Missing kDEFAULT flag

I am upgrading from TRT5 to TRT6 and reading through the documentation I see that the createNetwork() method is deprecated and createNetworkV2()
should be used. Reading the documentation it says Providing the kDEFAULT flag as parameter mimics the behaviour of createNetwork().
but inside the NetworkDefinitionCreationFlag I cannot find such flag kDEFAULT.

Before using the new API my code compiles and parses my UFF file correctly, however with the new code it breaks with an error like this

all concat input tensors must have the same dimensions except on the concatenation axis (0), but dimensions mismatched at input 1 at index 1. Input 0 shape: [1,96,46,82], Input 1 shape: [1,384,46,82]

Can you help me figuring out where this kDEFAULT is or what value to pass to createNetworkV2()
to have the same behavior as createNetwork()?

Hi,

You can check this sample for the usage of createNetworkV2():
/usr/src/tensorrt/samples/sampleDynamicReshape/

Based on the document here:
https://docs.nvidia.com/deeplearning/sdk/tensorrt-api/c_api/classnvinfer1_1_1_i_builder.html#a6ad627fece8e6f1aa3b4fea6471068ea

You can declare createNetworkV2() like following to achieve the similar behavior of createNetwork():

INetworkDefinition* network = builder->createNetworkV2(0U);

Thanks.

Thank you @AastaLLL I missed that.

Nonetheless, I strongly believe that if the description for createNetworkV2() mentions kDEFAULT it should exists as part of the enum NetworkDefinitionCreationFlag