can't run paddleocr converted model on onnxruntime

[pxike]

**Please fill in the following information completely so that we can solve the problem quickly, thank you! **

Problem Description paddleOCR converted to onnx model with dynamic input is being run in onnx . first the image needs preprocessing

def preprocess_image(image_path , it = (0, 3, 1, 2) ):
    with open(image_path, 'rb') as file:
        image_data = file.read()
    image_bytes = np.frombuffer(image_data, dtype=np.uint8)
    image = cv2.imdecode(image_bytes, cv2.IMREAD_COLOR) 
    # I would expect your image to be HWC at this point: [H, W, 3]
    image_array = image.astype(np.float32) / 255.0
    image_array = np.expand_dims(image_array, axis=0) # [1, H, W ,C]
    # notice that i reordered the dimensions on transpose ! 
    # N = 1 and C = 3 is moved to the first axis 
    image_array = np.transpose(image_array, it) 
    image_array
    return image_array

Input names: ['x']
Output names: ['softmax_2.tmp_0']
Input shape :['p2o.DynamicDimension.0',
 3,
 'p2o.DynamicDimension.1',
 'p2o.DynamicDimension.2']

When running the model in the runtime ;

{
	"name": "Fail",
	"message": "[ONNXRuntimeError] : 1 : FAIL : Non-zero status code returned while running Concat node. Name:'p2o.Concat.2' Status Message: concat.cc:157 onnxruntime::ConcatBase::PrepareForCompute Non concat axis dimensions must match: Axis 2 has mismatched dimensions of 1 and 31",
	"stack": "---------------------------------------------------------------------------
Fail                                      Traceback (most recent call last)
Cell In[12], line 3
      1 print(x)
      2 image_array = preprocess_image(r\"images\\2.jpg\")
----> 3 outputs = session.run(None, {session.get_inputs()[0].name: image_array})
      4 print(\"success\")

File c:\\Users\\hamza.kharbouch\\Desktop\\pp\\.conda\\lib\\site-packages\\onnxruntime\\capi\\onnxruntime_inference_collection.py:220, in Session.run(self, output_names, input_feed, run_options)
    218     output_names = [output.name for output in self._outputs_meta]
    219 try:
--> 220     return self._sess.run(output_names, input_feed, run_options)
    221 except C.EPFail as err:
    222     if self._enable_fallback:

Fail: [ONNXRuntimeError] : 1 : FAIL : Non-zero status code returned while running Concat node. Name:'p2o.Concat.2' Status Message: concat.cc:157 onnxruntime::ConcatBase::PrepareForCompute Non concat axis dimensions must match: Axis 2 has mismatched dimensions of 1 and 31"
}

More information :

• Inference engine for deployment: None
• Why you need to convert to ONNX format: CPU memory leaks in normal runtime for paddleOCR and fastDeploy https://github.com/PaddlePaddle/PaddleOCR/issues/11639
• Paddle2ONNX version: 1.2.1

Error report screenshot

other information

[Zheng-Bicheng]

Sorry for the late response. Could you provide me with the model? I will troubleshoot the issue.

[Zheng-Bicheng]

@pxike

[pxike]

Hello @Zheng-Bicheng , it's this model https://drive.google.com/file/d/1tCa3qzHzHEGq_KqzpAdCt5VLH-4Ym58k/view?usp=sharing which is converted from https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_rec_infer.tar