Selur's Little Message Board

Hello Selur,

   I'm trying compiling a filter with torch_tensorrt but I found a compatibility problem regarding torch_tensorrt and TensorRT 10.0.
   Looking to NVIDIA migration guide

   The TensorRT 8.x function build_engine() is substitute by the function build_serialized_network()
   Also the TensorRT 8.x parameters INetworkDefinition.has_explicit_precision and  INetworkDefinition.has_implicit_batch_dimension() were removed.

   But all these parameters are used in the torch_tensorrt 2.4.0.dev installed in Hybrid.
   See for example:

  torch_tensorrt\fx\fx2trt.py, rows: 211, 213, 257

  This version is targeting TensorRT 8.x and not the installed version TensorRT 10.0

  I'm wondering if the current Hybrid environment need to be changed to get  torch_tensorrt working.

Dan

I managed to don't use tensorrt.fx and now I'm able to compile (till now only fp32).

Dan