check.py script not working

[prathik-naidu]

Hi,

I installed torch2trt with the plugins, running python setup.py install --plugins which seemed to work fine. However, I get the error,

AttributeError: 'Tensor' object has no attribute '_trt'

which is happening at the following line:

mark_outputs(self, torch_outputs, names)
    305
    306         for i, torch_output in enumerate(torch_outputs):
--> 307             trt_tensor = torch_output._trt
    308             trt_tensor.name = names[i]
    309             trt_tensor.location = torch_device_to_trt(torch_output.device)

Seems like this is an issue that's been discussed before (https://github.com/NVIDIA-AI-IOT/torch2trt/issues/15) due to convertors not being implemented. Is this something you came across/any idea why this error is happening on the check script with this Upsample net?

Also I'm not using a Jetson device (trying to run this on an RTX 2080 Ti)

Thanks!