[Torch] improve TensorRT fallback & dynamic #177
Closed
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
tanyokwok
force-pushed
the
features/dev_trt
branch
from
3a0dfef
to
17e5e0d
Compare
qiuxiafei
reviewed
Mar 21, 2022
tanyokwok
force-pushed
the
features/dev_trt
branch
from
26822f8
to
1cabb27
Compare
This was referenced Mar 22, 2022
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Outdated suggestions cannot be applied.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 pull request mainly improves dynamic shape optimization and fallback of TensorRT.
Previously, we only support user configuration of dynamic shapes by setting min/max/opts shapes, which are lists of ints.
And TorchBlade will automatically generate test data according to the input shapes. With the test data, TorchBlade can trace the intermediate tensors' shapes and data types which is valuable to the optimization.
In this PR we would like to support inferencing dynamic shape setting with user input test data. Because some network's computations could be different according to different inputs, such as detection and decoder networks. And a randomly generated test data might output empty results.
In this PR we improve the fallback performance of TensorRT as well.
In this PR we added some debug utilities.