{"id":17272645,"url":"https://github.com/codertimo/korquad-question-generation","last_synced_at":"2025-04-14T08:34:27.396Z","repository":{"id":47256748,"uuid":"315890057","full_name":"codertimo/KorQuAD-Question-Generation","owner":"codertimo","description":"question generation model with KorQuAD dataset","archived":false,"fork":false,"pushed_at":"2021-09-06T13:10:32.000Z","size":935,"stargazers_count":38,"open_issues_count":0,"forks_count":12,"subscribers_count":2,"default_branch":"main","last_synced_at":"2025-03-27T22:07:12.580Z","etag":null,"topics":["gpt2","korquad","question-generation"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/codertimo.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null}},"created_at":"2020-11-25T09:28:41.000Z","updated_at":"2024-07-03T07:22:54.000Z","dependencies_parsed_at":"2022-09-03T11:50:25.541Z","dependency_job_id":null,"html_url":"https://github.com/codertimo/KorQuAD-Question-Generation","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":"codertimo/python-template","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/codertimo%2FKorQuAD-Question-Generation","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/codertimo%2FKorQuAD-Question-Generation/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/codertimo%2FKorQuAD-Question-Generation/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/codertimo%2FKorQuAD-Question-Generation/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/codertimo","download_url":"https://codeload.github.com/codertimo/KorQuAD-Question-Generation/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":248846276,"owners_count":21170939,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["gpt2","korquad","question-generation"],"created_at":"2024-10-15T08:49:06.550Z","updated_at":"2025-04-14T08:34:27.357Z","avatar_url":"https://github.com/codertimo.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Question Generation(QG) Model with KorQuAD\n\n학습된 [SKT-AI/KoGPT2](https://github.com/SKT-AI/KoGPT2) 모델을 기반으로 질문 생성 QG(Question Generation) 모델을 만들었습니다. \nQG 모델을 만들기 위해 Question Answering 데이터셋인 KorQuAD v1.0을 사용하였습니다.\n\n## 사용 방법\n\n### 데이터 준비\n\n학습/평가/생성을 위해서 KorQuAD v1.0 데이터셋을 다운 받습니다.\n\n```shell\nmake prepare-dataset\n```\n\n### 학습\n\n다음 커맨드를 이용해서 학습을 수행할 수 있습니다.\n\n```shell\npython -m scripts.run_fine_tune --train-batch-size 16 --eval-batch-size 16 --epochs 5\n```\n\n### 성능 평가 (dev 셋 PPL 측정)\n\n```shell\nMODEL_PATH = \"artifacts/gpt2_xxxxxxxx/gpt2_step_x.pth\"\npython -m scripts.run_evaluation --model-path $MODEL_PATH --batch-size 50\n```\n\n### 질문 생성 (dev 셋에 대해서 질문 생성)\n\n![Decoding 결과](docs/decoded_examples.png)\n\n[Question Generation POC 스프레드 시트](https://docs.google.com/spreadsheets/d/1-PQKFTfBhyH-K0EBa03_KkPMgL86wF_rn_813Hk-LNQ): KorQuAD v1.0 dev 셋에 대해서 decoding 한 결과 입니다.\n\nbeam-search 를 기반으로 decoding 되었으며, beam_size 는 5를 사용하였습니다.\n\n```shell\nMODEL_PATH = \"artifacts/gpt2_xxxxxxxx/gpt2_step_x.pth\"\npython -m scripts.run_generate --model-path $MODEL_PATH --output-path decoded.tsv\n```\n\n### 학습된 QG 모델 다운로드\n\n- [모델 weight](https://drive.google.com/file/d/1t6ChVQwp3MapJDVEBCnyJcdVurIvn7y7/view?usp=sharing)\n- [학습 로그](https://drive.google.com/file/d/1bMi_AA5nhTt72iEIzcg5gbyN5AYdZxI8/view?usp=sharing)\n\n## Author\n\nby Junseong Kim (Scatter Lab, Pingpong AI) codertimo@gmail.com\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fcodertimo%2Fkorquad-question-generation","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fcodertimo%2Fkorquad-question-generation","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fcodertimo%2Fkorquad-question-generation/lists"}