A collection of Korean Text Datasets ready to use using Tensorflow-Datasets.

Overview

tfds-korean

A collection of Korean Text Datasets ready to use using Tensorflow-Datasets.

TensorFlow-Datasets를 이용한 한국어/한글 데이터셋 모음입니다.

Dataset Catalog | pypi

PyPI - License PyPI Test Python

Usage

Installation

pip install tfds-korean

Loading dataset

import tensorflow_datasets as tfds
import tfds_korean.nsmc # register nsmc dataset

ds = tfds.load('nsmc')

train_ds = ds['train'].batch(32)
test_ds = ds['test'].batch(128)

# define model
# ....
# ....

model.fit(train_ds)
model.evaluate(test_ds)

See Dataset Catalog page for dataset list and details of each dataset.

Examples

Licenses

The license for this repository and licenses for datasets are applied separately. It is recommended to use each dataset after checking the dataset's website.

본 레포지토리의 라이선스와 데이터셋의 라이선스는 별도로 적용됩니다. 데이터셋을 사용하기 전 각 데이터셋의 라이선스와 웹 사이트를 확인 후 사용하시길 권해드리며, 본 라이브러리는 데이터셋을 호스팅하거나 배포하지 않는 점을 참고부탁드립니다.

Comments
  • [Dataset Request] sae4k

    [Dataset Request] sae4k

    Dataset Information

    • Dataset Name:
    • Prefered code name(e.g. korean_chatbot_qa_data): sae4k
    • Dataset description:
    • Homepage: https://github.com/warnikchow/sae4k
    • Citation:

    Additional Context

    dataset request 
    opened by jeongukjae 2
  • [Dataset Request] namuwiki corpus

    [Dataset Request] namuwiki corpus

    Dataset Information

    • Dataset Name: namuwiki corpus
    • Prefered code name(e.g. korean_chatbot_qa_data):
    • Dataset description:
    • Homepage: https://github.com/jeongukjae/namuwiki-corpus
    • Citation:
    • License:

    Additional Context

    문장 단위 분절해놓은 나무위키 코퍼스

    dataset request 
    opened by jeongukjae 1
  • [Dataset Request] korean wikipedia corpus

    [Dataset Request] korean wikipedia corpus

    Dataset Information

    • Dataset Name: 한국어 위키피디아 코퍼스
    • Prefered code name(e.g. korean_chatbot_qa_data): korean_wikipedia_corpus
    • Dataset description:
    • Homepage: https://github.com/jeongukjae/korean-wikipedia-corpus
    • Citation:
    • License:

    Additional Context

    kowikitext도 충분히 좋지만, 문장단위로 사용할 때 불편한 점이 있다. 그래서 문장단위로 이미 나누어진 말뭉치를 한국어 위키피디아 덤프에서 하나 생성. (kss로 분절)

    FeaturesDict({
        'content': Sequence(Text(shape=(), dtype=tf.string)),
        'title': Text(shape=(), dtype=tf.string),
    })
    

    요런식으로 content가 TensorSpec(shape=[None], dtype=tf.string)인 텐서값을 가지도록 만들어주면 distillation이나 문장 단위 unsupervised learning이나 할 때 편할 것 같아요.

    dataset request before-release 
    opened by jeongukjae 1
  • [Dataset Request] KLUE

    [Dataset Request] KLUE

    Dataset Information

    • Dataset Name: KLUE
    • Prefered code name(e.g. korean_chatbot_qa_data): klue_dp, klue_mrc, ...
    • Dataset description:
    • Homepage:
    • Citation:
    • License:

    Additional Context

    https://github.com/KLUE-benchmark/KLUE https://arxiv.org/pdf/2105.09680v1.pdf

    • [x] dp @jeongukjae
    • [x] mrc @harrydrippin
    • [x] ner @jeongukjae
    • [x] nli @jeongukjae
    • [x] re @jeongukjae
    • [x] sts @jeongukjae
    • [x] wos @jeongukjae
    • [x] ynat @jeongukjae
    dataset request before-release 
    opened by jeongukjae 1
  • [Dataset Request] namuwikitext

