smartPatent-mContriever-lora

The model is fine-tuned on the customed Korean Patent Retrieval system.

Training Data

Two types of datasets are used as training data: queries automatically generated through GPT-4 and patent titles that are linked to existing patent abstracts.

Usage

from transformers import AutoTokenizer, AutoModel, AutoModelForSequenceClassification
import torch
from transformers import AutoModel, AutoTokenizer
from peft import PeftModel, PeftConfig

def get_model(peft_model_name):
    config = PeftConfig.from_pretrained(peft_model_name)
    base_model = AutoModel.from_pretrained(config.base_model_name_or_path)
    model = PeftModel.from_pretrained(base_model, peft_model_name)
    model = model.merge_and_unload()
    model.eval()
    return model

# Load the tokenizer and model
tokenizer = AutoTokenizer.from_pretrained('facebook/mcontriever-msmarco')
model = get_model('hanseokOh/smartPatent-mContriever-lora')

Info

• Developed by: hanseokOh
• Model type: information retriever
• Language(s) (NLP): Korean
• Finetuned from model [optional]: mContriever-msmarco

Model Sources [optional]

• Repository: https://github.com/hanseokOh/PatentSearch