General OCR Theory: Towards OCR-2.0 via a Unified End-to-end Model
[😊 Hugging Face] [📜 Paper] [🌟 GitHub]"🔥🔥🔥This is the official online demo of GOT-OCR-2.0 model!!!"
Demo Guidelines
You need to upload your image below and choose one mode of GOT, then click "Submit" to run GOT model. More characters will result in longer wait times.
- plain texts OCR & format texts OCR: The two modes are for the image-level OCR.
- plain multi-crop OCR & format multi-crop OCR: For images with more complex content, you can achieve higher-quality results with these modes.
- plain fine-grained OCR & format fine-grained OCR: In these modes, you can specify fine-grained regions on the input image for more flexible OCR. Fine-grained regions can be coordinates of the box, red color, blue color, or green color.
Choose one mode of GOT
If you choose the mode with format, the mathpix result will be automatically rendered as follows:
examples
| upload your image | Choose one mode of GOT | fine-grained type | color list | input box: [x1,y1,x2,y2] |
|---|