Open
1 of 3 issues completedDescription
TL;DR
This story tracks all the known perf bugs and fixes for them in Torch-TensorRT compared to ONNX-TRT and TS-TRT
Goal(s)
Close the gap between ONNX-TensorRT and Torch-TensorRT. Identify the existing performance culprit during engine compilation and record them here.
Tasks
### Tasks