- Stars
- 596
- Forks
- 38
- + today
- +3
- Created
- 1mo
Ranking data as of November 2, 2026 (UTC).
Star History
Today, hour by hour
Overview
Xiaomi-OCR-0 It ranks #522 on GitTiger, gaining +3 stars on November 2, 2026 (UTC).
The project is written in Python and has 38 forks. It was created 1mo ago.
What it is
Xiaomi-OCR-0 is a Python project whose README summary describes a single model of about 0.8 billion parameters for document parsing and OCR-related understanding. The repository description is only the project name, no topics are listed, and the license field reads Apache-2.0.
Capabilities described
The README highlights benchmark results, which are the project's own claims and were not independently checked here: - Document parsing results on OmniDocBench variants, where the README says it placed first or second in the comparisons shown on its page. - Formula recognition on UniMER-Test and table recognition on PubTabNet. - Historical handwriting and calligraphy, scored on a Chronicles-OCR set covering several script styles. - An averaged score across five further evaluations, cut off in the supplied summary.
Getting started
No installation, model download or inference instructions were supplied. Read the repository README for usage details.
git clone https://github.com/xiaomi-research/Xiaomi-OCR-0.git
cd Xiaomi-OCR-0
# see README for setupFrequently Asked Questions
What is Xiaomi-OCR-0?
According to the README summary, it is a unified model of roughly 0.8 billion parameters built for document parsing and OCR-related understanding.
What tasks does Xiaomi-OCR-0 report results on?
The README summary reports benchmarks for document parsing, formula recognition, table recognition and historical handwriting. These are the project's own claims.
How do I run Xiaomi-OCR-0?
No setup or inference steps were supplied in the evidence, so consult the repository README for instructions.