All my operators are GenericOperator (Windows)

Following the tutorial, I load my model using model = aidge_onnx.load_onnx("MNIST_model/MLP_MNIST.onnx") looking at the print it seems that every operator is loaded as a GenericOperator.

Available backends :
{'cpu'}
- _Flatten_output_0 (Flatten | GenericOperator)
	- axis : 1
- _fc1_Gemm_output_0 (Gemm | GenericOperator)
	- alpha : 1.0
	- transB : 1
	- beta : 1.0
- _Relu_output_0 (Relu | GenericOperator)
- _fc2_Gemm_output_0 (Gemm | GenericOperator)
	- alpha : 1.0
	- transB : 1
	- beta : 1.0
- _Relu_1_output_0 (Relu | GenericOperator)
- output (Gemm | GenericOperator)
	- alpha : 1.0
	- transB : 1
	- beta : 1.0

This leads to a crash during prediction as it cannot do shape inference.

Thanks in advance. :)