    [Dataset Request] namuwikitext

    Dataset Information

    • Dataset Name: Wikitext format dataset of Namuwiki
    • Prefered code name(e.g. korean_chatbot_qa_data): namuwikitext
    • Dataset description: 나무위키의 덤프 데이터를 바탕을 제작한 wikitext 형식의 텍스트 파일입니다. 학습 및 평가를 위하여 위키페이지 별로 train (99%), dev (0.5%), test (0.5%) 로 나뉘어져있습니다.
    • Homepage: https://github.com/lovit/namuwikitext
    • Citation:

    Additional Context

    https://github.com/lovit/namuwikitext/issues/10

    README에 있는 데이터셋 개수와 맞지 않아 이렇게 이슈 작성을 해놓았는데, 답변은 없는 상황임. 일단 Korpora에 있는 대로 추가해놓고 나중에 다시 수정하는 것이 좋지 않을까

    dataset request 
    opened by jeongukjae 1
  • [Dataset Request] KorQuAD

    [Dataset Request] KorQuAD

    Dataset Information

    • Dataset Name: KorQuAD 1.0
    • Prefered code name(e.g. korean_chatbot_qa_data): korquad_10
    • Dataset description: KorQuAD 1.0은 한국어 Machine Reading Comprehension을 위해 만든 데이터셋입니다. 모든 질의에 대한 답변은 해당 Wikipedia article 문단의 일부 하위 영역으로 이루어집니다. Stanford Question Answering Dataset(SQuAD) v1.0과 동일한 방식으로 구성되었습니다.
    • Homepage: https://korquad.github.io/KorQuad%201.0/
    • Citation:

    Dataset Information

    • Dataset Name: KorQuAD 2.0
    • Prefered code name(e.g. korean_chatbot_qa_data): korquad_20
    • Dataset description: KorQuAD 2.0은 KorQuAD 1.0에서 질문답변 20,000+ 쌍을 포함하여 총 100,000+ 쌍으로 구성된 한국어 Machine Reading Comprehension 데이터셋 입니다. KorQuAD 1.0과는 다르게 1~2 문단이 아닌 Wikipedia article 전체에서 답을 찾아야 합니다. 매우 긴 문서들이 있기 때문에 탐색 시간에 대한 고려가 필요할 것 입니다. 또한 표와 리스트도 포함되어 있기 때문에 HTML tag를 통한 문서의 구조 이해도 필요합니다. 이 데이터셋을 통해서 다양한 형태와 길이의 문서들에서도 기계독해가 가능해질 것 입니다.
    • Homepage: https://korquad.github.io
    • Citation:

    Additional Context

    일단은 KorQuAD 1.0만 추가해놓고 2.0은 추후에 추가해도 무방할 듯

    dataset request before-release 
    opened by jeongukjae 1
  • [Dataset Request] 한국해양대학교 NER 데이터셋

    [Dataset Request] 한국해양대학교 NER 데이터셋

    Dataset Information

    • Dataset Name: 한국해양대학교 자연언어처리 연구실 NER 데이터셋
    • Prefered code name(e.g. korean_chatbot_qa_data): kmounlp_ner
    • Dataset description: 한국어 개체명 정의 및 표지 표준화 기술보고서와 이를 기반으로 제작된 개체명 형태소 말뭉치
    • Homepage: https://github.com/kmounlp/NER
    • Citation:

    Additional Context

    보고서: https://github.com/kmounlp/NER/blob/master/NER%20Guideline%20(ver%201.0).pdf

    dataset request 
    opened by jeongukjae 1
  • Add CONTRIBUTING.md

    Add CONTRIBUTING.md

    • [ ] 프로젝트에서 사용하는 언어에 대한 설명. 사용법/데이터셋 설명은 가능하면 영어로 적되, 이슈/PR 소통은 한국어로 하는게 좋지 않을까?
    • [ ] 데이터셋 추가하는 법
    • [ ] 이슈/PR/Discussion 간단한 설명
    • [ ] 추가로 같이 관리하고 싶은 분들에 대한 설명
    • [ ] 데이터셋 라이선스에 대한 문제에 대한 설명
    documentation before-release 
    opened by jeongukjae 1
  • 현재 wikitext의 문제점을 카탈로그에 적어두기

