Code

Tensorrt8#285

Code

Merged

ceccocats merged 71 commits intomasterceccocats/tkDNN:masterfrom tensorrt8ceccocats/tkDNN:tensorrt8Copy head branch name to clipboardMar 30, 2022

+7,419-1,700

Commits

Commits on Apr 25, 2021

• Merge pull request #1 from ceccocats/master

  Show description for 9e3977d
  perseusdg

  authored
  9e3977d

  Copy full SHA for 9e3977d

Commits on Jun 5, 2021

• Merge branch 'ceccocats:master' into master

  perseusdg

  committed
  23fbcd1

  Copy full SHA for 23fbcd1

• Centernet fix for windows

  

  perseusdg

  committed
  710cb54

  Copy full SHA for 710cb54

Commits on Jun 18, 2021

• Merge branch 'ceccocats:master' into master

  perseusdg

  authored
  7f9b10c

  Copy full SHA for 7f9b10c

Commits on Jun 28, 2021

• ActivationLeaky IpluginV2

  mive93

  committed
  b86a93e

  Copy full SHA for b86a93e

Commits on Jul 20, 2021

• Merge branch 'ceccocats:master' into master

  perseusdg

  authored
  4ed247d

  Copy full SHA for 4ed247d

Commits on Jul 21, 2021

• Merge branch 'ceccocats:master' into master

  perseusdg

  authored
  548e87f

  Copy full SHA for 548e87f

Commits on Jul 24, 2021

• Merge branch 'ceccocats:master' into master

  perseusdg

  authored
  e58ddad

  Copy full SHA for e58ddad

• minor fixes for windows ,for release v0.6

  

  perseusdg

  committed
  1412aa6

  Copy full SHA for 1412aa6

Commits on Jul 25, 2021

• fix for downloading weights(utils.cpp) when using docker

  

  Harshvardhan Chandirasekar

  committed
  bb902f5

  Copy full SHA for bb902f5

Commits on Aug 4, 2021

• fix return type for bool functions

  

  perseusdg

  committed
  d2ce966

  Copy full SHA for d2ce966

Commits on Aug 16, 2021

• Updates to build libkernel.so under TensorRT 8

  

  perseusdg

  committed
  eba78e7

  Copy full SHA for eba78e7

Commits on Aug 21, 2021

• Yolo3Detection.cpp doesn't build yet,issues with dependecies of preprocessing and postprocessing on IPluginFactory

  

  perseusdg

  committed
  3d8b1ac

  Copy full SHA for 3d8b1ac

Commits on Aug 28, 2021

• Mnist works at the moment with trt8,others like yolo4tiny and mobilenet generate the engine files but crash after throwing nvifer1::CudaRuntimeError and when demo is being run ,it doesnt deserializ…

  Show description for 2ffe070
  

  perseusdg

  committed
  2ffe070

  Copy full SHA for 2ffe070

Commits on Aug 30, 2021

• update tensorrt8 branch

  

  perseusdg

  committed
  de83ae5

  Copy full SHA for de83ae5

Commits on Sep 4, 2021

• tkDNN can now deserialize tensorrt-8 engine (both through test_* and trtexec)

  Show description for 65ba5c9
  

  Harshvardhan Chandirasekar

  committed
  65ba5c9

  Copy full SHA for 65ba5c9

Commits on Sep 6, 2021

• test_yolo4 netRT->destroy()

  

  perseusdg

  committed
  504dee5

  Copy full SHA for 504dee5

Commits on Sep 7, 2021

• tkdnn(trt8) runs now minus the detections ,[executionContext.cpp::enqueueInternal::312] Error Code 3: Internal Error (Parameter check failed at: runtime/api/executionContext.cpp::enqueueInternal::3…

  Show description for 9fa116c
  

  perseusdg

  committed
  9fa116c

  Copy full SHA for 9fa116c

• tkDNN works with trt8!!,need to test int8 and mobilenet,dla_cnet (fps seems to be a bit low 350 on trt8 compared to 396 on trt7)

