Datasets:
2.1.0
Browse files- MultiPL-E.py +10 -15
- README.md +0 -0
MultiPL-E.py
CHANGED
|
@@ -5,20 +5,15 @@ from pathlib import Path
|
|
| 5 |
logger = datasets.logging.get_logger(__name__)
|
| 6 |
|
| 7 |
_CITATION = """\
|
| 8 |
-
@
|
| 9 |
-
|
| 10 |
-
|
| 11 |
-
|
| 12 |
-
Nguyen, Sydney and Phipps-Costin, Luna and Pinckney, Donald and
|
| 13 |
-
Yee, Ming-Ho and Zi, Yangtian and Anderson, Carolyn Jane and
|
| 14 |
-
Feldman, Molly Q and Guha, Arjun and
|
| 15 |
Greenberg, Michael and Jangda, Abhinav},
|
| 16 |
-
title = {A Scalable and
|
| 17 |
-
|
| 18 |
-
|
| 19 |
-
|
| 20 |
-
}
|
| 21 |
-
"""
|
| 22 |
|
| 23 |
_DESCRIPTION = """\
|
| 24 |
MultiPL-E is a dataset for evaluating large language models for code \
|
|
@@ -72,7 +67,7 @@ class MultiPLE(datasets.GeneratorBasedBuilder):
|
|
| 72 |
srcdata=srcdata,
|
| 73 |
language=language,
|
| 74 |
variation=variation,
|
| 75 |
-
version=datasets.Version("2.
|
| 76 |
for srcdata in _SRCDATA
|
| 77 |
for language in _LANGUAGES
|
| 78 |
for variation in _VARIATIONS
|
|
@@ -103,7 +98,7 @@ class MultiPLE(datasets.GeneratorBasedBuilder):
|
|
| 103 |
|
| 104 |
def _split_generators(self, dl_manager: datasets.DownloadManager):
|
| 105 |
files = dl_manager.download(
|
| 106 |
-
f"https://raw.githubusercontent.com/nuprl/MultiPL-E/
|
| 107 |
)
|
| 108 |
return [
|
| 109 |
datasets.SplitGenerator(
|
|
|
|
| 5 |
logger = datasets.logging.get_logger(__name__)
|
| 6 |
|
| 7 |
_CITATION = """\
|
| 8 |
+
@article{cassano:multipl-e,
|
| 9 |
+
author = {Cassano, Federico and Gouwar, John and Nguyen, Daniel and Nguyen, Sydney and
|
| 10 |
+
Phipps-Costin, Luna and Pinckney, Donald and Yee, Ming-Ho and Zi, Yangtian and
|
| 11 |
+
Anderson, Carolyn Jane and Feldman, Molly Q and Guha, Arjun and
|
|
|
|
|
|
|
|
|
|
| 12 |
Greenberg, Michael and Jangda, Abhinav},
|
| 13 |
+
title = {{MultiPL-E}: A Scalable and Polyglot Approach to Benchmarking Neural Code Generation},
|
| 14 |
+
journal = "{IEEE} Transactions of Software Engineering (TSE)",
|
| 15 |
+
year = 2023
|
| 16 |
+
}"""
|
|
|
|
|
|
|
| 17 |
|
| 18 |
_DESCRIPTION = """\
|
| 19 |
MultiPL-E is a dataset for evaluating large language models for code \
|
|
|
|
| 67 |
srcdata=srcdata,
|
| 68 |
language=language,
|
| 69 |
variation=variation,
|
| 70 |
+
version=datasets.Version("2.1.0"))
|
| 71 |
for srcdata in _SRCDATA
|
| 72 |
for language in _LANGUAGES
|
| 73 |
for variation in _VARIATIONS
|
|
|
|
| 98 |
|
| 99 |
def _split_generators(self, dl_manager: datasets.DownloadManager):
|
| 100 |
files = dl_manager.download(
|
| 101 |
+
f"https://raw.githubusercontent.com/nuprl/MultiPL-E/11b407bd2dd98c8204afea4d20043faf2145c20c/prompts/{self.config.srcdata}-{self.config.language}-{self.config.variation}.json"
|
| 102 |
)
|
| 103 |
return [
|
| 104 |
datasets.SplitGenerator(
|
README.md
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|