    현재 wikitext의 문제점을 카탈로그에 적어두기

    https://github.com/jeongukjae/tfds-korean/issues/12#issuecomment-826358469

    위와 같은 이유로 "필터를 해서 사용해라" 혹은 "중간에 빈 example이 있다" 정도는 적어두는 편이 좋은 듯

    documentation 
    opened by jeongukjae 0
  • [Dataset Request] sci-news-sum-kr-50

    [Dataset Request] sci-news-sum-kr-50

    Dataset Information

    • Dataset Name:
    • Prefered code name(e.g. korean_chatbot_qa_data): sci_news_sum_kr_50
    • Dataset description:
    • Homepage: https://github.com/theeluwin/sci-news-sum-kr-50
    • Citation:

    Additional Context

    dataset request 
    opened by jeongukjae 0
  • [Dataset Request] kowikitext

    [Dataset Request] kowikitext

    Dataset Information

    • Dataset Name: 한국어 wikitext
    • Prefered code name(e.g. korean_chatbot_qa_data): kowikitext
    • Dataset description: Wikitext format Korean corpus
    • Homepage: https://github.com/lovit/kowikitext
    • Citation:

    Additional Context

    이것도 #12 와 같은 문제점이 존재하는 것으로 보이는데, 일단은 Korpora 방식을 따라감. 이 데이터셋도 heading을 기준으로 split할 경우 = 분류~~~ =같은 행들이 존재하여 정확히 문서 단위로 복구가 불가능함.

    dataset request 
    opened by jeongukjae 0
  • [Dataset Request] korean_unsmile_dataset

    [Dataset Request] korean_unsmile_dataset

    Dataset Information

    • Dataset Name:
    • Prefered code name(e.g. korean_chatbot_qa_data):
    • Dataset description:
    • Homepage: https://github.com/smilegate-ai/korean_unsmile_dataset
    • Citation:
    • License:

    Additional Context

    dataset request 
    opened by jeongukjae 0
  • 데이터셋 카탈로그 빌더 특정 데이터셋 스킵가능하게 수정

    데이터셋 카탈로그 빌더 특정 데이터셋 스킵가능하게 수정

    현재 모든 데이터셋이 로컬에 존재해야 카탈로그를 빌드할 수 있는데, 이게 너무 부담이 된다. 현재 develop 기준만 해도 대략 30GB를 로컬에 들고 있어야 한다.

    데이터셋 버전이 바뀌지 않는다면 카탈로그를 다시 빌드해야하는 때는 build_catalog.py 스크립트가 변경될 때 뿐이라서 특정 데이터셋 페이지 & index 페이지만 빌드해도 되도록 수정해두자. 물론 전체 데이터셋에 대한 카탈로그 빌드도 가능하게 유지.

    documentation 
    opened by jeongukjae 0
  • [Dataset Request] Korean Single Speaker Speech Dataset

    [Dataset Request] Korean Single Speaker Speech Dataset

    Dataset Information

    • Dataset Name: Korean Single Speaker Speech Dataset
    • Prefered code name(e.g. korean_chatbot_qa_data):
    • Dataset description:
    • Homepage: https://www.kaggle.com/bryanpark/korean-single-speaker-speech-dataset
    • Citation:
    • License:

    Additional Context

    dataset request 
    opened by jeongukjae 0
  • [Dataset Request] 세종코퍼스

    [Dataset Request] 세종코퍼스

    Dataset Information

    • Dataset Name:
    • Prefered code name(e.g. korean_chatbot_qa_data): sejong_corpus
    • Dataset description:
    • Homepage: https://ithub.korean.go.kr/user/total/database/corpusManager.do
    • Citation:
    • License:

    Additional Context

    세종 코퍼스: https://ithub.korean.go.kr/user/total/database/corpusManager.do 세종 코퍼스 - 병렬: https://ithub.korean.go.kr/user/total/database/etcManager.do

    라이선스가 상업적 이용이 어렵더라도 이용하기에 좋은 말뭉치라 생각해서 일단은 추가하는 게 좋을 것 같아요.

    dataset request 
    opened by jeongukjae 0
  • [Dataset Request] kcbert

    [Dataset Request] kcbert

    Dataset Information

    • Dataset Name:
    • Prefered code name(e.g. korean_chatbot_qa_data): kcbert
    • Dataset description:
    • Homepage: https://github.com/Beomi/KcBERT
    • Citation:

