https://github.com/swe-bench/SWE-bench
[ICLR 2024] SWE-bench: Can Language Models Resolve Real-world Github Issues?
https://github.com/swe-bench/SWE-bench
benchmark language-model software-engineering
Last synced: about 2 months ago
JSON representation
[ICLR 2024] SWE-bench: Can Language Models Resolve Real-world Github Issues?
- Host: GitHub
- URL: https://github.com/swe-bench/SWE-bench
- Owner: swe-bench
- License: mit
- Created: 2023-10-04T01:22:46.000Z (about 2 years ago)
- Default Branch: main
- Last Pushed: 2024-12-07T21:38:45.000Z (10 months ago)
- Last Synced: 2024-12-07T22:24:32.610Z (10 months ago)
- Topics: benchmark, language-model, software-engineering
- Language: Python
- Homepage: https://www.swebench.com
- Size: 10.1 MB
- Stars: 2,058
- Watchers: 28
- Forks: 363
- Open Issues: 30
-
Metadata Files:
- Readme: README.md
- Changelog: CHANGELOG.md
- License: LICENSE
Awesome Lists containing this project
- jimsghstars - swe-bench/SWE-bench - [ICLR 2024] SWE-bench: Can Language Models Resolve Real-world Github Issues? (Python)
- StarryDivineSky - swe-bench/SWE-bench - bench是一个用于评估大型语言模型（LLMs）在解决真实世界GitHub问题能力的项目，它在ICLR 2024上发表。该基准测试包含从GitHub收集的真实软件错误修复问题，旨在衡量LLMs理解、推理和生成正确代码修复的能力。SWE-bench强调现实场景，避免了人为构造的简化问题。项目特色在于其问题的真实性和复杂性，挑战LLMs处理实际软件开发任务。SWE-bench提供了一个标准化的评估平台，可以比较不同LLMs在软件修复任务上的表现。研究人员可以使用SWE-bench来推动LLMs在软件工程领域的应用，并识别现有模型的局限性。该项目包含一个数据集，以及用于评估模型性能的工具和脚本。SWE-bench的目的是促进LLMs在自动化软件修复方面的研究和发展。它专注于评估模型生成正确补丁的能力，并提供详细的评估指标。使用SWE-bench，研究人员可以更深入地了解LLMs在实际软件开发环境中的表现。该项目为LLMs在软件工程领域的应用提供了一个有价值的资源。 (A01_文本生成_文本对话 / 大语言对话模型及数据)
README
[ Read the Docs ]
---
Code and data for the following works:
* [ICLR 2025] SWE-bench Multimodal: Do AI Systems Generalize to Visual Software Domains?
* [ICLR 2024 Oral] SWE-bench: Can Language Models Resolve Real-World GitHub Issues?
## 📰 News
* **[Jan. 13, 2025]**: We've integrated [SWE-bench Multimodal](https://swebench.github.io/multimodal) ([paper](https://arxiv.org/abs/2410.03859), [dataset](https://huggingface.co/datasets/SWE-bench/SWE-bench_Multimodal)) into this repository! Unlike SWE-bench, we've kept evaluation for the test split *private*. Submit to the leaderboard using [sb-cli](https://github.com/swe-bench/sb-cli/tree/main), our new cloud-based evaluation tool.
* **[Jan. 11, 2025]**: Thanks to [Modal](https://modal.com/), you can now run evaluations entirely on the cloud! See [here](https://github.com/swe-bench/SWE-bench/blob/main/docs/assets/evaluation.md#%EF%B8%8F-evaluation-with-modal) for more details.
* **[Aug. 13, 2024]**: Introducing *SWE-bench Verified*! Part 2 of our collaboration with [OpenAI Preparedness](https://openai.com/preparedness/). A subset of 500 problems that real software engineers have confirmed are solvable. Check out more in the [report](https://openai.com/index/introducing-swe-bench-verified/)!
