qwen2.5-0.5b-issam-sft

Qwen2.5-0.5B Fine-tuned on issam-data

This model is a fine-tuned version of Qwen/Qwen2.5-0.5B on the sh2124/issam-data dataset.

Training Details

• Base Model: Qwen/Qwen2.5-0.5B
• Dataset: sh2124/issam-data
• Training Type: Supervised Fine-tuning (SFT)
• Epochs: 3
• Training Loss: 3.94

Usage

from transformers import AutoTokenizer, AutoModelForCausalLM

model_name = "sh2124/qwen2.5-0.5b-issam-sft"
tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(model_name, trust_remote_code=True)

# Example usage
question = "아트봉봉과의 협업 내용은?"
formatted_input = f"질문: {question}\n답변:"
inputs = tokenizer(formatted_input, return_tensors="pt")
outputs = model.generate(**inputs, max_length=200, do_sample=True, temperature=0.7)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(response)

Model Performance

The fine-tuned model shows improved performance on domain-specific questions related to the training data, particularly for Korean language queries about collaboration and business topics.