jaehwan02/risolju-1.0-1.7b

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:2BQuant:BF16Ctx Length:32kPublished:May 24, 2026License:apache-2.0Architecture:Transformer Open Weights Warm

The jaehwan02/risolju-1.0-1.7b is a 1.7 billion parameter full model, fine-tuned from Qwen/Qwen3-1.7B. This model is specifically designed for experimenting with Korean character speech styles, focusing on North Korean-style Korean and character-based conversational responses. It serves as a language style and character simulation for dialogue, not endorsing any political system or ideology.

Loading preview...

Model Overview

The jaehwan02/risolju-1.0-1.7b is a 1.7 billion parameter language model developed by jaehwan02. It is built upon the Qwen/Qwen3-1.7B base model, which is released under the Apache License 2.0.

Key Capabilities

  • Korean Character Speech Simulation: The primary focus of this model is to experiment with and generate responses in specific Korean character speech styles.
  • North Korean-style Korean: It is fine-tuned to simulate North Korean-style Korean language and conversational patterns.
  • Dialogue Generation: Designed for conversational AI, providing character-specific responses.

Intended Use

This model is intended for experimental purposes related to language style and character simulation. It is explicitly stated that the model is not intended to support or promote any specific political system, individual, ideology, or action. Developers can use it to explore unique linguistic nuances and character-driven interactions in Korean.

Example Usage

To generate text with this model, you can use the mlx_lm.generate command, specifying the model path and a prompt. For instance, a prompt like "너 누구야?" (Who are you?) might elicit a response such as "내레 리설주 1.0입네다. 조선식 문화어와 구어체로 동무의 물음에 곧장 답하는 대화식 인공지능이우." (I am Risolju 1.0. I am a conversational AI that answers your questions directly in Korean cultural language and colloquialisms.)