Open
Description
After deploying OCRFlux-3B using vllm, when multiple images are passed in, only the json parsing result of the first image can be returned, and multiple images cannot be parsed at the same time. I tried another Nanonets-OCR-s, which supports parsing multiple images at the same time.
In addition, the parsing format of images containing merged cells is incorrect. Example input image:
返回结果：
| 1 | 1 | ||
| 2 | 21 | 22 | 23 |
| 3 | 3 | ||
| 4 | 41 | 42 | 43 |
| 5 | 5 | ||
| 6 | 6 |
The first row has only 2 cells.
Metadata
Metadata
Assignees
Labels
No labels