in the TensorRT engine Establish approach, some complex layer fusions can't be instantly identified. TensorRT-LLM optimizes these making use of plugins which are explicitly inserted to the community graph definition at https://tripsbookmarks.com/story17791047/a-secret-weapon-for-python-class-in-btm