Datasets:
Tasks:
Image Classification
Sub-tasks:
multi-class-image-classification
Languages:
English
Size:
100K<n<1M
ArXiv:
License:
Update rvl_cdip_easyocr.py
Browse files- rvl_cdip_easyocr.py +5 -5
rvl_cdip_easyocr.py
CHANGED
@@ -71,8 +71,11 @@ _CLASSES = [
|
|
71 |
"memo",
|
72 |
]
|
73 |
|
74 |
-
_IMAGES_DIR = "images/"
|
75 |
-
|
|
|
|
|
|
|
76 |
|
77 |
|
78 |
# class OCRConfig(datasets.BuilderConfig):
|
@@ -195,9 +198,6 @@ class RvlCdipEasyOcr(datasets.GeneratorBasedBuilder):
|
|
195 |
with open(labels_filepath, encoding="utf-8") as f:
|
196 |
data = f.read().splitlines()
|
197 |
|
198 |
-
# hardcoded to not get stuck in annoying IO and LFS problems in Hub
|
199 |
-
_OCR_DIR = "/cw/liir_data/NoCsBack/jordy/BDPC"
|
200 |
-
|
201 |
image_to_OCR = self._get_image_to_OCR(_OCR_DIR, split)
|
202 |
image_to_class_id = self._get_image_to_class_map(data)
|
203 |
|
|
|
71 |
"memo",
|
72 |
]
|
73 |
|
74 |
+
_IMAGES_DIR = "images/"
|
75 |
+
# hardcoded to not get stuck in annoying IO and LFS problems in Hub
|
76 |
+
_OCR_DIR = "/cw/liir_data/NoCsBack/jordy/BDPC"
|
77 |
+
_OCR_DIR = _OCR_DIR if os.path.exists(_OCR_DIR) else "data/"
|
78 |
+
|
79 |
|
80 |
|
81 |
# class OCRConfig(datasets.BuilderConfig):
|
|
|
198 |
with open(labels_filepath, encoding="utf-8") as f:
|
199 |
data = f.read().splitlines()
|
200 |
|
|
|
|
|
|
|
201 |
image_to_OCR = self._get_image_to_OCR(_OCR_DIR, split)
|
202 |
image_to_class_id = self._get_image_to_class_map(data)
|
203 |
|