flow_inference.evaluation

Evaluate OCR/HTR inference results stored in Hugging Face datasets.

This module loads an inference result dataset, selects evaluation rows with both ground-truth text and predictions, computes Character Error Rate (CER), uploads evaluation artifacts, and refreshes the dataset README with evaluation metadata.

Classes

Evaluation(evaluation_repo_name, hf_token[, ...])

Run CER evaluation for a Hugging Face inference result dataset.