Create handtest.py
Browse files- handtest.py +38 -0
handtest.py
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import datasets
|
| 2 |
+
|
| 3 |
+
_TAR_FILES=[
|
| 4 |
+
" tmp.tar.gz"
|
| 5 |
+
]
|
| 6 |
+
|
| 7 |
+
class ZipDataset(datasets.GeneratorBasedBuilder):
|
| 8 |
+
|
| 9 |
+
def _info(self):
|
| 10 |
+
return datasets.DatasetInfo(
|
| 11 |
+
description="TMP description",
|
| 12 |
+
homepage="google it",
|
| 13 |
+
citation="lmao",
|
| 14 |
+
license="dunno, tbh, assume the worst, k thx."
|
| 15 |
+
)
|
| 16 |
+
|
| 17 |
+
def _split_generators(self, dl_manager):
|
| 18 |
+
archive_path = dl_manager.download(_TAR_FILES[0])
|
| 19 |
+
return [
|
| 20 |
+
datasets.SplitGenerator(
|
| 21 |
+
name=datasets.Split.TRAIN,
|
| 22 |
+
gen_kwargs={
|
| 23 |
+
"images": dl_manager.iter_archive(archive_path),
|
| 24 |
+
},
|
| 25 |
+
),
|
| 26 |
+
]
|
| 27 |
+
|
| 28 |
+
def _generate_examples(self, images):
|
| 29 |
+
"""Generate images and labels for splits."""
|
| 30 |
+
for file_path, file_obj in images:
|
| 31 |
+
yield file_path, {
|
| 32 |
+
"image": {"path": file_path, "bytes": file_obj.read()},
|
| 33 |
+
}
|
| 34 |
+
|
| 35 |
+
|
| 36 |
+
|
| 37 |
+
#https://huggingface.co/datasets/oscar-corpus/OSCAR-2201/blob/main/OSCAR-2201.py
|
| 38 |
+
#https://huggingface.co/datasets/food101
|