Datasets:

Modalities:
Image
Text
Formats:
parquet
Languages:
Armenian
Size:
< 1K
ArXiv:
License:
Yeva commited on
Commit
5fece1a
·
verified ·
1 Parent(s): f976152

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +0 -3
README.md CHANGED
@@ -42,8 +42,6 @@ dataset_info:
42
  - name: bbox
43
  sequence: float32
44
  length: 4
45
- - name: rotation
46
- dtype: float32
47
  splits:
48
  - name: test
49
  num_bytes: 132031004.0
@@ -104,7 +102,6 @@ Each entry in `annotations has:
104
  | `label` | string | Region type (see label taxonomy below) |
105
  | `transcription` | string | Ground-truth text for the region (`""` where transcription was not collected) |
106
  | `bbox` | float[4] | Bounding box `[x1, y1, x2, y2]` in pixels |
107
- | `rotation` | float | Rotation angle in degrees |
108
 
109
  Regardless of whether the source page was born-digital or scanned (see the `Text layer`
110
  column in Sources below), `bbox` coordinates are always in pixel space matching
 
42
  - name: bbox
43
  sequence: float32
44
  length: 4
 
 
45
  splits:
46
  - name: test
47
  num_bytes: 132031004.0
 
102
  | `label` | string | Region type (see label taxonomy below) |
103
  | `transcription` | string | Ground-truth text for the region (`""` where transcription was not collected) |
104
  | `bbox` | float[4] | Bounding box `[x1, y1, x2, y2]` in pixels |
 
105
 
106
  Regardless of whether the source page was born-digital or scanned (see the `Text layer`
107
  column in Sources below), `bbox` coordinates are always in pixel space matching