    Additional Context

    이거 추가해두면 엄청 유용하게 쓸 수 있을 것 같다!!

    dataset request 
    opened by jeongukjae 4
  • [Dataset Request] KAIST Corpus

    [Dataset Request] KAIST Corpus

    Dataset Information

    • Dataset Name: kaist corpus
    • Prefered code name(e.g. korean_chatbot_qa_data): kaist_corpus
    • Dataset description:
    • Homepage: http://semanticweb.kaist.ac.kr/home/index.php/KAIST_Corpus
    • Citation:

    Additional Context

    wontfix dataset request 
    opened by jeongukjae 1
Releases(0.4.0)
  • 0.4.0(Sep 19, 2021)

    • Update KLUE dataset to 1.1.0 https://github.com/jeongukjae/tfds-korean/commit/e954ec4550ec5db015d3f93750e6763aca5a9b48
    • Reorder ClassLabel names of NLI datasets. https://github.com/jeongukjae/tfds-korean/commit/be3e8cba7b9d537969b9c08738dd6df36b0145bc
    Source code(tar.gz)
    Source code(zip)
  • 0.3.0(Jun 16, 2021)

    • add korean_wikipedia_corpus (https://jeongukjae.github.io/tfds-korean/datasets/korean_wikipedia_corpus.html)
    • add namuwiki_corpus (https://jeongukjae.github.io/tfds-korean/datasets/namuwiki_corpus.html)
    Source code(tar.gz)
    Source code(zip)
  • 0.2.0(Jun 6, 2021)

    • add KLUE benchmark datasets
    • update dataset catalog (https://github.com/jeongukjae/tfds-korean/commit/eb1c72d0a716aba7326276e77e8e6f94976bb579, https://github.com/jeongukjae/tfds-korean/commit/614616b82d0bbdaecbc4ec50e0cfc67b78b646c2)
    • fix klue_ner supervised key bug (https://github.com/jeongukjae/tfds-korean/commit/10f765f01b9f3952e298395779dcf8efeefde93a)
    Source code(tar.gz)
    Source code(zip)
  • 0.1.3(May 29, 2021)

  • 0.1.2(May 25, 2021)

  • 0.1.1(Apr 30, 2021)

  • 0.1.0(Apr 29, 2021)

    • Add kowikitext and namuwikitext dataset
    • Add missing licenses and bibtex.
    • Add license section in catalog page.
    • Add example links in catalog page.
    Source code(tar.gz)
    Source code(zip)
Owner
Jeong Ukjae
Machine Learning Engineer
Jeong Ukjae
💫 Industrial-strength Natural Language Processing (NLP) in Python

spaCy: Industrial-strength NLP spaCy is a library for advanced Natural Language Processing in Python and Cython. It's built on the very latest researc

Explosion 24.9k Jan 02, 2023
Code from the paper "High-Performance Brain-to-Text Communication via Handwriting"

Code from the paper "High-Performance Brain-to-Text Communication via Handwriting"

Francis R. Willett 305 Dec 22, 2022
pkuseg多领域中文分词工具; The pkuseg toolkit for multi-domain Chinese word segmentation

pkuseg:一个多领域中文分词工具包 (English Version) pkuseg 是基于论文[Luo et. al, 2019]的工具包。其简单易用,支持细分领域分词,有效提升了分词准确度。 目录 主要亮点 编译和安装 各类分词工具包的性能对比 使用方式 论文引用 作者 常见问题及解答 主要

LancoPKU 6k Dec 29, 2022
🕹 An esoteric language designed so that the program looks like the transcript of a Pokémon battle

PokéBattle is an esoteric language designed so that the program looks like the transcript of a Pokémon battle. Original inspiration and specification

Eduardo Correia 9 Jan 11, 2022
A library for end-to-end learning of embedding index and retrieval model

Poeem Poeem is a library for efficient approximate nearest neighbor (ANN) search, which has been widely adopted in industrial recommendation, advertis

54 Dec 21, 2022
Disfl-QA: A Benchmark Dataset for Understanding Disfluencies in Question Answering

Disfl-QA is a targeted dataset for contextual disfluencies in an information seeking setting, namely question answering over Wikipedia passages. Disfl-QA builds upon the SQuAD-v2 (Rajpurkar et al., 2

Google Research Datasets 52 Jun 21, 2022
ACL'22: Structured Pruning Learns Compact and Accurate Models

☕ CoFiPruning: Structured Pruning Learns Compact and Accurate Models This repository contains the code and pruned models for our ACL'22 paper Structur

Princeton Natural Language Processing 130 Jan 04, 2023
Toy example of an applied ML pipeline for me to experiment with MLOps tools.