* **[Jun. 27, 2024]**: We have an exciting update for SWE-bench - with support from [OpenAI's Preparedness](https://openai.com/preparedness/) team: We're moving to a fully containerized evaluation harness using Docker for more reproducible evaluations! Read more in our [report](https://github.com/swe-bench/SWE-bench/blob/main/docs/20240627_docker/README.md).
* **[Apr. 2, 2024]**: We have released [SWE-agent](https://github.com/SWE-agent/SWE-agent), which sets the state-of-the-art on the full SWE-bench test set! ([Tweet 🔗](https://twitter.com/jyangballin/status/1775114444370051582))
* **[Jan. 16, 2024]**: SWE-bench has been accepted to ICLR 2024 as an oral presentation! ([OpenReview 🔗](https://openreview.net/forum?id=VTF8yNQM66))
## 👋 Overview
SWE-bench is a benchmark for evaluating large language models on real world software issues collected from GitHub.
Given a *codebase* and an *issue*, a language model is tasked with generating a *patch* that resolves the described problem.
To access SWE-bench, copy and run the following code:
```python
from datasets import load_dataset
swebench = load_dataset('princeton-nlp/SWE-bench', split='test')
```
## 🚀 Set Up
SWE-bench uses Docker for reproducible evaluations.
Follow the instructions in the [Docker setup guide](https://docs.docker.com/engine/install/) to install Docker on your machine.
If you're setting up on Linux, we recommend seeing the [post-installation steps](https://docs.docker.com/engine/install/linux-postinstall/) as well.
Finally, to build SWE-bench from source, follow these steps:
```bash
git clone git@github.com:princeton-nlp/SWE-bench.git
cd SWE-bench
pip install -e .
```
Test your installation by running:
```bash
python -m swebench.harness.run_evaluation \
--predictions_path gold \
--max_workers 1 \
--instance_ids sympy__sympy-20590 \
--run_id validate-gold
```
> [!NOTE]
> If using a MacOS M-series or other ARM-based systems, add `--namespace ''` to the above script.
> By default, the evaluation script pulls images (built for Linux) from [DockerHub](https://hub.docker.com/u/swebench).
> Adding `--namespace ''` will cause evaluation images to be built locally instead.
## 💽 Usage
Evaluate patch predictions on SWE-bench Lite with the following command:
```bash
python -m swebench.harness.run_evaluation \
--dataset_name princeton-nlp/SWE-bench_Lite \
--predictions_path \
--max_workers \
--run_id
# use --predictions_path 'gold' to verify the gold patches
# use --run_id to name the evaluation run
# use --modal true to run on Modal
```
This command will generate docker build logs (`logs/build_images`) and evaluation logs (`logs/run_evaluation`) in the current directory.
The final evaluation results will be stored in the `evaluation_results` directory.
> [!WARNING]
> SWE-bench evaluation can be resource intensive
> We recommend running on an `x86_64` machine with at least 120GB of free storage, 16GB of RAM, and 8 CPU cores.
> We recommend using fewer than `min(0.75 * os.cpu_count(), 24)` for `--max_workers`.
>
> If running with Docker desktop, make sure to increase your virtual disk space to ~120 free GB. Set max_workers to be consistent with the above for the CPUs available to Docker.
>
> Support for `arm64` machines is experimental.
To see the full list of arguments for the evaluation harness, run:
```bash
python -m swebench.harness.run_evaluation --help
```
See the [evaluation tutorial](docs/guides/evaluation.md) for the full rundown on datasets you can evaluate.
If you're looking for non-local, cloud based evaluations, check out...
* [sb-cli](https://github.com/swe-bench/sb-cli), our tool for running evaluations automatically on AWS, or...
* Running SWE-bench evaluation on [Modal](https://modal.com/). Details [here](docs/guides/evaluation.md#Cloud-Based-Evaluation)
Additionally, you can also:
* [Train](https://github.com/swe-bench/SWE-bench/tree/main/swebench/inference/make_datasets) your own models on our pre-processed datasets. (🆕 Check out [SWE-smith](https://swesmith.com/), a dedicated toolkit for creating SWE training data.)
