Files
ocr-server/tests
zimoyin a7d4ff6454 feat: add enable_detection switch to skip text detection
enable_detection: false 跳过检测模型,直接把整图当一行文本送识别。
实测:单行图可正确识别;多行整图因识别模型固定缩放到高 48px 会返回空结果,已加告警。

Co-Authored-By: Claude Code <noreply@anthropic.com>
2026-10-13 15:59:25 +08:00
..
2026-10-11 02:26:26 +08:00