mmr/tkDNN
Archived
1
0
Fork 0
Code Issues 46 Pull Requests 12 Releases 4 Activity

Commit Graph

  • c95c2dbfbc rtinference nstep master Francesco Gatti 2023-08-09 02:28:25 +08:00
  • a503f0010b (tkDNN): Make Resnet-101-AP-GeM work with TRT dev_rtls Micaela Verucchi 2022-09-25 11:30:53 +02:00
  • 56da10df64 (tkDNN): Support Resnet-101-AP-GeM for CUDNN, TRT to be fixed Micaela Verucchi 2022-09-24 16:57:12 +02:00
  • 226875567b Update yolo4-mbzirc-objs with better training mbzirc Micaela Verucchi 2022-09-06 10:23:58 +02:00
  • 59ce9ed23b Add weights automatic download for yolo4-mbzirc-vessel and yolo4-mbzirc-objs Micaela Verucchi 2022-08-30 12:17:38 +02:00
  • 365e7d87a6 Add two new mbzirc network for vessels and objs Micaela Verucchi 2022-08-30 11:31:18 +02:00
  • 40dc3f7c2a Fix eigen version / kernels.h include AdamCuellar/master AdamCuellar/HEAD AdamCuellar/master Adam Cuellar 2022-08-25 13:50:30 -04:00
  • 48943b9fcf compile error fixed from header include properly engcang/master engcang/HEAD engcang/master engcang 2022-08-24 04:55:54 +09:00
  • 300b3fbc52 Add mbzirc net Micaela Verucchi 2022-08-22 16:49:47 +02:00
  • 1019e5f5ef Add installing libyaml-cpp-dev to docker/Dockerfile.base matthewgiarra/dockerfile.base-add-libyaml-cpp-dev matthewgiarra/dockerfile.base-add-libyaml-cpp-dev Matthew Giarra 2022-06-18 23:31:05 -04:00
  • cd58072ae9 Fixes for running master branch on windows perseusdg/master perseusdg/HEAD perseusdg/master Harshvardhan Chandirasekar 2022-04-17 11:01:57 -07:00
  • 7e4b5dbfa6 Merge pull request #13 from ceccocats/master perseusdg/tensorrt8 Harshvardhan Chandirasekar 2022-04-16 15:16:08 -07:00
  • a37c14d71d Fix bad merge Adam Cuellar 2022-04-11 15:57:40 -04:00
  • 31075be60b Merge remote-tracking branch 'origin/master' Adam Cuellar 2022-04-11 15:19:27 -04:00
  • a336126733 Revert "Merge branch 'master' into tensorrt8" perseusdg 2022-04-06 17:40:20 -04:00
  • ac0d6e928e Merge branch 'master' into tensorrt8 Harshvardhan Chandirasekar 2022-04-06 17:06:27 -04:00
  • a3d607a210 Add yolov4-csp-swish AdamCuellar 2021-09-15 08:24:13 -04:00
  • c8bb81b236 Update cmake Cuellar, Adam T 2021-10-04 19:41:35 +00:00
  • 69ff4f37f4 Add swish Cuellar, Adam T 2021-07-27 13:09:43 +00:00
  • d4f7b4ad8b remove using namespace useless v0.7 matthewgiarra/master matthewgiarra/HEAD Francesco Gatti 2022-03-30 22:14:06 +02:00
  • 5e71b99265 YoloRT save bias, mask and clasesName into RT file Francesco Gatti 2022-03-30 20:46:51 +02:00
  • fa9db167b8 version 0.7 Francesco Gatti 2022-03-30 17:48:02 +02:00
  • 30098ca6b4 Merge pull request #285 from ceccocats/tensorrt8 Francesco Gatti 2022-03-30 17:45:41 +02:00
  • df31375b67 Merge branch 'master' into tensorrt8 Francesco Gatti 2022-03-30 16:41:45 +02:00
  • 69bb7370a5 compile with tensorrt7 Francesco Gatti 2022-03-30 15:52:30 +02:00
  • 7c0620e391 test_all_test script save results in separate files matthewgiarra/tensorrt7 tensorrt7 Francesco Gatti 2022-03-30 15:47:41 +02:00
  • c690d537f1 Update READMEs, add README_depth, minors Micaela Verucchi 2022-03-30 15:40:25 +02:00
  • 40266a6c32 Fixed tensorrt8 branch to work jetpack 4.5 and tensorrt7 Harshvardhan Chandirasekar 2022-03-16 18:09:32 +05:30
  • 480b5a9c5a added monodepth2_1024 perseusdg 2022-02-22 14:50:39 +00:00
  • 3bcc32ffdc Fix nms thresh Micaela Verucchi 2022-02-15 19:57:57 +01:00
  • a0f54cde95 Improve Yolo, mobilenet and shelfnet preprocessing using GPU #56 matthewgiarra/gpu-prepocess gpu-prepocess Micaela Verucchi 2022-01-27 10:53:25 +01:00
  • c2825cc570 Add interleavedToPlanar kernels, normalize kernels and utiliy funcs Micaela Verucchi 2022-01-27 10:47:13 +01:00
  • b3bc93693f Revert "Merge branch 'tensorrt8-rds' into depth" perseusdg 2022-01-25 18:30:39 +05:30
  • ef564f134d Merge pull request #6 from perseusdg/depth Harshvardhan Chandirasekar 2022-01-24 21:55:06 +05:30