  

  perseusdg

  committed
  0347374

  Copy full SHA for 0347374

• DeformableConvRT.h fixes ,cmake cuda arch auto detection

  

  perseusdg

  committed
  18cc6ab

  Copy full SHA for 18cc6ab

Commits on Oct 18, 2021

• tensorrt7 support for ipluginv2

  

  perseusdg

  committed
  54e7af1

  Copy full SHA for 54e7af1

Commits on Oct 27, 2021

• mobilenet works with trt7(IPluginV2IOExt) ,need to test it with trt8

  perseusdg

  committed
  8c36dd0

  Copy full SHA for 8c36dd0

Commits on Oct 28, 2021

• TRT8 works with almost every nerual network now!!!!(including demo3d)

  perseusdg

  committed
  c5e66c6

  Copy full SHA for c5e66c6

• ReshapeRT.cpp fix

  perseusdg

  committed
  7a89a4a

  Copy full SHA for 7a89a4a

Commits on Oct 30, 2021

• windows debug fix

  perseusdg

  committed
  802c01b

  Copy full SHA for 802c01b

Commits on Nov 5, 2021

• Fix max elem (remove thrust) for segmentation

  Show description for d2d44e9
  

  mive93

  authored and

  Harshvardhan Chandirasekar

  committed
  d2d44e9

  Copy full SHA for d2d44e9

Commits on Nov 9, 2021

• Update README.md,windows.md and demo.cpp

  Show description for 744396f
  perseusdg

  committed
  744396f

  Copy full SHA for 744396f

Commits on Nov 10, 2021

• - CMakeLists.txt opencv cuda contrib autodetect

  Show description for c0e2097
  perseusdg

  committed
  c0e2097

  Copy full SHA for c0e2097

• Remove debug prints

  perseusdg

  committed
  936b680

  Copy full SHA for 936b680

• added assert for supportsFormat in plugins

  perseusdg

  committed
  7dd33cd

  Copy full SHA for 7dd33cd

• Migrate tkDNN max pooling plugin creation to pluginRegistry from the default method

  perseusdg

  committed
  9e328c0

  Copy full SHA for 9e328c0

Commits on Nov 12, 2021

• bug fixes for dla networks and ported optimization from different pull request

  perseusdg

  committed
  ee5000c

  Copy full SHA for ee5000c

• remove duplicated lines

  perseusdg

  committed
  d488c3b

  Copy full SHA for d488c3b

Commits on Nov 13, 2021

• fps fix

  perseusdg

  committed
  367061f

  Copy full SHA for 367061f

Commits on Nov 14, 2021

• Added num of layers in net and netRT,suggestions from pull request 270 on main repo

  perseusdg

  committed
  6f93609

  Copy full SHA for 6f93609

• rds_slam -> tensorrt8 port for rds

  perseusdg

  committed
  19f12d9

  Copy full SHA for 19f12d9

Commits on Nov 19, 2021

• Signed-off-by: perseusdg <43143075+perseusdg@users.noreply.github.com>

  Show description for a8c98e3
  perseusdg

  committed
  a8c98e3

  Copy full SHA for a8c98e3

Commits on Nov 23, 2021

• -Modified Dockerfile.base to cudagl

  Show description for 9cecc50
  perseusdg

  committed
  9cecc50

  Copy full SHA for 9cecc50

Commits on Nov 24, 2021

• Merge of perseusdg-tensorrt8 inside tkDNN

  Show description for 0707c26
  mive93

  committed
  0707c26

  Copy full SHA for 0707c26

• Update README and add print in demo

  Show description for bbae618
  mive93

  committed
  bbae618

  Copy full SHA for bbae618

• Fix merge problem

  Show description for a478124
  mive93

  committed
  a478124

  Copy full SHA for a478124

Commits on Jan 3, 2022

• performance improvements for yolo based networks,significant reduction in inference time can be seen yolo4tiny ,yolo4 and minor reduction in inference time can be seen in yolo4_berkeley_f1 and yolo…

  Show description for dcf4054
  perseusdg

  committed
  dcf4054

  Copy full SHA for dcf4054

• add reflection padding from github

  perseusdg

  committed
  6837644

  Copy full SHA for 6837644

• completed padding migrations from github

  perseusdg

  committed
  ba02266

  Copy full SHA for ba02266

Commits on Jan 4, 2022

• Added constant and zero padding with ISliceLayer,need to add them for tensorrt versions less than 8.2 using IPluginV2Ext instead of ISliceLayer since they dont seem to support reflect and zero

