mindspore-lab / mindocr

A toolbox of ocr models and algorithms based on MindSpore

Home Page:https://mindspore-lab.github.io/mindocr/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

MindOCR推理export.py时--model_name_or_config configs/det/dbnet/db++_r50_icdar15.yaml失败

tacyi opened this issue · comments

MindOCR推理export.py时 --model_name_or_config dbnetpp_resnet50成功,--model_name_or_config configs/det/dbnet/db++_r50_icdar15.yaml失败。原因是db++_r50_icdar15.yaml中的字段++校验出错。
具体报错信息如下:
[2023-11-30 18:51:59] mindocr.models.utils.load_model INFO - Finish loading model checkoint from /nfs_for_sync/Mindlab_data/dataset/mindocr_infer_ckpt/dbnetpp_resnet50-068166c2.ckpt. If no parameter fail-load warning displayed, all checkpoint params have been successfully loaded.
[2023-11-30 18:51:59] mindocr.export INFO - Set the AMP level of the model to be O0.
Traceback (most recent call last):
File "/data1/jenkins/workspace/source_code/mindocr//tools/export.py", line 171, in
export(**vars(args))
File "/data1/jenkins/workspace/source_code/mindocr//tools/export.py", line 88, in export
ms.export(net, x, file_name=output_path, file_format="MINDIR")
File "/usr/local/lib/python3.7/site-packages/mindspore/train/serialization.py", line 1575, in export
Validator.check_file_name_by_regular(file_name)
File "/usr/local/lib/python3.7/site-packages/mindspore/_checkparam.py", line 586, in check_file_name_by_regular
raise ValueError(f"{prim_name} '{target}' is illegal, it must be match regular '{reg}'.")
ValueError: The 'db++_r50_icdar15.mindir' is illegal, it must be match regular '^[0-9a-zA-Z@_-.:/\]+$'.

感谢您的反馈。该问题已在PR#664中修复,可通过拉取最新的代码获取。