ValueError while using ONNX DB Net to infer

[thsno02]

请提供下述完整信息以便快速定位问题/Please provide the following information to quickly locate the problem

• 系统环境/System Environment： CentOS 7
• 版本号/Version：Paddle：2.3.1-gpu PaddleOCR：2.6.0 问题相关组件/Related components：onnx, 1.13.0; onnxruntime-gpu, 1.13.1
• 运行指令/Command Code：python3 tools/infer_predict.py --use_onnx=True
• 完整报错/Complete Error Message：ValueError: This ORT build has ['TensorrtExecutionProvider', 'CUDAExecutionProvider', 'CPUExecutionProvider'] enabled. Since ORT 1.9, you are required to explicitly set the providers parameter when instantiating InferenceSession. For example, onnxruntime.InferenceSession(..., providers=['TensorrtExecutionProvider', 'CUDAExecutionProvider', 'CPUExecutionProvider'], ...)

[thsno02]

Workaround:

replace

PaddleOCR/tools/infer/utility.py

Line 184 in 23e034c

sess = ort.InferenceSession(model_file_path)

with

sess = ort.InferenceSession(model_file_path, providers=['CUDAExecutionProvider', 'CPUExecutionProvider'])

REF: neuralchen/SimSwap#176