AnonNo2
commited on
Commit
•
ede0033
1
Parent(s):
aad7903
fir commit
Browse files- my_dataset.py +46 -0
- test_datasets.jsonl +0 -0
my_dataset.py
ADDED
@@ -0,0 +1,46 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
|
2 |
+
|
3 |
+
|
4 |
+
from datasets import DatasetInfo, Features, Split, SplitGenerator, GeneratorBasedBuilder, Value, Sequence
|
5 |
+
import json
|
6 |
+
|
7 |
+
class MyDataset(GeneratorBasedBuilder):
|
8 |
+
def _info(self):
|
9 |
+
return DatasetInfo(
|
10 |
+
features=Features({
|
11 |
+
"questions": Sequence(Value("string")),
|
12 |
+
"answers": Sequence(Value("string"))
|
13 |
+
}),
|
14 |
+
supervised_keys=("questions", "answers"),
|
15 |
+
homepage="https://github.com/FreedomIntelligence/HuatuoGPT",
|
16 |
+
citation="...",
|
17 |
+
)
|
18 |
+
|
19 |
+
def _split_generators(self, dl_manager):
|
20 |
+
train_path = "train_datasets.jsonl"
|
21 |
+
validation_path = "validation_datasets.jsonl"
|
22 |
+
test_path = "test_datasets.jsonl"
|
23 |
+
|
24 |
+
return [
|
25 |
+
SplitGenerator(name=Split.TRAIN, gen_kwargs={"filepath": train_path}),
|
26 |
+
SplitGenerator(name=Split.VALIDATION, gen_kwargs={"filepath": validation_path}),
|
27 |
+
SplitGenerator(name=Split.TEST, gen_kwargs={"filepath": test_path}),
|
28 |
+
]
|
29 |
+
|
30 |
+
def _generate_examples(self, filepath):
|
31 |
+
with open(filepath, encoding="utf-8") as f:
|
32 |
+
for id_, row in enumerate(f):
|
33 |
+
# Process your data here and create a dictionary with the features.
|
34 |
+
# For example, if your data is in JSON format:
|
35 |
+
data = json.loads(row)
|
36 |
+
yield id_, {
|
37 |
+
"questions": data["questions"],
|
38 |
+
"answers": data["answers"],
|
39 |
+
}
|
40 |
+
|
41 |
+
if __name__ == '__main__':
|
42 |
+
from datasets import load_dataset
|
43 |
+
|
44 |
+
dataset = load_dataset("my_dataset.py")
|
45 |
+
|
46 |
+
print()
|
test_datasets.jsonl
ADDED
The diff for this file is too large to render.
See raw diff
|
|