Sagnik Ray Choudhury
commited on
Commit
·
1c0cd8c
1
Parent(s):
5c99d5b
first commit
Browse files- README.md +5 -0
- snli-bt.py +72 -0
- test.jsonl +3 -0
- train.jsonl +3 -0
- validation.jsonl +3 -0
README.md
CHANGED
@@ -1,3 +1,8 @@
|
|
|
|
1 |
---
|
2 |
license: afl-3.0
|
3 |
---
|
|
|
|
|
|
|
|
|
|
1 |
+
<<<<<<< HEAD
|
2 |
---
|
3 |
license: afl-3.0
|
4 |
---
|
5 |
+
=======
|
6 |
+
### Dataset Card for SNLI Back Translation
|
7 |
+
back translation of SNLI dataset: only use the test version
|
8 |
+
>>>>>>> 2afd18c (first commit)
|
snli-bt.py
ADDED
@@ -0,0 +1,72 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import json
|
2 |
+
|
3 |
+
import datasets
|
4 |
+
|
5 |
+
|
6 |
+
logger = datasets.logging.get_logger(__name__)
|
7 |
+
_URL = "https://huggingface.co/datasets/sagnikrayc/snli-bt/resolve/main"
|
8 |
+
_URLS = {
|
9 |
+
"train": f"{_URL}/train.jsonl",
|
10 |
+
"validation": f"{_URL}/validation.jsonl",
|
11 |
+
"test": f"{_URL}/test.jsonl",
|
12 |
+
}
|
13 |
+
|
14 |
+
|
15 |
+
class SnliBTConfig(datasets.BuilderConfig):
|
16 |
+
def __init__(self, **kwargs):
|
17 |
+
super(SnliBTConfig, self).__init__(**kwargs)
|
18 |
+
|
19 |
+
|
20 |
+
class SnliCF(datasets.GeneratorBasedBuilder):
|
21 |
+
"""SQUAD: The Stanford Question Answering Dataset. Version 1.1."""
|
22 |
+
|
23 |
+
BUILDER_CONFIGS = [
|
24 |
+
SnliBTConfig(
|
25 |
+
name="plain_text",
|
26 |
+
version=datasets.Version("1.0.0", ""),
|
27 |
+
description="Plain text",
|
28 |
+
),
|
29 |
+
]
|
30 |
+
|
31 |
+
def _info(self):
|
32 |
+
return datasets.DatasetInfo(
|
33 |
+
description="NA",
|
34 |
+
features=datasets.Features(
|
35 |
+
{
|
36 |
+
"idx": datasets.Value("string"),
|
37 |
+
"premise": datasets.Value("string"),
|
38 |
+
"hypothesis": datasets.Value("string"),
|
39 |
+
"label": datasets.Value("string"),
|
40 |
+
"_type": datasets.Value("string"),
|
41 |
+
}
|
42 |
+
),
|
43 |
+
# No default supervised_keys (as we have to pass both question
|
44 |
+
# and context as input).
|
45 |
+
supervised_keys=None
|
46 |
+
)
|
47 |
+
|
48 |
+
def _split_generators(self, dl_manager):
|
49 |
+
downloaded_files = dl_manager.download_and_extract(_URLS)
|
50 |
+
|
51 |
+
return [
|
52 |
+
datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": downloaded_files["train"]}),
|
53 |
+
datasets.SplitGenerator(
|
54 |
+
name=datasets.Split.VALIDATION, gen_kwargs={"filepath": downloaded_files["validation"]}
|
55 |
+
),
|
56 |
+
datasets.SplitGenerator(name=datasets.Split.TEST, gen_kwargs={"filepath": downloaded_files["test"]}),
|
57 |
+
]
|
58 |
+
|
59 |
+
def _generate_examples(self, filepath):
|
60 |
+
"""This function returns the examples in the raw (text) form."""
|
61 |
+
logger.info("generating examples from = %s", filepath)
|
62 |
+
with open(filepath, encoding="utf-8") as rf:
|
63 |
+
for idx, line in enumerate(rf):
|
64 |
+
if line:
|
65 |
+
_line = json.loads(line)
|
66 |
+
yield idx, {
|
67 |
+
"premise": _line["premise"],
|
68 |
+
"hypothesis": _line["hypothesis"],
|
69 |
+
"idx": _line["idx"],
|
70 |
+
"_type": _line["_type"],
|
71 |
+
"label": _line["label"]
|
72 |
+
}
|
test.jsonl
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2e0ffeb1198abdd717a20257ed3067ba93422b9028835d4d76017fec5041df01
|
3 |
+
size 3860362
|
train.jsonl
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2e0ffeb1198abdd717a20257ed3067ba93422b9028835d4d76017fec5041df01
|
3 |
+
size 3860362
|
validation.jsonl
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2e0ffeb1198abdd717a20257ed3067ba93422b9028835d4d76017fec5041df01
|
3 |
+
size 3860362
|