Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -45,4 +45,37 @@ configs:
|
|
| 45 |
data_files:
|
| 46 |
- split: test
|
| 47 |
path: data/test-*
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 48 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 45 |
data_files:
|
| 46 |
- split: test
|
| 47 |
path: data/test-*
|
| 48 |
+
license: cc-by-sa-4.0
|
| 49 |
+
language:
|
| 50 |
+
- code
|
| 51 |
+
task_categories:
|
| 52 |
+
- text2text-generation
|
| 53 |
+
tags:
|
| 54 |
+
- code-generation
|
| 55 |
+
pretty_name: DS-1000
|
| 56 |
---
|
| 57 |
+
|
| 58 |
+
This is a reupload of [DS-1000](https://huggingface.co/datasets/xlangai/DS-1000). The metadata dictionary has been extracted into columns and the categorical variables are now `ClassLabel` types, and the dataset is natively a parquet. The features are as follows:
|
| 59 |
+
|
| 60 |
+
| Column | Type |
|
| 61 |
+
|----------------------|-----------------------------------------------------------------------------------------------------------|
|
| 62 |
+
|problem_id |`Value(dtype='int64', id=None)` |
|
| 63 |
+
|prompt |`Value(dtype='string', id=None)` |
|
| 64 |
+
|reference_code |`Value(dtype='string', id=None)` |
|
| 65 |
+
|code_context |`Value(dtype='string', id=None)` |
|
| 66 |
+
|library_problem_id |`Value(dtype='int64', id=None)` |
|
| 67 |
+
|library |`ClassLabel(names=['Matplotlib', 'Numpy', 'Pandas', 'Pytorch', 'Scipy', 'Sklearn', 'Tensorflow'], id=None)`|
|
| 68 |
+
|test_case_cnt |`Value(dtype='int64', id=None)` |
|
| 69 |
+
|perturbation_type |`ClassLabel(names=['Difficult-Rewrite', 'Origin', 'Semantic', 'Surface'], id=None)` |
|
| 70 |
+
|perturbation_origin_id|`Value(dtype='int64', id=None)` |
|
| 71 |
+
|
| 72 |
+
All credits go to the original authors below.
|
| 73 |
+
|
| 74 |
+
---
|
| 75 |
+
|
| 76 |
+
<h1 align="center"> DS-1000 in simplified format </h1>
|
| 77 |
+
|
| 78 |
+
See testing code and more information in the [DS-1000 repo](https://github.com/xlang-ai/DS-1000/).
|
| 79 |
+
|
| 80 |
+
Reformatting credits: Yuhang Lai, Sida Wang
|
| 81 |
+
|