LandyGuo commited on
Commit
c8bee36
·
1 Parent(s): f3ce60b

fix training data

Browse files
Files changed (2) hide show
  1. dataset.py +38 -0
  2. train/train.jsonl +2 -2
dataset.py ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ mport os
2
+ import json
3
+ from datasets import DatasetInfo, GeneratorBasedBuilder, SplitGenerator, Split
4
+
5
+ class AReaLBoba2RLCode(GeneratorBasedBuilder):
6
+ VERSION = "1.0.0"
7
+
8
+ def _info(self):
9
+ return DatasetInfo(
10
+ description="AReaL-boba-2-RL-Code dataset (only train split)",
11
+ features={
12
+ "input": {"dtype": "string", "id": None},
13
+ "output": {"dtype": "string", "id": None},
14
+ },
15
+ supervised_keys=None,
16
+ )
17
+
18
+ def _split_generators(self, dl_manager):
19
+ data_dir = dl_manager.download_and_extract(self.config.data_dir or ".")
20
+ train_path = os.path.join(data_dir, "train", "train.jsonl")
21
+
22
+ return [
23
+ SplitGenerator(
24
+ name=Split.TRAIN,
25
+ gen_kwargs={"filepath": train_path},
26
+ ),
27
+ ]
28
+
29
+ def _generate_examples(self, filepath):
30
+ with open(filepath, encoding="utf-8") as f:
31
+ for idx, line in enumerate(f):
32
+ if not line.strip():
33
+ continue
34
+ data = json.loads(line)
35
+ yield idx, {
36
+ "input": data.get("input", ""),
37
+ "output": data.get("output", ""),
38
+ }
train/train.jsonl CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:02c7abb1a272ad13e381390560363d668cca1b00c39ede7b08775e6fdf2dbe94
3
- size 6869752293
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:07196b4fbc74e8961bc458261a5f1c361c9f7978482682218188937148342b3e
3
+ size 6869950785