• Stars
    star
    163
  • Rank 231,141 (Top 5 %)
  • Language
    Python
  • License
    MIT License
  • Created over 4 years ago
  • Updated over 1 year ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

Simple Chit-Chat based on KoGPT2

Simple Chit-Chat based on KoGPT2

Purpose

Architecture

ν•™μŠ΅ 데이터에 μ ν•©ν•˜κ²Œ Hello! GPT-2λ₯Ό μ‘μš©, μ•„λž˜μ™€ 같은 μ•„ν‚€ν…μ²˜λ₯Ό 섀계함

drawing

  • λ°μ΄ν„°μ˜ Q ν•„λ“œλ₯Ό <usr>λ°œν™”, A ν•„λ“œλ₯Ό <sys>λ°œν™” 그리고 감정 λ ˆμ΄λΈ”μ„ <sent>둜 맀핑해 P(<sys>|<usr>, <sent>)λ₯Ό μ΅œλŒ€ν™” ν•  수 μžˆλŠ” λͺ¨λΈμ„ ν•™μŠ΅ν•œλ‹€. 감정 λ ˆμ΄λΈ”μ€ 이곳의 μ •μ˜λ₯Ό λ”°λ₯Έλ‹€(μΌμƒλ‹€λ°˜μ‚¬ 0, 이별(λΆ€μ •) 1, μ‚¬λž‘(긍정) 2).

Install

git clone --recurse-submodules https://github.com/haven-jeon/KoGPT2-chatbot.git
cd KoGPT2-chatbot
pip3 install -r requirements.txt 

How to Train

CUDA_VISIBLE_DEVICES=0 python train_torch.py --gpus 1 --train --max_epochs 2

How to Chat!

# CUDA_VISIBLE_DEVICES=0 python train.py --chat
CUDA_VISIBLE_DEVICES=0 python train_torch.py --gpus 1 --chat
user > λŒ€λ°• 
Simsimi > 잘 될 κ±°μ˜ˆμš”.

user > λ„ˆ 이름이 λ­λ‹ˆ?
Simsimi > μ €λŠ” μœ„λ‘œλ΄‡μž…λ‹ˆλ‹€.

user > 어버이날엔 뭘 ν•˜μ£ ?
Simsimi > μ €λž‘ λ†€μ•„μš”.

user > 비가와
Simsimi > μš°μ‚° μ±™κΈ°μ„Έμš”.

user > λˆˆμ΄μ™€
Simsimi > λˆˆμ‚¬λžŒ λ§Œλ“€μ–΄λ³΄μ„Έμš”.

user > 천λ‘₯이 쳐
Simsimi > 저도 보고 μ‹Άμ–΄μš”.

user > 봄이왔어.
Simsimi > 봄은 항상 λ‘κ·Όκ±°λ €μš”.

user > λ°°νƒˆλ‚¬μ–΄.
Simsimi > μ–Όλ₯Έ 집에 κ°€μ„œ μ‰¬μ„Έμš”.

user > λ„ˆλ₯Ό λ§Œλ“  μ‚¬λžŒμ€?
Simsimi > 당신이 λ§Œλ“  μ‚¬λžŒμ΄μ—μš”.

Colab tutorials

  • PyTorch
    • Open In Colab

More Repositories

1

PyKoSpacing

Automatic Korean word spacing with Python
Python
315
star
2

KoNLP

R package for Korean NLP
Java
152
star
3

LegalQA

Korean LegalQA using SentenceKoBART
Python
83
star
4

KoSpacing

Automatic Korean word spacing with R
R
76
star
5

ko_en_neural_machine_translation

Korean English NMT(Neural Machine Translation) with Gluon
Jupyter Notebook
60
star
6

TensorFlow-Book-R

This is the unofficial code repository for Machine Learning with TensorFlow(R).
HTML
58
star
7

KoBART-chatbot

KoBART chatbot
Python
45
star
8

TrainKoSpacing

Automatic Korean word spacing with neural n-gram detector(NND)
Python
35
star
9

NIADic

NIA(National Information Society Agency) Hangul Dictionary
R
33
star
10

KoWordSpacing

Korean Word Spacing with RNN.
HTML
22
star
11

KoGPT2-subtasks

NSMC, KorSTS ... fine-tunings
Python
20
star
12

ko_data_science_docker

데이터 뢄석 λͺ¨λΈλ§μš© 도컀 이미지
Dockerfile
18
star
13

grad_cam_gluon

Grad CAM for Text Classification
Jupyter Notebook
16
star
14

HDKU

HDKU : Hangul Dubeolsik Keystroke Utils
Python
10
star
15

beer_recommander

keras based beer recommendation
HTML
8
star
16

korea_real_estate_analysis

뢀동산 데이터 뢄석 μ½”λ“œ 및 데이터
HTML
7
star
17

rdatamining

R Data Mining ꡐ윑자료
7
star
18

HanNanum-Analyzer

HanNanum Analyzer for KoNLP
Java
5
star
19

2014_Seoul_Mayoral_Election_Analysis

Bayesian Inference using Opinion Survey of Seoul Mayoral Election 2014
TeX
5
star
20

DeepLearning_with_R

Deep Learning with R
4
star
21

knitr_example

knitr example
R
4
star
22

Ruchardet

R port of 'universalchardet', that is the encoding detector library of Mozilla.
C++
4
star
23

BOPR

Bayesian online learning scheme for probit regression with R
R
3
star
24

R_based_visualization

Rλ‘œν•˜λŠ” 데이터 μ‹œκ°ν™”
3
star
25

introduction_to_most_usable_pkgs_in_project

μ‹€μ œ ν”„λ‘œμ νŠΈ ν•˜λ©΄μ„œ ν•„μˆ˜μΈ νŒ¨ν‚€μ§€λ“€ μ†Œκ°œ
R
3
star
26

reproducible-data-analysis-examples

reproducible data analysis examples
R
1
star
27

RscriptUtils

RscriptUtils, Tools to make developing Rscript easier
R
1
star
28

CIIA_Korean

Collective Intelligence in Action Examples for Korean
Java
1
star
29

Sejong

KoNLP static dictionaries and Sejong project resources for corpus linguistics.
R
1
star
30

calcifer-vtuber

1
star
31

GMenuNext

C++
1
star