AlexeyAB
|
dc7277f152
|
Increased speed on Tensor Cores for csresnext50-panet-spp.cfg model
|
2019-12-20 16:07:34 +03:00 |
Alexey
|
3004ee851c
|
Update Readme.md
|
2019-12-18 15:59:01 +03:00 |
AlexeyAB
|
7f25e50279
|
Minor fix
|
2019-12-18 15:34:02 +03:00 |
AlexeyAB
|
01ad1f5415
|
Added csresnext50-panet-spp.cfg
|
2019-12-18 15:23:51 +03:00 |
AlexeyAB
|
e43a1c424d
|
Fixed a bug in counters_per_class=
|
2019-12-18 03:53:40 +03:00 |
AlexeyAB
|
e33ecb785e
|
Added param counters_per_class=100,200,... <for each class> to [yolo] and [Gaussian_yolo] to solve class-imbalance-problem
|
2019-12-17 18:03:42 +03:00 |
AlexeyAB
|
63396082d7
|
Another one minor fix
|
2019-12-16 18:41:28 +03:00 |
AlexeyAB
|
972608c8ef
|
Minor fix
|
2019-12-16 18:32:20 +03:00 |
AlexeyAB
|
f4b2612b04
|
Added the ability to compile in C++ (Makefile USE_CPP=1)
|
2019-12-16 18:26:50 +03:00 |
AlexeyAB
|
1626f9d495
|
Some fixes
|
2019-12-16 17:53:58 +03:00 |
AlexeyAB
|
f831835125
|
Fixed DropBlock bug for Conv-networks. When are used: random=1 and [dropout] dropblock=1
|
2019-12-16 16:17:44 +03:00 |
AlexeyAB
|
f2fc239096
|
Stretch conv-kernel
|
2019-12-15 01:02:59 +03:00 |
AlexeyAB
|
b78aa3961b
|
Dropped out items are has compensated
|
2019-12-14 22:50:33 +03:00 |
AlexeyAB
|
5e035565ff
|
Rotate conv-kernel
|
2019-12-14 18:11:09 +03:00 |
AlexeyAB
|
a08c872564
|
Sway conv-kernel
|
2019-12-14 03:07:01 +03:00 |
AlexeyAB
|
7ae1ae5641
|
Fixed activation=NORM_CHAN gradient
|
2019-12-13 18:52:54 +03:00 |
AlexeyAB
|
764872a190
|
Minor fix
|
2019-12-13 03:17:55 +03:00 |
AlexeyAB
|
642c065c0e
|
Fixed DropBlock [dropout] dropblock_size_abs=7 dropblock_size_rel=0.5
|
2019-12-13 01:22:03 +03:00 |
AlexeyAB
|
d137d304c1
|
Fixed backward_scale_channels_kernel() for ASFF
|
2019-12-12 14:52:51 +03:00 |
AlexeyAB
|
1df3ddc7d6
|
Added DropBlock. Use [dropout] dropblock=1 dropblock_size=0.5 probability=0.1
|
2019-12-11 13:16:41 +03:00 |
AlexeyAB
|
a616fcd28e
|
Minor fix
|
2019-12-10 20:34:03 +03:00 |
AlexeyAB
|
dbe34d7865
|
Accelerated [Gaussian_yolo] layer
|
2019-12-10 01:37:35 +03:00 |
AlexeyAB
|
fc173454b1
|
Fixed time_step=1 for darknet.so/dll
|
2019-12-09 15:28:31 +03:00 |
AlexeyAB
|
c7e3ba3ed4
|
Fixed memory leak in ./darknet detector demo ...
