Commit Graph

1176 Commits

Author SHA1 Message Date
b6e15f1656 ZED 3D Camera support added to ./uselib (yolo_console_cpp.exe) example 2019-03-18 02:48:52 +03:00
7a854302ef Merge pull request #2542 from kossolax/patch-2
Segfault if height > width
2019-03-13 02:41:42 +03:00
dd27d67f58 Merge pull request #2536 from cenit/dev/cenit/cmakeConfig
improve darknet library handling for downstream projects, other small updates
2019-03-12 01:19:02 +03:00
4fc5c6e7b0 improve compatibility with latest pthreadVC renaming on vcpkg 2019-03-08 21:48:23 +01:00
17fed7c4e2 Segfault if height > width
./darknet classifier train data/sort.data cfg/alexnet.cfg -dont_show
alexnet
1
 compute_capability = 750, cudnn_half = 1 
layer     filters    size              input                output
   0 conv     96 11 x11 / 4   128 x 256 x   3   ->    30 x  62 x  96 0.130 BF
   1 max          3 x 3 / 2    30 x  62 x  96   ->    14 x  30 x  96 0.000 BF
   2 conv    256  5 x 5 / 1    14 x  30 x  96   ->    14 x  30 x 256 0.516 BF
   3 max          3 x 3 / 2    14 x  30 x 256   ->     6 x  14 x 256 0.000 BF
   4 conv    384  3 x 3 / 1     6 x  14 x 256   ->     6 x  14 x 384 0.149 BF
   5 conv    384  3 x 3 / 1     6 x  14 x 384   ->     6 x  14 x 384 0.223 BF
   6 conv    256  3 x 3 / 1     6 x  14 x 384   ->     6 x  14 x 256 0.149 BF
   7 max          3 x 3 / 2     6 x  14 x 256   ->     2 x   6 x 256 0.000 BF
   8 connected                            3072  ->  4096
   9 dropout       p = 0.50               4096  ->  4096
  10 connected                            4096  ->  4096
  11 dropout       p = 0.50               4096  ->  4096
  12 connected                            4096  ->  1261
  13 softmax                                        1261
  14 cost                                           1261
Total BFLOPS 1.167 
 Allocate additional workspace_size = 1940.23 MB 
Learning Rate: 0.01, Momentum: 0.9, Decay: 0.0005
893253
Loaded: 0.205056 seconds
Segmentation fault (core dumped)




 This PR fix this issue:
./darknet classifier train data/sort.data cfg/alexnet.cfg -dont_show
alexnet
1
 compute_capability = 750, cudnn_half = 1 
layer     filters    size              input                output
   0 conv     96 11 x11 / 4   128 x 256 x   3   ->    30 x  62 x  96 0.130 BF
   1 max          3 x 3 / 2    30 x  62 x  96   ->    14 x  30 x  96 0.000 BF
   2 conv    256  5 x 5 / 1    14 x  30 x  96   ->    14 x  30 x 256 0.516 BF
   3 max          3 x 3 / 2    14 x  30 x 256   ->     6 x  14 x 256 0.000 BF
   4 conv    384  3 x 3 / 1     6 x  14 x 256   ->     6 x  14 x 384 0.149 BF
   5 conv    384  3 x 3 / 1     6 x  14 x 384   ->     6 x  14 x 384 0.223 BF
   6 conv    256  3 x 3 / 1     6 x  14 x 384   ->     6 x  14 x 256 0.149 BF
   7 max          3 x 3 / 2     6 x  14 x 256   ->     2 x   6 x 256 0.000 BF
   8 connected                            3072  ->  4096
   9 dropout       p = 0.50               4096  ->  4096
  10 connected                            4096  ->  4096
  11 dropout       p = 0.50               4096  ->  4096
  12 connected                            4096  ->  1261
  13 softmax                                        1261
  14 cost                                           1261
Total BFLOPS 1.167 
 Allocate additional workspace_size = 1940.23 MB 
Learning Rate: 0.01, Momentum: 0.9, Decay: 0.0005
893253
Loaded: 0.509795 seconds
1, 0.000: 4.070450, 4.070450 avg, 0.010000 rate, 0.310202 seconds, 128 images
Loaded: 0.416345 seconds
2, 0.000: 4.069521, 4.070357 avg, 0.010000 rate, 0.055622 seconds, 256 images
Loaded: 0.451087 seconds



