@cui,xingxing I tried loading the ONNX model in Python ONNXRuntime and sent 2 dummy inputs of size:
1 x 3 x 400 x 300 (N x C x H x W) since the model does not specify the expected height and width of its inputs, and these are the sizes of "viprectification_deskLeft.png" and "viprectification_deskRight.png". Following is my python scirpt:
import onnxruntime as ort
import numpy as np
model_path = 'xfeat.onnx'
session = ort.InferenceSession(model_path)
input1_name = session.get_inputs()[0].name
input2_name = session.get_inputs()[1].name
# Create a dummy input tensor with the same shape as the model's input
# Replace this with your actual input data
dummy_input1 = np.random.rand(1,3,400,300).astype(np.float32)
dummy_input2 = np.random.rand(1,3,400,300).astype(np.float32)
# Run inference
outputs = session.run(None, {input1_name: dummy_input1, input2_name: dummy_input2})
Unfortunately, I run into the following error in ONNX Runtime, upon running the above script:
---------------------------------------------------------------------------Fail Traceback (most recent call last) Cell In[7], line 14 11 dummy_input2 = np.random.rand(1,3,400,300).astype(np.float32) 13 # Run inference---> 14 outputs = session.run(None, {input1_name: dummy_input1, input2_name: dummy_input2}) 16 # Print the output 17 print("Model output:", outputs[0]) File ~\AppData\Roaming\Python\Python311\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 TopK node. Name:'/TopK_3' Status Message: k argument [6400] should not be greater than specified axis dim value [3072]
Please let me know the correct input sizes that work for this model, in ONNX Runtime?
Since you already reached out to us through MathWorks Technical Support, please directly reply to them with the requested information.