  • f996659845 Merge remote-tracking branch 'origin/depth' into depth perseusdg 2022-01-24 21:48:35 +05:30
  • afdad8e661 Moved batchnorm and test_monodepth2_new_format layer to dev Harshvardhan Chandirasekar 2022-01-24 21:41:59 +05:30
  • 71366befc2 Merge pull request #5 from perseusdg/depth Harshvardhan Chandirasekar 2022-01-24 21:44:37 +05:30
  • 3e86671c50 Moved batchnorm and test_monodepth2_new_format layer to dev Harshvardhan Chandirasekar 2022-01-24 21:41:59 +05:30
  • 00f06f7bcc Merge branch 'tensorrt8-rds' into depth Harshvardhan Chandirasekar 2022-01-20 22:05:54 +05:30
  • 53725c88a9 Merge pull request #3 from perseusdg/tensorrt8 Harshvardhan Chandirasekar 2022-01-20 21:12:49 +05:30
  • 6a133d8dec -Replaced ISliceLayer based paddings(reflection,constant and zero) with the IPluginV2 version for TensorRT >= 8.2.0 -Fixed demo build issue on Windows Harshvardhan Chandirasekar 2022-01-20 00:52:22 +05:30
  • decd73d298 Fix saving result video in demoDepth, add automatic download for monodepth2 weights Micaela Verucchi 2022-01-19 10:22:06 -08:00
  • 907df27e07 Fix monodepth2, add demoDepth: - Fix monodepth2 network, now works with both cuDNN and tensorRT - Substitute cuDNN ELU with tkDNN one - add DepthNN class - add demoDepth demo, now only works with monodepth2 net Micaela Verucchi 2022-01-18 21:34:16 -08:00
  • 2f80c3c38d Delete codeql-analysis.yml Harshvardhan Chandirasekar 2022-01-11 13:27:55 +05:30
  • 9ba6d84282 Create codeql-analysis.yml Harshvardhan Chandirasekar 2022-01-11 13:26:53 +05:30
  • 19e41a8b99 monodepth2 new format (test) conv + independent batch norm perseusdg 2022-01-11 12:38:36 +05:30
  • f58b6456bc Merge branch 'ceccocats:master' into master Adam Cuellar 2022-01-10 15:07:37 -05:00
  • 061bc79a69 Added BatchNorm to NetworkRT (conver_layer) perseusdg 2022-01-10 22:40:31 +05:30
  • bcf0c4eab3 Added BatchNorm Layer (Testing still needs to be done) perseusdg 2022-01-10 18:43:12 +05:30
  • e1eac2d42a 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 2022-01-09 23:03:26 +05:30
  • 3023926695 fix constantpadding commit perseusdg 2022-01-09 22:17:43 +05:30
  • 4d8f99b441 depth->tensorrt8 patches perseusdg 2022-01-09 19:24:38 +05:30
  • cabebc95a2 added individual layer names in monodepth2 perseusdg 2022-01-09 13:45:44 +05:30
  • 3b58fbcb82 Added ConstantPadding plugin for TensorRT < 8.2 perseusdg 2022-01-07 13:05:09 +05:30
  • f189efcbb1 Reflection Padding native plugin fix ,forgot to added input_dim.c in the plugin creator perseusdg 2022-01-07 05:41:26 +00:00
  • 7298dcfb2f Added monodepth2.cpp perseusdg 2022-01-06 16:30:11 +00:00
  • b75cecb105 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 2022-01-04 09:03:04 +05:30
  • ba022663f1 completed padding migrations from github perseusdg 2022-01-03 19:49:32 +05:30
  • 6837644eb2 add reflection padding from github perseusdg 2022-01-03 18:58:13 +05:30
  • dcf4054bc6 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 yolo4_berkeley - tested with a batchsize of 1 and 2 and on gtx 1070,it is possible that the performance improvement is more signficant in newer hardware perseusdg 2022-01-03 13:52:16 +05:30
  • 04de9908a6 Add yolo4-csp for crowdhuman dataset, add shelfnet for coco-stuff dataset, fix minor in demo.cpp Micaela Verucchi 2021-11-29 17:33:57 +01:00
  • 9cbac460bc Update README.md Micaela Verucchi 2021-11-25 11:32:38 +01:00
  • 24cdb4c4a7 Update Readme Micaela Verucchi 2021-11-24 18:20:26 +01:00
  • a4781244f4 Fix merge problem Micaela Verucchi 2021-11-24 18:16:15 +01:00
  • bbae618118 Update README and add print in demo Micaela Verucchi 2021-11-24 18:08:41 +01:00
  • 0707c26bbd Merge of perseusdg-tensorrt8 inside tkDNN Micaela Verucchi 2021-11-24 17:59:27 +01:00
