Too much text on a single PDF page may result in no output at all. (maybe solved)

## Example:
[ocean-2 Extract\[3\].pdf](https://github.com/user-attachments/files/23237613/ocean-2.Extract.3.pdf)

ran in vllm.

## Problem:
If there are too many text in one page or in one picture, then there will be no output.

## Solution (maybe) :
1. increaseing the max_model_len.
2. manully Split one page of content into multiple pages.