shleeeee/mistral-ko-OpenOrca-wiki-v1

TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:8kLicense:otherArchitecture:Transformer Cold

The shleeeee/mistral-ko-OpenOrca-wiki-v1 is a fine-tuned Mistral-7B-v0.1 model developed by shleeeee (Seunghyeon Lee) and oopsung (Sungwoo Park). This model is specifically optimized for Korean language tasks, leveraging fine-tuning on 2000 ko-OpenOrca datasets. It is designed for general text generation and understanding in Korean, building upon the Mistral architecture.

Loading preview...

Model Overview

shleeeee/mistral-ko-OpenOrca-wiki-v1 is a specialized language model developed by shleeeee (Seunghyeon Lee) and oopsung (Sungwoo Park). It is a fine-tuned variant of the Mistral-7B-v0.1 architecture, specifically adapted for Korean language processing.

Key Characteristics

  • Base Model: Mistral-7B-v0.1.
  • Language Focus: Optimized for Korean language tasks.
  • Fine-tuning Data: Trained on 2000 samples from the ko-OpenOrca dataset.
  • Training Details: Fine-tuned with 2 epochs and a train_batch size of 4.
  • LoRA Target Modules: q_proj, k_proj, v_proj, o_proj, gate_proj were targeted during LoRA fine-tuning.
  • Prompt Format: Utilizes the standard Mistral prompt template: <s>[INST]{['instruction']}[/INST]{['output']}</s>.

Usage

This model can be easily integrated into applications using the Hugging Face transformers library for text generation tasks. It is suitable for various Korean natural language processing applications where a Mistral-based model with Korean language proficiency is required.

Limitations

Evaluation details are currently pending, so performance metrics and specific strengths/weaknesses are yet to be fully documented.