Batch Inputs for Scripted Model in C++

[kg512]

Hi,
I have used https://github.com/facebookresearch/detectron2/tree/main/tools/deploy to generate a scripted faster rcnn and I am able to run it in C++ for a batch size 1.

The detectron2 documentation mentions that dynamic batch inputs are supported for scripted models. However, I am struggling with changing https://github.com/facebookresearch/detectron2/blob/main/tools/deploy/torchscript_mask_rcnn.cpp to support batch inputs for the scripted model. I tried creating a stack with multiple inputs and sending it to model forward but it led to an error.

    c10::Stack stack_batch{
        std::make_tuple(dic),
        std::make_tuple(dic)};

    auto output = model.forward(stack_batch);
    
    Error: Expected at most 2 argument(s) for operator 'forward', but received 3 argument(s). Declaration: 
    forward(__torch__.ScriptableAdapter self, (Dict(str, Tensor)) inputs) -> (Dict(str, Tensor)[][])

Any suggestions?