yujunjun/olmocr
1
0
Fork 0
mirror of https://github.com/allenai/olmocr.git synced 2025-10-15 18:22:16 +00:00
Code Issues Packages Projects Releases Wiki Activity

Big bug fix, moving the prompt to match how training was done, 2.3 point boost on olmocr-bench

Browse Source
This commit is contained in:
Jake Poznanski 2025-05-14 19:51:00 +00:00
parent 28966b9f14
commit ffee4c9740
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
olmocr/pipeline.py
View File

@ -138,8 +138,8 @@ async def build_page_query(local_pdf_path: str, page: int, target_longest_image_
{
"role": "user",
"content": [
{"type": "text", "text": build_finetuning_prompt(anchor_text)},
{"type": "image_url", "image_url": {"url": f"data:image/png;base64,{image_base64}"}},
{"type": "text", "text": build_finetuning_prompt(anchor_text)},
],
}
],