|
2019-12-09 15:21:22 +03:00 |
AlexeyAB
|
2a873f3448
|
Added label_smooth_eps=0.1 for [net] layer for Label Smoothing for Classifier
|
2019-12-08 17:25:01 +03:00 |
AlexeyAB
|
318919e1cb
|
Added label_smooth_eps=0.1 for [yolo] and [Gaussian_yolo] layer for Label Smoothing
|
2019-12-08 16:39:28 +03:00 |
AlexeyAB
|
67e94dc1de
|
Minor fix
|
2019-12-08 01:03:56 +03:00 |
AlexeyAB
|
ffb4a1df67
|
Compile fix
|
2019-12-08 00:26:23 +03:00 |
AlexeyAB
|
3aa2e45ad3
|
Compile fix
|
2019-12-07 17:42:01 +03:00 |
AlexeyAB
|
13f064f1be
|
Improved mosaic=1 for Detector
|
2019-12-07 17:35:30 +03:00 |
AlexeyAB
|
87f36b79f5
|
Another version of mosaic=1 for Detector
|
2019-12-07 16:09:26 +03:00 |
AlexeyAB
|
4f52ba1a25
|
Remove eps=0.0001 from gradient of normalize_channels_softmax
|
2019-12-07 00:22:06 +03:00 |
AlexeyAB
|
c9c745ccf1
|
Added [convolutional] activation=normalize_channels_softmax
|
2019-12-06 23:25:12 +03:00 |
AlexeyAB
|
5d13aad887
|
Fixed types in the image_opencv.cpp/h that could lead to a memory leak
|
2019-12-06 20:37:08 +03:00 |
AlexeyAB
|
ebf1d7a73b
|
Minor fixes
|
2019-12-06 03:46:34 +03:00 |
AlexeyAB
|
aaba21bdd3
|
Added ./darknet detector valid for KITTI and BDD (were previously MSCOCO and ImageNet)
|
2019-12-05 19:12:46 +03:00 |
AlexeyAB
|
27da7f85c7
|
Minor fix
|
2019-12-05 16:14:34 +03:00 |
AlexeyAB
|
2207acd9c4
|
Minor fix
|
2019-12-05 15:26:44 +03:00 |
AlexeyAB
|
142fcdeb1e
|
Added blur=1 for Classifier training if OPENCV=1 (removes textures like in stylized-ImageNet)
|
2019-12-04 23:37:06 +03:00 |
AlexeyAB
|
efc5478a23
|
Minor fix
|
2019-12-04 17:46:40 +03:00 |
AlexeyAB
|
5ddf9c74a5
|
Added mosaic=1 data augmentation for Detector. Fixed blur=. Fixed [scale_channels] for ASFF.
|
2019-12-04 17:42:32 +03:00 |
AlexeyAB
|
c6f0f7b22c
|
reduced pinned memory step
|
2019-12-04 15:36:07 +03:00 |
AlexeyAB
|
649abac372
|
Fixed blur= parameter for train detector
|
2019-12-04 02:02:22 +03:00 |
AlexeyAB
|
5d0352f961
|
Fixed CPU-RAM + GPU-processing bug for Dropout-layer
|
2019-12-03 18:35:57 +03:00 |
Alexey
|
213b82a1bd
|
Merge pull request #2912 from cenit/dev/cenit/setup
add setup prerequisite script scaffolding
|
2019-12-03 13:45:59 +03:00 |
Stefano Sinigardi
|
99e0e57dc3
|
move setup scripts to scripts subfolder
|
2019-12-03 07:22:00 +01:00 |
AlexeyAB
|
e5d464d3d0
|
Added mosaic=1 data augmentation for Classifier
|
2019-12-02 17:41:45 +03:00 |
AlexeyAB
|
acb578cb37
|
Fixed training with GPU=1 CUDNN=0
|
2019-12-02 16:46:13 +03:00 |
AlexeyAB
|
a95bc3f5b6
|
Fixed memory leak for ./darknet classifier predict ... < list.txt
|
2019-12-02 14:58:35 +03:00 |
AlexeyAB
|
9064755c3b
|
Minor fix
|
2019-12-02 03:11:22 +03:00 |