "1    [送料無料, サマータイヤ, 代引不可, 中古, ブラック, diy, レディース, 工具,...   \n",
+       "2       [としあき, 無念, name, 投稿日, id, 16, 名前, 柳宗理, no, 11]   \n",
+       "3    [ワンピース, 5cm, レディース, 着丈, 肩幅, 素材, 格安通販, シューズ, 袖丈...   \n",
+       "4    [ベンジャミン, フランクリン, passion, thee, nベンジャミン, 全業種, ...   \n",
+       "..                                                 ...   \n",
+       "337  [abuse, you, counselling, emotional, addiction...   \n",
+       "338  [京都の道, snorkeling, その１, 中の池, k7, silfra, 今だけ特別...   \n",
+       "339  [実印, いつ使う, 件のレビュー例えば, いつ使うは, しっかりした会社, 印鑑, 実印の...   \n",
+       "340  [galaxy, s7, samsung, edge, i9195i, s8, 3i9200...   \n",
+       "341  [ﾟдﾟ, 対価, 労働, 産業別組合, 工会, 約款, union, 契約書, 規約, 労...   \n",
+       "\n",
+       "                                   Representative_Docs  \n",
+       "0    [Створення сайту - Сторінка 419 - Форум\\nЧетве...  \n",
+       "1    [上品なスタイル 【5/1(土)クーポン&ワンダフルデー 4本1台分!!】 215/45R1...  \n",
+       "2    [ハニーセレクト日曜昼の部テンプレセット髪型全然使ってなかったけど - ふたろぐばこ−二次元...  \n",
+       "3    [非売品 入学式 セレモニー 秋冬 秋 他と被らない 冬 小さいサイズ スカート セット 卒...  \n",
+       "4    [it's ok with me 意味\\t9\\n英語で「It's okay.（イッツオーケー...  \n",
+       "..                                                 ...  \n",
+       "337  [スピリチュアルカウンセリングは、魂の向上を目的とした、至高神からのヒーリングで魂を整えて頂...  \n",
+       "338  [オアフ島（ホノルル） 福岡発 ◎今だけ無料で海の見える部屋へアップグレード！◎シェラトン・...  \n",
+       "339  [冊子の「契約内容のお知らせ」ページをめくると、登録情報の変更シートがあります。\\n, 今回...  \n",
+       "340  [ S8 PlusとS9 Plus - bajatyoutube.com\\n2019/0...  \n",
+       "341  [ただし、中小企業の事業主等、労働者以外でも業務の実態や災害の発生状況からみて、労働者に準じ...  \n",
+       "\n",
+       "[342 rows x 5 columns]"
+      ]
+     },
+     "execution_count": 7,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "topic_model.get_topic_info()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "ft",
+   "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": "3.11.5"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}