Fix bug in download script
Browse files- EgoSpeak.py +18 -8
EgoSpeak.py
CHANGED
@@ -31,14 +31,24 @@ class EgoSpeak(GeneratorBasedBuilder):
|
|
31 |
)
|
32 |
|
33 |
def _split_generators(self, dl_manager):
|
34 |
-
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
|
41 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
42 |
|
43 |
def _generate_examples(self, data_dir):
|
44 |
"""Generate examples from the dataset directory structure."""
|
|
|
31 |
)
|
32 |
|
33 |
def _split_generators(self, dl_manager):
|
34 |
+
# Use the Hugging Face cache directory as the default
|
35 |
+
data_dir = os.path.abspath(dl_manager.manual_dir or dl_manager.cache_dir)
|
36 |
+
|
37 |
+
# If no data is available in the cache or provided manually, raise an error
|
38 |
+
if not os.listdir(data_dir):
|
39 |
+
raise ValueError(
|
40 |
+
"Dataset requires data to be present in the Hugging Face cache or manually specified. "
|
41 |
+
"Please provide the `data_dir` parameter or ensure the dataset is cached correctly. "
|
42 |
+
"For example: `load_dataset('kjunh/EgoSpeak', data_dir='/path/to/dataset')`."
|
43 |
+
)
|
44 |
+
|
45 |
+
return [
|
46 |
+
SplitGenerator(
|
47 |
+
name=Split.TRAIN,
|
48 |
+
gen_kwargs={"data_dir": data_dir},
|
49 |
+
),
|
50 |
+
]
|
51 |
+
|
52 |
|
53 |
def _generate_examples(self, data_dir):
|
54 |
"""Generate examples from the dataset directory structure."""
|