为什么启用了ocr_config的OCRModelType.MEDIUM， 还是运行PP-OCRv6_rec_small

为什么启用了ocr_config的OCRModelType.MEDIUM， 还是运行PP-OCRv6_rec_small

代码是：
from rapidocr import EngineType as OCREngineType, OCRVersion, ModelType as OCRModelType
ocr_config = {
"Det.engine_type": OCREngineType.ONNXRUNTIME,
"Det.model_type": OCRModelType.MEDIUM,
"Det.ocr_version": OCRVersion.PPOCRV6,
"Rec.engine_type": OCREngineType.ONNXRUNTIME,
"Rec.model_type": OCRModelType.MEDIUM,
"Rec.ocr_version": OCRVersion.PPOCRV6,
}

image_config = {
"extract_original_image": True,
}
layout_config = {
"markdown_ignore_labels": []
}
table_config = {
"skip_text_in_image": False,
"use_compare_table": True, # 启用表格结果比较（同时跑有线/无线并比对），默认 False
"table_image_enable": True, # 表格内图片识别
"extract_original_image": True # 是否提取表格内原始图片，默认 False
}

engine = RapidDoc(table_config=table_config, image_config=image_config, layout_config=layout_config,
ocr_config=ocr_config)

ocr_config已经设置了OCRModelType.MEDIUM， 但是日志里没有显示medium的相关内容：

[INFO] 2026-08-12 11:01:26,716 [RapidOCR] base.py:23: Using engine_name: onnxruntime
[INFO] 2026-08-12 11:01:26,718 [RapidOCR] main.py:63: Using D:\ProgramData\PyCharmProject\test-python\venv\lib\site-packages\rapid_doc\resources\PP-OCRv6_det_small.onnx
[INFO] 2026-08-12 11:01:26,765 [RapidOCR] base.py:23: Using engine_name: onnxruntime
[INFO] 2026-08-12 11:01:26,768 [RapidOCR] download_file.py:60: File exists and is valid: D:\ProgramData\PyCharmProject\test-python\venv\Lib\site-packages\rapidocr\models\ch_ppocr_mobile_v2.0_cls_mobile.onnx
[INFO] 2026-08-12 11:01:26,768 [RapidOCR] main.py:63: Using D:\ProgramData\PyCharmProject\test-python\venv\Lib\site-packages\rapidocr\models\ch_ppocr_mobile_v2.0_cls_mobile.onnx
[INFO] 2026-08-12 11:01:26,857 [RapidOCR] base.py:23: Using engine_name: onnxruntime
[INFO] 2026-08-12 11:01:26,858 [RapidOCR] main.py:63: Using D:\ProgramData\PyCharmProject\test-python\venv\lib\site-packages\rapid_doc\resources\PP-OCRv6_rec_small.onnx
[INFO] 2026-08-12 11:01:28,308 [RapidOCR] base.py:23: Using engine_name: onnxruntime
[INFO] 2026-08-12 11:01:28,309 [RapidOCR] main.py:63: Using D:\ProgramData\PyCharmProject\test-python\venv\lib\site-packages\rapid_doc\resources\PP-OCRv6_det_small.onnx
[INFO] 2026-08-12 11:01:29,149 [RapidOCR] base.py:23: Using engine_name: onnxruntime
[INFO] 2026-08-12 11:01:29,152 [RapidOCR] download_file.py:60: File exists and is valid: D:\ProgramData\PyCharmProject\test-python\venv\Lib\site-packages\rapidocr\models\ch_ppocr_mobile_v2.0_cls_mobile.onnx
[INFO] 2026-08-12 11:01:29,153 [RapidOCR] main.py:63: Using D:\ProgramData\PyCharmProject\test-python\venv\Lib\site-packages\rapidocr\models\ch_ppocr_mobile_v2.0_cls_mobile.onnx
[INFO] 2026-08-12 11:01:29,831 [RapidOCR] base.py:23: Using engine_name: onnxruntime
[INFO] 2026-08-12 11:01:29,832 [RapidOCR] main.py:63: Using D:\ProgramData\PyCharmProject\test-python\venv\lib\site-packages\rapid_doc\resources\PP-OCRv6_rec_small.onnx

[INFO] 2026-08-12 11:02:03,266 [RapidOCR] base.py:23: Using engine_name: openvino
[INFO] 2026-08-12 11:02:03,268 [RapidOCR] main.py:56: Using D:\ProgramData\PyCharmProject\test-python\venv\lib\site-packages\rapid_doc\resources\pp-ocrv4_mobile_seal_det.onnx
[INFO] 2026-08-12 11:02:03,269 [RapidOCR] device_config.py:50: Using OpenVINO config: {}
[INFO] 2026-08-12 11:02:03,483 [RapidOCR] base.py:23: Using engine_name: onnxruntime
[INFO] 2026-08-12 11:02:03,484 [RapidOCR] download_file.py:60: File exists and is valid: D:\ProgramData\PyCharmProject\test-python\venv\Lib\site-packages\rapidocr\models\ch_ppocr_mobile_v2.0_cls_mobile.onnx
[INFO] 2026-08-12 11:02:03,484 [RapidOCR] main.py:63: Using D:\ProgramData\PyCharmProject\test-python\venv\Lib\site-packages\rapidocr\models\ch_ppocr_mobile_v2.0_cls_mobile.onnx
[INFO] 2026-08-12 11:02:03,519 [RapidOCR] base.py:23: Using engine_name: openvino
[INFO] 2026-08-12 11:02:03,519 [RapidOCR] main.py:56: Using D:\ProgramData\PyCharmProject\test-python\venv\lib\site-packages\rapid_doc\resources\PP-OCRv6_rec_small.onnx
[INFO] 2026-08-12 11:02:03,520 [RapidOCR] device_config.py:50: Using OpenVINO config: {}

Start with the OCR initialization entries shown in main.py and base.py, then follow how the supplied Det.model_type and Rec.model_type values are handled. Compare that selection with the PP-OCRv6_rec_small paths in the logs; done means OCRModelType.MEDIUM selects and reports the expected medium model for both detection and recognition.