AIdenU/LLAMA-2-13b-ko-Y24-DPO_v2.0

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:13BQuant:FP8Ctx Length:4kLicense:apache-2.0Architecture:Transformer Open Weights Warm

AIdenU/LLAMA-2-13b-ko-Y24-DPO_v2.0 is a 13 billion parameter language model developed by AIdenU, fine-tuned using Direct Preference Optimization (DPO). This model is based on the LLAMA-2 architecture and is specifically optimized for Korean language processing. It is designed for generating high-quality, contextually relevant responses in Korean.

Loading preview...

AIdenU/LLAMA-2-13b-ko-Y24-DPO_v2.0 Overview

AIdenU/LLAMA-2-13b-ko-Y24-DPO_v2.0 is a 13 billion parameter language model, a DPO-fine-tuned variant of the AIdenU/LLAMA-2-13b-ko-Y24_v2.0 base model. This model leverages the robust LLAMA-2 architecture, enhanced through Direct Preference Optimization to improve its response quality and alignment.

Key Capabilities

  • Korean Language Proficiency: Specifically optimized for understanding and generating text in Korean.
  • DPO Fine-tuning: Benefits from Direct Preference Optimization, which typically leads to more helpful, harmless, and honest outputs compared to standard fine-tuning methods.
  • Causal Language Modeling: Capable of generating coherent and contextually relevant text based on given prompts.

Usage

This model is suitable for various natural language processing tasks requiring strong Korean language capabilities, including:

  • Conversational AI: Building chatbots or virtual assistants that interact in Korean.
  • Content Generation: Creating Korean text for articles, summaries, or creative writing.
  • Question Answering: Providing informative answers to questions posed in Korean.

Popular Sampler Settings

Top 3 parameter combinations used by Featherless users for this model. Click a tab to see each config.

temperature
top_p
top_k
frequency_penalty
presence_penalty
repetition_penalty
min_p