Ko-DialoGPT Open-Source Korean Dialogue Model - Free Deployment to Generate Natural Korean Dialogue Responses

Ko DialoGPT

Developed by byeongal

This is a Korean dialogue generation model based on the GPT-2 architecture, specifically designed to generate natural Korean dialogue responses.

Dialogue System

Transformers

Korean#Korean dialogue generation #Multi-turn dialogue processing #GPT-2 architecture optimization

Downloads 19

Release Time : 3/2/2022

Model Overview

This model is a Korean dialogue generation model based on the GPT-2 architecture, capable of generating coherent Korean dialogue responses based on user input.

Model Features

Korean Dialogue Optimization

Specially optimized for Korean dialogue scenarios, capable of generating natural and fluent Korean dialogue responses.

Context Awareness

Can remember and utilize dialogue history context to generate more coherent dialogue responses.

Based on GPT-2 Architecture

Utilizes the powerful language understanding and generation capabilities of GPT-2 to ensure high-quality dialogue generation.

Model Capabilities

Korean dialogue generation

Context-aware dialogue

Multi-turn dialogue processing

Use Cases

Chatbot

Korean Customer Service Bot

Used in Korean customer service scenarios to automatically answer customer questions

Improves customer service efficiency and reduces labor costs

Korean Social Chatbot

Used as a Korean conversation partner on social platforms

Provides a natural Korean dialogue experience

Language Learning

Korean Learning Assistant

Helps Korean learners practice conversations

Provides an authentic Korean dialogue environment

🚀 Ko-DialoGPT

A Korean conversational model based on GPT2, suitable for building chatbots and other conversational applications.

🚀 Quick Start

This section will guide you on how to use the Ko - DialoGPT model.

💻 Usage Examples

Basic Usage

from transformers import PreTrainedTokenizerFast, GPT2LMHeadModel
import torch


device = 'cuda' if torch.cuda.is_available() else 'cpu'

tokenizer = PreTrainedTokenizerFast.from_pretrained('byeongal/Ko-DialoGPT')
model = GPT2LMHeadModel.from_pretrained('byeongal/Ko-DialoGPT').to(device)

past_user_inputs = []
generated_responses = []

while True:
    user_input = input(">> User:")
    if user_input == 'bye':
        break
    text_idx = tokenizer.encode(user_input + tokenizer.eos_token, return_tensors='pt')
    for i in range(len(generated_responses)-1, len(generated_responses)-3, -1):
        if i < 0:
            break
        encoded_vector = tokenizer.encode(generated_responses[i] + tokenizer.eos_token, return_tensors='pt')
        if text_idx.shape[-1] + encoded_vector.shape[-1] < 1000:
            text_idx = torch.cat([encoded_vector, text_idx], dim=-1)
        else:
            break
        encoded_vector = tokenizer.encode(past_user_inputs[i] + tokenizer.eos_token, return_tensors='pt')
        if text_idx.shape[-1] + encoded_vector.shape[-1] < 1000:
            text_idx = torch.cat([encoded_vector, text_idx], dim=-1)
        else:
            break
    text_idx = text_idx.to(device)
    inference_output = model.generate(
            text_idx,
            max_length=1000,
            num_beams=5,
            top_k=20,
            no_repeat_ngram_size=4,
            length_penalty=0.65,
            repetition_penalty=2.0,
        )
    inference_output = inference_output.tolist()
    bot_response = tokenizer.decode(inference_output[0][text_idx.shape[-1]:], skip_special_tokens=True)
    print(f"Bot: {bot_response}")
    past_user_inputs.append(user_input)
    generated_responses.append(bot_response)

📚 Documentation

Reference

📄 License

This project is licensed under the CC - BY - NC - SA 4.0 license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご