File size: 2,119 Bytes
907871c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
619569d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
907871c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
---
annotations_creators:
- expert-generated
language_creators:
- expert-generated
language:
- code
license:
- mit
multilinguality:
- monolingual
source_datasets:
- original
task_categories:
- text2text-generation
task_ids: []
pretty_name: OpenAI HumanEval-Infilling
tags:
- code-generation
configs:
- config_name: MultiLine
  data_files:
  - split: test
    path: MultiLine/test.jsonl

- config_name: RandomSpan
  data_files:
  - split: test
    path: RandomSpan/test.jsonl

- config_name: RandomSpanLight
  data_files:
  - split: test
    path: RandomSpanLight/test.jsonl

- config_name: SingleLine
  data_files:
  - split: test
    path: SingleLine/test.jsonl
---

# HumanEval-Infilling


## Dataset Description

- **Repository:**  https://github.com/openai/human-eval-infilling
- **Paper:** https://arxiv.org/pdf/2207.14255

## Dataset Summary

[HumanEval-Infilling](https://github.com/openai/human-eval-infilling) is a benchmark for infilling tasks, derived from [HumanEval](https://huggingface.co/datasets/openai_humaneval) benchmark for the evaluation of code generation models.

## Dataset Structure
To load the dataset you need to specify a subset. By default `HumanEval-SingleLineInfilling` is loaded.

```python
from datasets import load_dataset
ds = load_dataset("humaneval_infilling", "HumanEval-RandomSpanInfilling")

DatasetDict({
    test: Dataset({
        features: ['task_id', 'entry_point', 'prompt', 'suffix', 'canonical_solution', 'test'],
        num_rows: 1640
    })
})
```

## Subsets

This dataset has 4 subsets: HumanEval-MultiLineInfilling, HumanEval-SingleLineInfilling, HumanEval-RandomSpanInfilling, HumanEval-RandomSpanInfillingLight.
The single-line, multi-line, random span infilling and its light version have 1033, 5815, 1640 and 164 tasks, respectively.

## Citation

```
@article{bavarian2022efficient,
  title={Efficient Training of Language Models to Fill in the Middle},
  author={Bavarian, Mohammad and Jun, Heewoo and Tezak, Nikolas and Schulman, John and McLeavey, Christine and Tworek, Jerry and Chen, Mark},
  journal={arXiv preprint arXiv:2207.14255},
  year={2022}
}
```