hash2430

pitchtron

TTS for pitch-accented language. Korean dialect DB.

dv3_world

Deep voice 3 + WORLD vocoder.

emotion_tts

adaptive_voice_cloning

BERT_sentiment_classification

BERT-based binary sentiment classification + Byte-pair encoding (BPE) implemented from scratch.

French_EN_NMT

French to English neural machine translation trained on multi30k dataset.

Korean_Intent_classifier

Question_answering

This simple question answering model achieves F1 score of 80.56 on dev set of SQuAD

ai-deadlines

:alarm_clock: AI conference deadline countdowns

AwesomeKorean_Data

한국어 데이터 세트 링크

DistilKoBERT

Distillation of KoBERT from SKTBrain (Lightweight KoBERT)

emotiontts_open_db

음성DB 공개

Episodic-DG

This is the repo for the paper "Episodic Training for Domain Generalization" https://arxiv.org/abs/1902.00113

flowtron

Auto-regressive flow-based generative network for text to speech synthesis

GoEmotions-Korean

Korean version of GoEmotions Dataset 😍😢😱

hash2430.github.io

KoBERT

Korean BERT pre-trained cased (KoBERT)

KoBERT-nsmc

Naver movie review sentiment classification with KoBERT

kochat

Opensource Korean chatbot framework based on deep learning 💬

KoELECTRA

Pretrained ELECTRA Model for Korean

kor_pratrain_LM

https://ailabs.enliple.com/

korean-hate-speech

Korean HateSpeech Dataset

korean_toxic_detection

LASER

Language-Agnostic SEntence Representations

mellotron

Mellotron: a multispeaker voice synthesis model based on Tacotron 2 GST that can make a voice emote and sing without emotive or singing training data

ML-Interview

Resources I used for ML Engineer, Applied Scientist and Quant Researcher interviews.

nlp-startups

국내 자연어 처리 기술을 연구 및 개발하는 스타트업 목록

nlpaug

Data augmentation for NLP

sacred

Sacred is a tool to help you configure, organize, log and reproduce experiments developed at IDSIA.

Word2vec

Word2vec implementation with naive softmax loss and negative sampling softmax.