castorini/rankllama-v1-13b-lora-passage

RankLLaMA-13B-Passage

Fine-Tuning LLaMA for Multi-Stage Text Retrieval. Xueguang Ma, Liang Wang, Nan Yang, Furu Wei, Jimmy Lin, arXiv 2023

This model is fine-tuned from LLaMA-2-13B using LoRA for passage reranking.

Training Data

The model is fine-tuned on the training split of MS MARCO Passage Ranking datasets for 1 epoch. Please check our paper for details.

Usage

Below is an example to compute the similarity score of a query-passage pair

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

def get_model(peft_model_name):
    config = PeftConfig.from_pretrained(peft_model_name)
    base_model = AutoModelForSequenceClassification.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('meta-llama/Llama-2-13b-hf')
model = get_model('castorini/rankllama-v1-13b-lora-passage')

# Define a query-passage pair
query = "What is llama?"
title = "Llama"
passage = "The llama is a domesticated South American camelid, widely used as a meat and pack animal by Andean cultures since the pre-Columbian era."

# Tokenize the query-passage pair
inputs = tokenizer(f'query: {query}', f'document: {title} {passage}', return_tensors='pt')

# Run the model forward
with torch.no_grad():
    outputs = model(**inputs)
    logits = outputs.logits
    score = logits[0][0]
    print(score)

Citation

If you find our paper or models helpful, please consider cite as follows:

@article{rankllama,
      title={Fine-Tuning LLaMA for Multi-Stage Text Retrieval}, 
      author={Xueguang Ma and Liang Wang and Nan Yang and Furu Wei and Jimmy Lin},
      year={2023},
      journal={arXiv:2310.08319},
}