Datasets:
Tasks:
Translation
Modalities:
Text
Formats:
parquet
Languages:
code
Size:
10K - 100K
ArXiv:
Tags:
code-to-code
License:
Add paper
Browse files
README.md
CHANGED
@@ -69,13 +69,15 @@ dataset_info:
|
|
69 |
## Dataset Description
|
70 |
|
71 |
- **Homepage:** https://github.com/microsoft/CodeXGLUE/tree/main/Code-Code/code-to-code-trans
|
|
|
72 |
|
73 |
### Dataset Summary
|
74 |
|
75 |
CodeXGLUE code-to-code-trans dataset, available at https://github.com/microsoft/CodeXGLUE/tree/main/Code-Code/code-to-code-trans
|
76 |
|
77 |
The dataset is collected from several public repos, including Lucene(http://lucene.apache.org/), POI(http://poi.apache.org/), JGit(https://github.com/eclipse/jgit/) and Antlr(https://github.com/antlr/).
|
78 |
-
|
|
|
79 |
|
80 |
### Supported Tasks and Leaderboards
|
81 |
|
|
|
69 |
## Dataset Description
|
70 |
|
71 |
- **Homepage:** https://github.com/microsoft/CodeXGLUE/tree/main/Code-Code/code-to-code-trans
|
72 |
+
- **Paper:** https://arxiv.org/abs/2102.04664
|
73 |
|
74 |
### Dataset Summary
|
75 |
|
76 |
CodeXGLUE code-to-code-trans dataset, available at https://github.com/microsoft/CodeXGLUE/tree/main/Code-Code/code-to-code-trans
|
77 |
|
78 |
The dataset is collected from several public repos, including Lucene(http://lucene.apache.org/), POI(http://poi.apache.org/), JGit(https://github.com/eclipse/jgit/) and Antlr(https://github.com/antlr/).
|
79 |
+
|
80 |
+
We collect both the Java and C# versions of the codes and find the parallel functions. After removing duplicates and functions with the empty body, we split the whole dataset into training, validation and test sets.
|
81 |
|
82 |
### Supported Tasks and Leaderboards
|
83 |
|