I'll tell you later if training is converging.
2019-03-06 10:19:13 +01:00
8804479dbd [appveyor] explode matrix to workaround time bomb 2019-03-05 18:35:07 +01:00
2c8ff8db0d create cmake targets for proper lib export 2019-03-05 18:34:53 +01:00
b751bac175 Fixed mAP calculation for CRNN layer during training 2019-03-05 02:15:44 +03:00
ef803e05df Merge pull request #2514 from cenit/dev/cenit/appveyor
fix cuda build on appveyor, remove old unsupported config on macOS
2019-03-04 17:04:41 +03:00
fe9eaa9872 improve build script based on user feedback 2019-03-04 10:22:13 +01:00
540beb52ec get & label OTB dataset 2019-03-04 04:05:17 +03:00
8fd3e7ad25 switch debug/release build order, better futureproof for next cmakeTargets work 2019-03-03 23:35:38 +01:00
40cdaaaa43 add artifacts on appveyor 2019-03-03 19:57:40 +01:00
80c9fa76a4 refresh build scripts 2019-03-03 16:55:15 +01:00
d5cb0b490c scripts/README.md 2019-03-03 15:09:41 +03:00
80d0582ee4 fix cuda build on appveyor, remove old unsupported config on macOS 2019-03-03 13:07:49 +01:00
42b8c0751e Update Readme.md 2019-03-03 13:18:32 +03:00
dd0fdfb684 Compile fix 2019-03-02 03:37:55 +03:00
75f2a3e7cf Added object Detection & Tracking using conv-rnn layer on frames from video 2019-03-02 03:32:24 +03:00
50956447f8 Fixed issue with killed mjpeg_stream 2019-03-01 19:01:14 +03:00
ec1b66ace2 Merge branch 'master' of github.com:AlexeyAB/darknet 2019-03-01 16:56:21 +03:00
e0596a1046 Added cmd/sh scripts for downloading ILSVRC2012 train & valid for Windows 2019-03-01 16:56:00 +03:00
aa24da9fba Merge pull request #2476 from cenit/dev/cenit/CI
[CI] add scripts for travis and appveyor
2019-03-01 16:19:18 +03:00
cfb80e39cd Fixed http_stream.cpp 2019-03-01 15:49:23 +03:00
2c0d857110 [appveyor-cuda] add another env variable 2019-03-01 12:03:41 +01:00
6c7540420b [opencv-cuda] handle properly cuda enabled but opencv built without cuda support 2019-03-01 11:23:31 +01:00
278df5e89e [appveyor] disable building opencv[cuda], it's too slow 2019-03-01 11:11:25 +01:00
6dbfeee159 [appveyor] fix CUDA paths again 2019-03-01 09:27:29 +01:00
52312f5f8f [appveyor/cuda] extract all files from CUDA archive 2019-02-28 21:04:58 +01:00
9b09abe122 Fixed convolutional-layer when it is used as base for crnn-layer 2019-02-28 20:47:22 +03:00
e8c3905e54 Functions for fixing nan 2019-02-28 20:38:13 +03:00
2acbfe9592 [ci] try cuda builds on windows, skip building uselib_track if opencv does not have cuda support 2019-02-28 18:13:17 +01:00
139ddab7fa remove unused cpu function in .cu file 2019-02-28 18:13:17 +01:00
3b43d32dff [ci] add tests for included libs, add link path for linux 2019-02-28 18:13:17 +01:00
caeb0aee05 [CI] add CUDA build tests in matrix 2019-02-28 18:13:09 +01:00
2a874b49a0 [CI] trying to improve resiliency 2019-02-28 18:13:09 +01:00
ef88900eb0 [CMakeLists.txt] remove references to C since CXX is the only way to build the project 2019-02-28 18:13:09 +01:00
56a6e0cdba [travis/appveyor] add CI, remove broken empty include path from CMakeLists.txt 2019-02-28 18:12:55 +01:00
051b8b7d6f Added a compilation for AVX = 1 in CircleCI 2019-02-28 17:04:44 +03:00
e82505d20e Fixed random values 2019-02-28 16:55:30 +03:00
f37ea75b28 Merge pull request #2461 from cenit/dev/cenit/cmake_quickfix
[CMakeLists.txt] fallback not enabled on windows
2019-02-22 20:08:40 +03:00
fd5312f9c8 [build.ps1] fix wrong path, enable builds only if env variable is present 2019-02-22 16:58:50 +01:00
cccc90b340 CMakeLists.txt temporary fix for CUDA 2019-02-22 18:47:10 +03:00
78b707fa09 fallback enabled only not on windows, need to investigate better the failure anyway 2019-02-22 16:43:37 +01:00
d1ee38ca05 Use both $(cudnn) and $(CUDNN) environment variables in darknet.vcxproj & yolo_cpp_dll.vcxproj 2019-02-22 18:08:24 +03:00
7872cbf198 Update Readme.md 2019-02-22 17:53:27 +03:00
954feac226 Update Readme.md 2019-02-22 17:03:36 +03:00
14650fb649 Merge pull request #2412 from cenit/dev/cenit/yoloAlexey_cmake
add CMakeLists.txt and necessary modules
2019-02-22 16:41:18 +03:00
da9c0be371 [CMakeLists.txt] add a compatibility include directive to workaround some regressions in enable_language(CUDA) 2019-02-22 14:22:10 +01:00
d8291858d5 [FindCUDNN] add another hint for newly suggested ENV variable 2019-02-22 09:14:41 +01:00