I'm using a TorchScript Model on Pytorch C++ Frontend.
The model in Python returns an output
dict as Dict[str, List[torch.Tensor]]
.
When I use it in C++, it returns a c10::Dict<c10::IValue, c10::IValue>
. What is the equivalent of this Python code:
value_a = output['key_a']
value_b = output['key_b']
in C++ to get value from c10::Dict
?
I have tried this but it does not work.
torch::IValue key_a("key_a");
torch::IValue key_b("key_b");
c10::IValue value_a = output[key_a];
c10::IValue value_b = output[key_b];
std::cout << value_a << std::endl;
std::cout << value_b << std::endl;
And error:
error: type 'c10::Dict<c10::IValue, c10::IValue>' does not provide a subscript operator
You can find header file of c10:Dict
here. What you want is at
method (defined here), so:
auto value_a = output.at(key_a);
Should do the trick.
Also you don't have to create torch::IValue key_ay("key_a")
explicitly, this should be sufficient:
auto value_a = output.at("key_a");