language: en 

datasets:
- 37 popular Python code repositories
- See princeton-nlp/SWE-bench train split
- See the `make_datasets` documentation on SWE-bench's [GitHub](https://github.com/princeton-nlp/SWE-bench/tree/main/inference/make_datasets) for details on formatting input.

---

# SWE-Llama

SWE-Llama are variants of the [CodeLlama](https://arxiv.org/abs/2308.12950) model fine-tuned on software engineering tasks extracted from real-world GitHub issues and pull requests. They were introduced and evaluated on the SWE-bench benchmark in this [paper](https://arxiv.org/abs/2310.06770).

## Model Details

- **Architecture:** Transformer, based on [CodeLlama](https://arxiv.org/abs/2308.12950) architecture 
- **Parameters:** 7 billion for SWE-Llama-7b, 13 billion for SWE-Llama-13b
- **Objective:** Generating patches to resolve GitHub issues, conditioned on issue description and code context

## Training Data

SWE-Llama was fine-tuned on 19,000 issues and pull requests collected from 37 popular Python code repositories on GitHub, disjoint from those used in SWE-bench.

## Training Procedure

- Fine-tuned only the attention matrices using LoRA method
- Trained for 4 epochs with a batch size of 32
- Selected best checkpoint based on validation perplexity

## Evaluation Results

When evaluated on the SWE-bench benchmark:

- SWE-Llama-7b achieved 3.0% issue resolution rate using oracle context retrieval
- SWE-Llama-13b achieved 4.0% issue resolution rate using oracle context retrieval

## BibTeX Entry
```tex
@misc{jimenez2023swebench,
      title={SWE-bench: Can Language Models Resolve Real-World GitHub Issues?}, 
      author={Carlos E. Jimenez and John Yang 
        and Alexander Wettig and Shunyu Yao 
        and Kexin Pei and Ofir Press and Karthik Narasimhan},
      year={2023},
      eprint={2310.06770},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}
```