How It Works
Required Fields
The model returns one entry for each correction, with these fields:
- Type of edit: One of
insertion,deletion,replacement,punctuation,capitalization, oritalicize - Original text: The text as it appeared before the edit
- Corrected text: The intended version after applying the edit
- Line number: The line on which the edit occurs. Use line 0 for titles or headings, and start counting full lines of body text from line 1
- Page: A known identifier for the image (e.g.,
"001.png"), provided alongside the input and not extracted by the model
Common editing marks include:
- Insertions: Indicated by caret marks (
^) or added words between lines - Deletions: Shown using strikethroughs or crossed-out text
- Replacements: Circled, underlined, or bracketed text with substitutions nearby
- Punctuation edits: Handwritten punctuation added, removed, or modified
- Capitalization: Case changes marked explicitly or via notation
- Italicize: Text that should be formatted in italics, typically indicated by underlining or special notation
Example Input
Sample page from Little Dorrit with editorial marks.
Expected Output
This page contains ten corrections. The expected annotation is shown below:
{
"image": "001.png",
"page_number": 5,
"source": "Little Dorrit",
"annotator": "pairsys",
"annotation_date": "2025-04-04",
"verified": true,
"edits": [
{
"type": "punctuation",
"original_text": "church bells",
"corrected_text": "church bells,",
"line_number": 2,
"page": "001.png"
},
{
"type": "punctuation",
"original_text": "wine bottles",
"corrected_text": "wine-bottles",
"line_number": 11,
"page": "001.png"
},
{
"type": "punctuation",
"original_text": "got through",
"corrected_text": "got, through",
"line_number": 14,
"page": "001.png"
},
{
"type": "punctuation",
"original_text": "iron bars fashioned",
"corrected_text": "iron bars, fashioned",
"line_number": 14,
"page": "001.png"
},
{
"type": "punctuation",
"original_text": "grating where",
"corrected_text": "grating, where",
"line_number": 17,
"page": "001.png"
},
{
"type": "punctuation",
"original_text": "outside and",
"corrected_text": "outside; and",
"line_number": 29,
"page": "001.png"
},
{
"type": "punctuation",
"original_text": "intact in",
"corrected_text": "intact, in",
"line_number": 30,
"page": "001.png"
},
{
"type": "capitalization",
"original_text": "indian ocean",
"corrected_text": "Indian Ocean",
"line_number": 31,
"page": "001.png"
},
{
"type": "punctuation",
"original_text": "was waiting to be fed looking",
"corrected_text": "was waiting to be fed; looking",
"line_number": 36,
"page": "001.png"
},
{
"type": "punctuation",
"original_text": "bars that",
"corrected_text": "bars, that",
"line_number": 36,
"page": "001.png"
}
]
}Latest benchmark run: