MMOCR

A comprehensive OCR toolbox and benchmark that supports text detection, recognition, and downstream tasks like key information extraction. It includes multiple state-of-the-art models and provides utilities for performance assessment, visualization, and data conversion.

from benchthing import Bench

bench = Bench("mmocr")

bench.run(
    benchmark="mmocr",
    task_id="1",
    models=yourOCRModels
)

result = bench.get_result("1")