* Run [inference](docs/reference/inference.md) on existing models (both local and API models). The inference step is where you give the model a repo + issue and have it generate a fix.
* Run SWE-bench's [data collection procedure](https://github.com/swe-bench/SWE-bench/blob/main/swebench/collect/) ([tutorial](docs/guides/collection.md)) on your own repositories, to make new SWE-Bench tasks.
* ⚠️ We are temporarily pausing support for queries around creating SWE-bench instances. Please see the note in the tutorial.
## ⬇️ Downloads
| Datasets | Models | RAG |
| - | - | - |
| [💿 SWE-bench](https://huggingface.co/datasets/SWE-bench/SWE-bench) | [🦙 SWE-Llama 13b](https://huggingface.co/princeton-nlp/SWE-Llama-13b) | [🤗 "Oracle" Retrieval](https://huggingface.co/datasets/SWE-bench/SWE-bench_oracle) |
| [💿 SWE-bench Lite](https://huggingface.co/datasets/SWE-bench/SWE-bench_Lite) | [🦙 SWE-Llama 13b (PEFT)](https://huggingface.co/princeton-nlp/SWE-Llama-13b-peft) | [🤗 BM25 Retrieval 13K](https://huggingface.co/datasets/SWE-bench/SWE-bench_bm25_13K) |
| [💿 SWE-bench Verified](https://huggingface.co/datasets/SWE-bench/SWE-bench_Verified) | [🦙 SWE-Llama 7b](https://huggingface.co/princeton-nlp/SWE-Llama-7b) | [🤗 BM25 Retrieval 27K](https://huggingface.co/datasets/SWE-bench/SWE-bench_bm25_27K) |
| [💿 SWE-bench Multimodal](https://huggingface.co/datasets/SWE-bench/SWE-bench_Multimodal) | [🦙 SWE-Llama 7b (PEFT)](https://huggingface.co/princeton-nlp/SWE-Llama-7b-peft) | [🤗 BM25 Retrieval 40K](https://huggingface.co/datasets/SWE-bench/SWE-bench_bm25_40K) |
| | | [🤗 BM25 Retrieval 50K (Llama tokens)](https://huggingface.co/datasets/SWE-bench/SWE-bench_bm25_50k_llama) |
## 💫 Contributions
We would love to hear from the broader NLP, Machine Learning, and Software Engineering research communities, and we welcome any contributions, pull requests, or issues!
To do so, please either file a new pull request or issue and fill in the corresponding templates accordingly. We'll be sure to follow up shortly!
Contact person: [Carlos E. Jimenez](http://www.carlosejimenez.com/) and [John Yang](https://john-b-yang.github.io/) (Email: carlosej@princeton.edu, johnby@stanford.edu).
## ✍️ Citation & license
MIT license. Check `LICENSE.md`.
If you find our work helpful, please use the following citations.
```bibtex
@inproceedings{
jimenez2024swebench,
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 R Narasimhan},
booktitle={The Twelfth International Conference on Learning Representations},
year={2024},
url={https://openreview.net/forum?id=VTF8yNQM66}
}
@inproceedings{
yang2024swebenchmultimodal,
title={{SWE}-bench Multimodal: Do AI Systems Generalize to Visual Software Domains?},
author={John Yang and Carlos E. Jimenez and Alex L. Zhang and Kilian Lieret and Joyce Yang and Xindi Wu and Ori Press and Niklas Muennighoff and Gabriel Synnaeve and Karthik R. Narasimhan and Diyi Yang and Sida I. Wang and Ofir Press},
booktitle={The Thirteenth International Conference on Learning Representations},
year={2025},
url={https://openreview.net/forum?id=riTiq3i21b}
}
```
## Our Other Projects
