[TensorRT EP] Add new provider option to exclude ops from running on TRT #23705
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
jywu-msft
reviewed
Feb 15, 2025
onnxruntime/core/providers/tensorrt/tensorrt_execution_provider_helper.cc
Outdated
Show resolved
Hide resolved
|
This PR also modifies NonZero and NMS unit-tests to accommodate TRT's output. |
|
I've tested on Windows with FasterRCNN/MaskRCNN and see no perf regression on latency after enabling DDS. |
jywu-msft
approved these changes
Feb 21, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR removes the implicit filtering-out DDS ops from running on TRT. In other words, by default, DDS nodes will be run by TRT if it supports.
Moreover, it adds new provider option
trt_op_types_to_exclude:trt_op_types_to_exclude="NonMaxSuppression,NonZero,RoiAlignl"(This PR basically adds back featurethat previously being held to merge.)
[Note]
There may be potential performance issues in TRT 10 when running models that contain DDS operations such as NonMaxSuppression, NonZero, and RoiAlign (e.g., Faster-RCNN).
If user encounters significant performance degradation, we suggest specifying those DDS ops to be excluded from running by TRT, i.e. trt_op_types_to_exclude="NonMaxSuppression,NonZero,RoiAlign". Those DDS nodes will be run by CUDA EP or CPU.