shleeeee/mistral-ko-OpenOrca-2000

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

The shleeeee/mistral-ko-OpenOrca-2000 model, developed by shleeeee (Seunghyeon Lee) and oopsung (Sungwoo Park), is a fine-tuned version of the Mistral-7B-v0.1 architecture. This model is specifically optimized for Korean language tasks, having been fine-tuned on 2000 ko-OpenOrca datasets. Its primary strength lies in generating Korean text, leveraging the Mistral architecture's efficiency for language understanding and generation in a Korean context.

Loading preview...

Overview

The shleeeee/mistral-ko-OpenOrca-2000 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 the Korean language. The fine-tuning process involved using 2000 entries from the ko-OpenOrca dataset, enhancing its performance and understanding of Korean linguistic nuances.

Key Characteristics

  • Base Model: Mistral-7B-v0.1
  • Language Focus: Optimized for Korean language generation and understanding.
  • Fine-tuning Data: Utilizes 2000 ko-OpenOrca datasets.
  • Training Details: Fine-tuned with train_batch of 4 and 2 epochs.
  • LoRA Target Modules: q_proj, k_proj, v_proj, o_proj, gate_proj were targeted during LoRA fine-tuning.
  • Prompt Template: Employs the standard Mistral prompt format: <s>[INST]{['instruction']}[/INST]{['output']}</s>.

Good For

  • Korean Text Generation: Ideal for applications requiring high-quality text output in Korean.
  • Korean Language Understanding: Suitable for tasks involving comprehension and processing of Korean text.
  • Research and Development: Provides a strong base for further experimentation and fine-tuning on specific Korean NLP tasks.

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