mary-lev/llm-ocr
LLM-powered OCR evaluation and correction package that supports multiple language models for OCR processing and text correction tasks.
This tool helps researchers, archivists, and historians accurately convert scanned historical documents or images into digital text, even from challenging sources like old books. You provide images (like JPEGs) and their corresponding ALTO XML layout files, and the system outputs highly accurate, corrected text. It's designed for anyone working with physical documents that need precise digital conversion for analysis or archiving.
No commits in the last 6 months.
Use this if you need to extract and correct text from scanned documents, especially those with historical or complex layouts, and want to leverage advanced AI models for superior accuracy.
Not ideal if you only need basic OCR for modern, clean documents or if you prefer not to use external Large Language Model services.
Stars
4
Forks
1
Language
Python
License
MIT
Category
Last pushed
Jun 24, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/transformers/mary-lev/llm-ocr"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
FuxiaoLiu/LRV-Instruction
[ICLR'24] Mitigating Hallucination in Large Multi-Modal Models via Robust Instruction Tuning
kiyoshisasano/llm-failure-atlas
A graph-based failure modeling and deterministic detection system for LLM agent runtimes.
gwasiakshay/llm-eval-benchmark
LLM evaluation & benchmarking framework using LLM-as-a-judge scoring, multi-model comparison,...
useentropy/llmkit
LLM Kit - Python Large Language Model Kit for generating data of your choice
flamehaven01/CRoM-EfficientLLM
A Python toolkit to optimize LLM context by intelligently selecting, re-ranking, and...