llama3.1_korean_v0.1_sft_by_aidx

Introduction

The LLAMA3.1_KOREAN_V0.1_SFT_BY_AIDX model is designed to understand and apply Korean language and cultural contexts. It is based on the LLAMA3.1 Instruct foundation model and trained using a variety of Korean datasets to align with Korean societal values and cultural nuances.

Architecture

The model uses the Meta-LLAMA/Llama-3.1-8B-Instruct as its base. It is specialized for text generation tasks, particularly in Korean and English languages.

Training

The model was trained on a proprietary dataset totaling 3.6GB, comprising 2.33 million records of Q&A, summaries, and classifications. This includes 1.33 million multiple-choice questions across 53 domains like history, finance, and law, and 1.3 million subjective questions across 38 domains. The data is designed to understand Korean societal values and emotions.

Guide: Running Locally

To run the model locally, follow these steps:

1. Install the transformers library.

2. Load the tokenizer and model using the following code:

   from transformers import AutoModel, AutoTokenizer
   
   tokenizer = AutoTokenizer.from_pretrained("SEOKDONG/llama3.1_korean_v0.1_sft_by_aidx")
   model = AutoModel.from_pretrained("SEOKDONG/llama3.1_korean_v0.1_sft_by_aidx")
   
   input_text = """ 「국민건강보험법」제44조, 「국민건강보험법 시행령」제19조,「약관의 규제에 관한 법률」제5조, 「상법」제54조 참조 판단 해줘"""
   inputs = tokenizer(input_text, return_tensors="pt")
   
   with torch.no_grad():
       outputs = model.generate(**inputs, max_length=1024, temperature=0.5, do_sample=True, repetition_penalty=1.15)
   result = tokenizer.decode(outputs[0], skip_special_tokens=True)
   print(result)
   

3. Consider using a cloud GPU service, such as AWS, Google Cloud, or Azure, for optimal performance.

License

The model is licensed under the Apache-2.0 License.