jordiclive commited on
Commit
8acf52d
1 Parent(s): 59131df

Update uspto2.py

Browse files
Files changed (1) hide show
  1. uspto2.py +7 -7
uspto2.py CHANGED
@@ -26,37 +26,38 @@ class PileDomainDataset(GeneratorBasedBuilder):
26
 
27
  def _split_generators(self, dl_manager):
28
  #snapshot_download(repo_id="Multi-Domain-Expert-Layers/uspto", repo_type="dataset")
29
- snapshot_download(repo_id="jordiclive/uspto2", repo_type="dataset")
 
30
 
31
  return [
32
  datasets.SplitGenerator(
33
  name=datasets.Split.TRAIN,
34
  gen_kwargs={
35
- "data_dir": "data/train",
36
  "split": None,
37
  },
38
  ),
39
  datasets.SplitGenerator(
40
  name="validation_pile",
41
  gen_kwargs={
42
- "data_dir": "data/val",
43
  "split": "pile",
44
  },
45
  ),
46
  datasets.SplitGenerator(
47
  name="validation_domain",
48
  gen_kwargs={
49
- "data_dir": "data/val",
50
  "split": "domain",
51
  },
52
  ),
53
  datasets.SplitGenerator(
54
  name="test_pile",
55
- gen_kwargs={"data_dir": "data/test", "split": "pile"},
56
  ),
57
  datasets.SplitGenerator(
58
  name="test_domain",
59
- gen_kwargs={"data_dir": "data/test", "split": "domain"},
60
  ),
61
  ]
62
 
@@ -74,4 +75,3 @@ class PileDomainDataset(GeneratorBasedBuilder):
74
  data = json.loads(line)
75
  idx += 1
76
  yield idx, data
77
-
 
26
 
27
  def _split_generators(self, dl_manager):
28
  #snapshot_download(repo_id="Multi-Domain-Expert-Layers/uspto", repo_type="dataset")
29
+ # dl_manager.download_and_extract("https://huggingface.co/datasets/Multi-Domain-Expert-Layers/uspto/resolve/main/uspto.tar.gz")
30
+ dl_path = snapshot_download(repo_id="jordiclive/my_uspto2", repo_type="dataset")
31
 
32
  return [
33
  datasets.SplitGenerator(
34
  name=datasets.Split.TRAIN,
35
  gen_kwargs={
36
+ "data_dir": os.path.join(dl_path, "data/train"),
37
  "split": None,
38
  },
39
  ),
40
  datasets.SplitGenerator(
41
  name="validation_pile",
42
  gen_kwargs={
43
+ "data_dir": os.path.join(dl_path, "data/val"),
44
  "split": "pile",
45
  },
46
  ),
47
  datasets.SplitGenerator(
48
  name="validation_domain",
49
  gen_kwargs={
50
+ "data_dir": os.path.join(dl_path, "data/val"),
51
  "split": "domain",
52
  },
53
  ),
54
  datasets.SplitGenerator(
55
  name="test_pile",
56
+ gen_kwargs={"data_dir": os.path.join(dl_path, "data/test"), "split": "pile"},
57
  ),
58
  datasets.SplitGenerator(
59
  name="test_domain",
60
+ gen_kwargs={"data_dir": os.path.join(dl_path, "data/test"), "split": "domain"},
61
  ),
62
  ]
63
 
 
75
  data = json.loads(line)
76
  idx += 1
77
  yield idx, data