---
configs:
- config_name: default
  data_files:
  - split: dedup_C
    path: data/dedup_C-*
  - split: dup_C
    path: data/dup_C-*
  - split: dedup_DecomC
    path: data/dedup_DecomC-*
  - split: dup_DecomC
    path: data/dup_DecomC-*
  - split: dedup_demiStripped
    path: data/dedup_demiStripped-*
  - split: dup_demiStripped
    path: data/dup_demiStripped-*
  - split: no_fun_demiStripped
    path: data/no_fun_demiStripped-*
  - split: dup_stripped
    path: data/dup_stripped-*
  - split: dedup_stripped
    path: data/dedup_stripped-*
dataset_info:
  features:
  - name: id
    dtype: int64
  - name: docstring_tokens
    sequence: string
  - name: code_tokens
    sequence: string
  - name: fun_name
    dtype: string
  - name: repo
    dtype: string
  - name: starting
    dtype: string
  - name: partition
    dtype: string
  - name: __index_level_0__
    dtype: int64
  splits:
  - name: dedup_C
    num_bytes: 167770495
    num_examples: 79673
  - name: dup_C
    num_bytes: 348707539
    num_examples: 214587
  - name: dedup_DecomC
    num_bytes: 330052224
    num_examples: 79673
  - name: dup_DecomC
    num_bytes: 614158883
    num_examples: 214587
  - name: dedup_demiStripped
    num_bytes: 316991021
    num_examples: 79673
  - name: dup_demiStripped
    num_bytes: 590234671
    num_examples: 214587
  - name: no_fun_demiStripped
    num_bytes: 606914210
    num_examples: 214587
  - name: dup_stripped
    num_bytes: 60563000
    num_examples: 14245
  - name: dedup_stripped
    num_bytes: 40485701
    num_examples: 7826
  download_size: 592873091
  dataset_size: 3075877744
license: apache-2.0
task_categories:
- summarization
tags:
- code
- Reverse Engineering
- Binary
- Code Summarization
size_categories:
- 100K<n<1M
---
# Dataset Card for "Capybara"

## Dataset Description

- **Repository: https://github.com/AISE-TUDelft/Capybara-BinT5**
- **Paper: https://huggingface.co/papers/2301.01701**
- **Point of Contact: https://huggingface.co/aalkaswan**
- **Raw Data: https://zenodo.org/records/7229809**

### Dataset Summary

Dataset used to train [BinT5](https://huggingface.co/collections/AISE-TUDelft/bint5-65bd006a8c90bd5c97485244). Please refer to the paper for more information. 


### Citation Information

```
@inproceedings{alkaswan2023extending,
  title={Extending Source Code Pre-Trained Language Models to Summarise Decompiled Binaries},
  author={Al-Kaswan, Ali and Ahmed, Toufique and Izadi, Maliheh and Sawant, Anand Ashok and Devanbu, Premkumar and van Deursen, Arie},
  booktitle={2023 IEEE International Conference on Software Analysis, Evolution and Reengineering (SANER)},
  pages={260--271},
  year={2023},
  organization={IEEE}
}
```