I want to do inference using openvino.
But I got an error while using openvino.
Any way to solve it?enter code here
model = keras.models.load_model('/resnet50.h5')
onnx_model, _ = tf2onnx.convert.from_keras(model, opset=16)
onnx.save(onnx_model, '/t1_model.onnx')
ie = IECore()
net = ie.read_network("/t1_model.onnx")
input_name = list(net.input_info.keys())[0]
output_name = list(net.outputs.keys())[0]
net.input_info[input_name].precision = 'FP32'
net.outputs[output_name].precision = 'FP32'
exec_net = ie.load_network(network=net, device_name='CPU')
I faced these problems.
RuntimeError: Check 'std::get<0>(valid)' failed at C:\j\workspace\private-ci\ie\build-windows-vs2019@3\b\repos\openvino\src\inference\src\ie_core.cpp:1414:
InferenceEngine::Core::LoadNetwork doesn't support inputs having dynamic shapes. Use ov::Core::compile_model API instead. Dynamic inputs are :{ input:'input_1,input_1', shape={?,256,256,3}}
input_shape = (None, 256,256,3)
The IECore API doesn't support dynamic shapes so you need to make your model static before you load it into the plugin. You can use the reshape() method on the imported model.
As an alternative you can switch to the 2022.1 version of OV where the dynamic shapes are supported. You have to switch from IECore to Core, read_network -> read_model, load_network -> compile_model.