  perseusdg

  committed
  b75cecb

  Copy full SHA for b75cecb

Commits on Jan 6, 2022

• Added monodepth2.cpp

  perseusdg

  committed
  7298dcf

  Copy full SHA for 7298dcf

Commits on Jan 7, 2022

• Reflection Padding native plugin fix ,forgot to added input_dim.c in the plugin creator

  perseusdg

  committed
  f189efc

  Copy full SHA for f189efc

• Added ConstantPadding plugin for TensorRT < 8.2

  perseusdg

  committed
  3b58fbc

  Copy full SHA for 3b58fbc

Commits on Jan 9, 2022

• added individual layer names in monodepth2

  perseusdg

  committed
  cabebc9

  Copy full SHA for cabebc9

• depth->tensorrt8 patches

  perseusdg

  committed
  4d8f99b

  Copy full SHA for 4d8f99b

• fix constantpadding commit

  perseusdg

  committed
  3023926

  Copy full SHA for 3023926

Commits on Jan 10, 2022

• Added a seperate layer file for batchnorm,to support an independent batchnorm class in order to pass necessary parameters read from the bin file.

  perseusdg

  committed
  e1eac2d

  Copy full SHA for e1eac2d

• Added BatchNorm Layer (Testing still needs to be done)

  perseusdg

  committed
  bcf0c4e

  Copy full SHA for bcf0c4e

• Added BatchNorm to NetworkRT (conver_layer)

  perseusdg

  committed
  061bc79

  Copy full SHA for 061bc79

Commits on Jan 11, 2022

• monodepth2 new format (test) conv + independent batch norm

  perseusdg

  committed
  19e41a8

  Copy full SHA for 19e41a8

Commits on Jan 19, 2022

• Fix monodepth2, add demoDepth:

  Show description for 907df27
  mive93

  committed
  907df27

  Copy full SHA for 907df27

• Fix saving result video in demoDepth, add automatic download for monodepth2 weights

  Show description for decd73d
  mive93

  committed
  decd73d

  Copy full SHA for decd73d

• -Replaced ISliceLayer based paddings(reflection,constant and zero) with the IPluginV2 version for TensorRT >= 8.2.0

  Show description for 6a133d8
  perseusdg

  committed
  6a133d8

  Copy full SHA for 6a133d8

Commits on Jan 20, 2022

• Merge pull request #3 from perseusdg/tensorrt8

  Show description for 53725c8
  perseusdg

  authored
  53725c8

  Copy full SHA for 53725c8

• Merge branch 'tensorrt8-rds' into depth

  perseusdg

  authored
  00f06f7

  Copy full SHA for 00f06f7

Commits on Jan 24, 2022

• Moved batchnorm and test_monodepth2_new_format layer to dev

  Show description for 3e86671
  perseusdg

  committed
  3e86671

  Copy full SHA for 3e86671

• Merge pull request #5 from perseusdg/depth

  Show description for 71366be
  perseusdg

  authored
  71366be

  Copy full SHA for 71366be

• Moved batchnorm and test_monodepth2_new_format layer to dev

  Show description for afdad8e
  perseusdg

  committed
  afdad8e

  Copy full SHA for afdad8e

• Merge remote-tracking branch 'origin/depth' into depth

  perseusdg

  committed
  f996659

  Copy full SHA for f996659

• Merge pull request #6 from perseusdg/depth

  Show description for ef564f1
  perseusdg

  authored
  ef564f1

  Copy full SHA for ef564f1

Commits on Jan 25, 2022

• Revert "Merge branch 'tensorrt8-rds' into depth"

  Show description for b3bc936
  perseusdg

  committed
  b3bc936

  Copy full SHA for b3bc936

Commits on Feb 22, 2022

• added monodepth2_1024

  

  perseusdg

  committed
  480b5a9

  Copy full SHA for 480b5a9

Commits on Mar 16, 2022

• Fixed tensorrt8 branch to work jetpack 4.5 and tensorrt7

  Show description for 40266a6
  

  perseusdg

  authored and

  perseusdg

  committed
  40266a6

  Copy full SHA for 40266a6

Commits on Mar 30, 2022

• Update READMEs, add README_depth, minors

  Show description for c690d53
  mive93

  committed
  c690d53

  Copy full SHA for c690d53

• compile with tensorrt7

  ceccocats

  committed
  69bb737

  Copy full SHA for 69bb737

• Merge branch 'master' into tensorrt8

  ceccocats

  committed
  df31375

  Copy full SHA for df31375