  • 9cecc5051a -Modified Dockerfile.base to cudagl -Changed demo to take in input from demoConfig.yaml file -Readme changes for demo.md perseusdg 2021-11-24 03:39:57 +05:30
  • be5864748a Use yaml config file for the demo instead of param list Micaela Verucchi 2021-11-23 16:29:40 +01:00
  • 75c3cb0038 Add script to compare times_rtinference.csv files Micaela Verucchi 2021-11-23 13:01:58 +01:00
  • 55df97afe1 Fix warnings, upgrade to C++14 Micaela Verucchi 2021-11-23 13:01:07 +01:00
  • a8c98e3c31 Signed-off-by: perseusdg <43143075+perseusdg@users.noreply.github.com> perseusdg 2021-11-19 22:40:52 +05:30
  • 19f12d9c15 rds_slam -> tensorrt8 port for rds perseusdg 2021-11-15 00:08:37 +05:30
  • 6f936096ae Added num of layers in net and netRT,suggestions from pull request 270 on main repo perseusdg 2021-11-14 07:13:42 +05:30
  • 367061fea2 fps fix perseusdg 2021-11-13 18:11:46 +05:30
  • d488c3bd17 remove duplicated lines perseusdg 2021-11-13 02:12:50 +05:30
  • ee5000ccca bug fixes for dla networks and ported optimization from different pull request perseusdg 2021-11-13 02:00:49 +05:30
  • 9e328c0daa Migrate tkDNN max pooling plugin creation to pluginRegistry from the default method perseusdg 2021-11-10 20:09:57 +05:30
  • 7dd33cd118 added assert for supportsFormat in plugins perseusdg 2021-11-10 14:37:16 +05:30
  • 936b680f2f Remove debug prints perseusdg 2021-11-10 13:45:44 +05:30
  • c0e2097397 - CMakeLists.txt opencv cuda contrib autodetect - Updated Docker to cuda-11.3+cudnn-8.2.1+TensorRT-8.0.34,Ubuntu to 20.04 and OpenCV to 4.5.4 - Updated OpenCV4 to 4.5.4 in install_OpenCV4.sh - Updated README.md perseusdg 2021-11-10 12:30:57 +05:30
  • 31af5396bb Merge branch 'ceccocats:master' into master Adam Cuellar 2021-11-09 16:50:40 -05:00
  • 744396fb0e Update README.md,windows.md and demo.cpp Small fixes in DeformableConvRT.cpp perseusdg 2021-11-09 13:37:26 +05:30
  • d2d44e9e92 Fix max elem (remove thrust) for segmentation Micaela Verucchi 2021-11-04 19:28:55 +01:00
  • 33bbcd216c Fix max elem (remove thrust) for segmentation perseusdg/IPlugin Micaela Verucchi 2021-11-04 19:28:55 +01:00
  • d6fb6c6af4 Fix max elem (remove thrust) for segmentation Micaela Verucchi 2021-11-04 19:28:55 +01:00
  • 802c01bd8f windows debug fix perseusdg 2021-10-30 19:29:21 +05:30
  • 7a89a4a573 ReshapeRT.cpp fix perseusdg 2021-10-28 23:43:12 +05:30
  • c5e66c6bf6 TRT8 works with almost every nerual network now!!!!(including demo3d) perseusdg 2021-10-28 23:35:37 +05:30
  • 8c36dd0431 mobilenet works with trt7(IPluginV2IOExt) ,need to test it with trt8 perseusdg 2021-10-27 18:00:33 +05:30
  • 54e7af11ed tensorrt7 support for ipluginv2 perseusdg 2021-10-18 18:07:57 +05:30
  • 46daa78ca9 Merge branch 'ceccocats:master' into master Adam Cuellar 2021-10-04 15:57:48 -04:00
  • e2975555ef Merge branch 'master' of https://github.com/AdamCuellar/tkDNN Cuellar, Adam T 2021-10-04 19:47:42 +00:00
  • bacabb4b42 Update cmake Cuellar, Adam T 2021-10-04 19:41:35 +00:00
  • eca10ac0a8 Update weights Micaela Verucchi 2021-09-19 00:16:31 +02:00
  • 2e8ffb9d1d Add yolov4-csp-swish AdamCuellar 2021-09-15 08:24:13 -04:00
  • 18cc6abbb6 DeformableConvRT.h fixes ,cmake cuda arch auto detection perseusdg 2021-09-07 23:57:52 +05:30
  • 03473743c4 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 2021-09-07 22:00:28 +05:30
  • 9fa116ce4a 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::312, condition: mDeviceMemorySize == 0 || mExecutionResources->getDeviceMemory() != nullptr - is the error that causes it perseusdg 2021-09-07 07:20:46 +05:30
  • 504dee5016 test_yolo4 netRT->destroy() perseusdg 2021-09-06 15:35:57 +05:30
  • 65ba5c9844 tkDNN can now deserialize tensorrt-8 engine (both through test_* and trtexec) but demo has issues in yolo::computeDetections Harshvardhan Chandirasekar 2021-09-05 01:06:56 +05:30