TensorRT Python: Will subtensor/slicing operation be supported in the future?

Hi,

subtensor/slicing is quite common in some new models. For example in numpy we can do like this:

a = b[2:5]

Currently I do not know how to use it in TensorRT. Will it be supported in the future?

Thank you!

Hi,

TensorRT doesn’t have slice layer support currently.
We are checking the possibility but can’t disclosure more information here.

Thanks.