Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

TensorRT OSS 22.04 release #1923

Merged
merged 25 commits into from
Apr 14, 2022
Merged

TensorRT OSS 22.04 release #1923

merged 25 commits into from
Apr 14, 2022

Conversation

rajeevsrao
Copy link
Collaborator

@rajeevsrao rajeevsrao commented Apr 13, 2022

Commit used by the 22.04 TensorRT NGC container.

Changelog

Added

  • TensorRT Engine Explorer v0.1.0 README
  • Detectron 2 Mask R-CNN R50-FPN python sample
  • Model export script for sampleOnnxMnistCoordConvAC

Changed

  • Updated base TensorRT version to 8.2.4.2
  • Updated copyright headers with SPDX identifiers
  • Updated onnx-graphsurgeon v0.3.17 CHANGELOG
  • PyramidROIAlign plugin refactor and bug fixes
  • Fixed MultilevelCropAndResize crashes on Windows
  • #1583 - sublicense ieee/half.h under Apache2
  • Updated demo/BERT performance tables for rel-8.2
  • #1774 Fix python hangs at IndexErrors when TF is imported after TensorRT
  • Various bugfixes in demos - BERT, Tacotron2 and HuggingFace GPT/T5 notebooks
  • Cleaned up sample READMEs

Removed

  • sampleNMT removed from samples

@lgtm-com
Copy link

lgtm-com bot commented Apr 13, 2022

This pull request introduces 66 alerts and fixes 14 when merging 80ec139695f3222d12dcffdad0452512b603b432 into 46253b6 - view on LGTM.com

new alerts:

  • 19 for 'import *' may pollute namespace
  • 14 for Unused import
  • 10 for Wrong name for an argument in a class instantiation
  • 6 for Except block handles 'BaseException'
  • 5 for Unused local variable
  • 3 for `__init__` method calls overridden method
  • 2 for Wrong number of arguments in a call
  • 1 for First parameter of a method is not named 'self'
  • 1 for Use of the return value of a procedure
  • 1 for Constant in conditional expression or statement
  • 1 for Unreachable code
  • 1 for Module is imported with 'import' and 'import from'
  • 1 for Variable defined multiple times
  • 1 for Wrong number of arguments in a class instantiation

fixed alerts:

  • 8 for Module is imported with 'import' and 'import from'
  • 1 for Testing equality to None
  • 1 for Unused local variable
  • 1 for Unused import
  • 1 for Unreachable code
  • 1 for Nested loops with same variable
  • 1 for Wrong number of arguments in a call

@lgtm-com
Copy link

lgtm-com bot commented Apr 13, 2022

This pull request introduces 66 alerts and fixes 14 when merging 684aab4ad057ce01588c290844e713521e7f00ec into 46253b6 - view on LGTM.com

new alerts:

  • 19 for 'import *' may pollute namespace
  • 14 for Unused import
  • 10 for Wrong name for an argument in a class instantiation
  • 6 for Except block handles 'BaseException'
  • 5 for Unused local variable
  • 3 for `__init__` method calls overridden method
  • 2 for Wrong number of arguments in a call
  • 1 for First parameter of a method is not named 'self'
  • 1 for Use of the return value of a procedure
  • 1 for Constant in conditional expression or statement
  • 1 for Unreachable code
  • 1 for Module is imported with 'import' and 'import from'
  • 1 for Variable defined multiple times
  • 1 for Wrong number of arguments in a class instantiation

fixed alerts:

  • 8 for Module is imported with 'import' and 'import from'
  • 1 for Testing equality to None
  • 1 for Unused local variable
  • 1 for Unused import
  • 1 for Unreachable code
  • 1 for Nested loops with same variable
  • 1 for Wrong number of arguments in a call

CHANGELOG.md Outdated Show resolved Hide resolved
ttyio and others added 24 commits April 13, 2022 16:52
Signed-off-by: Vincent Huang <vincenth@nvidia.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
- Add Windows Load Library Support
- Add --verbose flag to demoBERT scripts for perf testing
- Fix demoBERT INT8 failure

Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
- Enable timing cache for tacotron2 test
- Update convert_onnx2trt.py

Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
Apache2 is a more restrictive license than MIT, and requires that the
attributions to original author to be preserved.

Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
The MultilevelCropAndResize plugin was causing a crash only on Windows.
On further investigation, it was revealed that the crash was due to the
user's model only providing 4 FP inputs to the plugin instead of the
required 5 FPs as specified by the plugin. The reason for the crash to
only occur in Windows and not in Linux is still not well understood.

Although there is an assertion to check whether the correct number of
inputs are provided, it is a C-assertion, which does not get hit
in release builds. It was verified that the assertion is indeed hit in
the debug build. As such, this issue is expected to be preventable after the
completion of TRT-15919 (https://jirasw.nvidia.com/browse/TRT-15919), when
PLUGIN_ASSERT will replace C-assertions.

As a tentative resolution, the README of the MultilevelCropAndResize plugin
has been updated to highlight that the plugin expects 5 FPs. Missing
documentation for the `image_size` attribute has also been added.

Signed-off-by: skarunaratne <skarunaratne@nvidia.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
Unpacking an trt.tensorrt.Dims object causes Python to hang when
Tensorflow is imported after TensorRT, only on Windows.

Signed-off-by: samurdhi karunaratne <skarunaratne@nvidia.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
Co-authored-by: Samurdhi Karunaratne <97725867+samurdhikaru@users.noreply.github.com>
Signed-off-by: Rajeev Rao <rajeevrao@nvidia.com>
@lgtm-com
Copy link

lgtm-com bot commented Apr 14, 2022

This pull request introduces 66 alerts and fixes 14 when merging 413cfc3 into 46253b6 - view on LGTM.com

new alerts:

  • 19 for 'import *' may pollute namespace
  • 14 for Unused import
  • 10 for Wrong name for an argument in a class instantiation
  • 6 for Except block handles 'BaseException'
  • 5 for Unused local variable
  • 3 for `__init__` method calls overridden method
  • 2 for Wrong number of arguments in a call
  • 1 for First parameter of a method is not named 'self'
  • 1 for Use of the return value of a procedure
  • 1 for Constant in conditional expression or statement
  • 1 for Unreachable code
  • 1 for Module is imported with 'import' and 'import from'
  • 1 for Variable defined multiple times
  • 1 for Wrong number of arguments in a class instantiation

fixed alerts:

  • 8 for Module is imported with 'import' and 'import from'
  • 1 for Testing equality to None
  • 1 for Unused local variable
  • 1 for Unused import
  • 1 for Unreachable code
  • 1 for Nested loops with same variable
  • 1 for Wrong number of arguments in a call

@rajeevsrao rajeevsrao merged commit f4a8635 into NVIDIA:main Apr 14, 2022
@rajeevsrao rajeevsrao deleted the dev/main branch August 23, 2023 22:26
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

9 participants