Toy Machine Learning Pipeline Table of Contents About Getting Started ML task description and evaluation procedure Dataset description Repository stru

Shreya Shankar 190 Dec 21, 2022
The source code of HeCo

HeCo This repo is for source code of KDD 2021 paper "Self-supervised Heterogeneous Graph Neural Network with Co-contrastive Learning". Paper Link: htt

Nian Liu 106 Dec 27, 2022
p-tuning for few-shot NLU task

p-tuning_NLU Overview 这个小项目是受乐于分享的苏剑林大佬这篇p-tuning 文章启发,也实现了个使用P-tuning进行NLU分类的任务, 思路是一样的,prompt实现方式有不同,这里是将[unused*]的embeddings参数抽取出用于初始化prompt_embed后

3 Dec 29, 2022
Translation to python of Chris Sims' optimization function

pycsminwel This is a locol minimization algorithm. Uses a quasi-Newton method with BFGS update of the estimated inverse hessian. It is robust against

Gustavo Amarante 1 Mar 21, 2022
Code for the paper in Findings of EMNLP 2021: "EfficientBERT: Progressively Searching Multilayer Perceptron via Warm-up Knowledge Distillation".

This repository contains the code for the paper in Findings of EMNLP 2021: "EfficientBERT: Progressively Searching Multilayer Perceptron via Warm-up Knowledge Distillation".

Chenhe Dong 28 Nov 10, 2022
Finds snippets in iambic pentameter in English-language text and tries to combine them to a rhyming sonnet.

Sonnet finder Finds snippets in iambic pentameter in English-language text and tries to combine them to a rhyming sonnet. Usage This is a Python scrip

Marcel Bollmann 11 Sep 25, 2022
This is a modification of the OpenAI-CLIP repository of moein-shariatnia

This is a modification of the OpenAI-CLIP repository of moein-shariatnia

Sangwon Beak 2 Mar 04, 2022
使用Mask LM预训练任务来预训练Bert模型。训练垂直领域语料的模型表征,提升下游任务的表现。

Pretrain_Bert_with_MaskLM Info 使用Mask LM预训练任务来预训练Bert模型。 基于pytorch框架,训练关于垂直领域语料的预训练语言模型,目的是提升下游任务的表现。 Pretraining Task Mask Language Model,简称Mask LM,即

Desmond Ng 24 Dec 10, 2022
SimCSE: Simple Contrastive Learning of Sentence Embeddings

SimCSE: Simple Contrastive Learning of Sentence Embeddings This repository contains the code and pre-trained models for our paper SimCSE: Simple Contr

Princeton Natural Language Processing 2.5k Jan 07, 2023
TunBERT is the first release of a pre-trained BERT model for the Tunisian dialect using a Tunisian Common-Crawl-based dataset.

TunBERT is the first release of a pre-trained BERT model for the Tunisian dialect using a Tunisian Common-Crawl-based dataset. TunBERT was applied to three NLP downstream tasks: Sentiment Analysis (S

InstaDeep Ltd 72 Dec 09, 2022
Easy to start. Use deep nerual network to predict the sentiment of movie review.

Easy to start. Use deep nerual network to predict the sentiment of movie review. Various methods, word2vec, tf-idf and df to generate text vectors. Various models including lstm and cov1d. Achieve f1

1 Nov 19, 2021
A sentence aligner for comparable corpora

About Yalign is a tool for extracting parallel sentences from comparable corpora. Statistical Machine Translation relies on parallel corpora (eg.. eur

Machinalis 128 Aug 24, 2022
Python SDK for working with Voicegain Speech-to-Text

Voicegain Speech-to-Text Python SDK Python SDK for the Voicegain Speech-to-Text API. This API allows for large vocabulary speech-to-text transcription

Voicegain 3 Dec 14, 2022