OrgLion-ML
/
darknet
mirror of https://github.com/AlexeyAB/darknet.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity
2,187 Commits 1 Branch 7 Tags 361 MiB
Commit Graph

2187 Commits

Author SHA1 Message Date
Stefano Sinigardi da15b141f0 trying to fix cuda path 2021-01-23 22:03:43 +01:00
Stefano Sinigardi 63caceab0d add manual trigger for github actions 2021-01-23 17:04:23 +01:00
Stefano Sinigardi e72a11a04c update README 2021-01-23 16:29:51 +01:00
Stefano Sinigardi 1735d57ad1 rework integrated libs logic 2021-01-23 16:06:36 +01:00
Stefano Sinigardi a8013ebcfb do not default on ninja on windows 2021-01-23 09:02:33 +01:00
Stefano Sinigardi 4de12a1e7e default on opencv4.5 on vcpkg 2021-01-23 09:02:22 +01:00
Alexey 64efa721ed
Merge pull request #4674 from nog642/fix 
Minor fix
 2021-01-21 22:30:15 +03:00
Stefano Sinigardi 00d578e327
update GitHub actions (#7095) 
* update to run-vcpkg@v6

* bump to cmake 3.18

* use newer cuda syntax, fail if cuda enabled but not found

* update win build script

* improve travis with non-cuda-enabled builds

* improve windows building process
 2021-01-21 17:36:00 +01:00
Ievgen Popovych f57636ad65
CMake CUDA configuration improvements (#7164) 
* CMake: Fail generation if CUDA was requested but not found

..instead of resorting to CPU build silently.

This makes builds predictable and
removes the confusion of `ENABLE_CUDA` option being merely a suggestion.

Signed-off-by: Ievgen Popovych <ievgenp@seetrue.ai>

* CMake: Be more precise with ZED SDK messages

CMake generation with `ENABLE_CUDA=OFF` and `ENABLE_ZED_CAMERA=OFF`
would spit out "ZED SDK not enabled, since it requires CUDA" messages
even though ZED SDK was explicitly disabled.
Fix this by being a bit more precise with messages.

Signed-off-by: Ievgen Popovych <ievgenp@seetrue.ai>

* CMake: Honor CUDNN_ENABLE option

Setting CUDNN_ENABLE to OFF would not disable CUDNN support if
CUDNN libraries are available.

Fix this behavior by only looking for CUDNN if it was requested (CUDNN_ENABLE)
and predicate actual settings on CUDNN_FOUND.

Signed-off-by: Ievgen Popovych <ievgenp@seetrue.ai>

* github: actions: Explicitly set CUDA state in CI
 2021-01-21 17:31:21 +01:00
Alexey 103d301ccb
Update README.md 2021-01-09 05:49:43 +03:00
Alexey b11af9e93b
Update README.md 2021-01-05 16:48:02 +03:00
Alexey 0968058b54
Merge pull request #7167 from zzpbuaa/zzpbuaa-patch-1 
Fix bug #6897 OSError: exception: access violation writing
 2021-01-04 01:31:44 +03:00
Alexey 95339f2df5
Update README.md 2020-12-27 04:05:20 +03:00
Alexey f83231fa1d
Update README.md 2020-12-26 17:45:07 +03:00
BlackHill de Eagle dd02b915be
Fix bug OSError: exception: access violation writing #6897 
Avoid read and write image at the same time.
 2020-12-23 15:18:56 +08:00
Alexey c5b8bc7f24
Update README.md 2020-12-21 20:39:35 +03:00
Alexey 84fa42a966
Update README.md 2020-12-18 05:45:09 +03:00
Alexey d18d73ff5f
Readme.md for Xilinx Zynq Ultrascale+ (DPU) ZCU102/ZCU104 2020-12-18 02:25:53 +03:00
AlexeyAB a298f94255 Minor fix for cfg-files 2020-12-15 07:09:58 +03:00
AlexeyAB e7d029c119 Merge branch 'master' of https://github.com/AlexeyAB/darknet 2020-12-15 01:34:25 +03:00
AlexeyAB 360206f04f Fixed equidistant point threshold 2020-12-15 01:34:12 +03:00
Alexey 972b25c5c8
Update README.md 2020-12-14 18:19:49 +03:00
AlexeyAB a488131a11 Fix scale_xy for CPU 2020-12-13 05:02:24 +03:00
AlexeyAB b25c2c6cbd Final training fix 2020-12-12 05:52:58 +03:00
AlexeyAB 2da4f0de84 Reverted some changes in yolo_layer.c to fix a bug 2020-12-12 04:37:31 +03:00
AlexeyAB 1eda55d791 Fixed EMA. Fixed display of GIoU / IoU loss. 2020-12-11 21:53:34 +03:00
AlexeyAB 6be487b51f Fixed cspx-p7-mish.cfg 2020-12-11 05:56:07 +03:00
AlexeyAB b5ff7f4cd3 Save weights for each 1000 iterations if max_batches < 10000 2020-12-10 23:13:26 +03:00
Alexey 1b936ab9ad
Update README.md 2020-12-10 22:56:54 +03:00
AlexeyAB a72bb059d8 Show GIoU/IoU loss instead of RMSE by default 2020-12-10 22:12:45 +03:00
Alexey 4ee3be7e68
Update README.md 2020-12-07 17:27:02 +03:00
Alexey 66dfa5bba4
Update README.md 2020-12-07 17:22:17 +03:00
AlexeyAB ef279575c5 Fixed yolov4-csp.cfg, yolov4x-mish.cfg 2020-12-07 02:10:09 +03:00
AlexeyAB c47b24a81d Fixed new_coords=1, yolov4-csp.cfg, yolov4x-mish.cfg 2020-12-07 00:08:30 +03:00
AlexeyAB 4709f6156c Use default gradient for new_coords=1 2020-12-04 21:01:24 +03:00
AlexeyAB 8d6e56e7b4 Fixed new_coords=1 training 2020-12-04 03:06:08 +03:00
AlexeyAB 312fd2e99a Fix letter_box for yolov4x-mish/csp.cfg: letter_box = net.letter_box; And force for Detection: if (net.letter_box) letter_box = 1; 2020-12-03 02:33:53 +03:00
AlexeyAB 560fda3494 [net] weights_reject_freq=1001 2020-12-02 14:42:05 +03:00
AlexeyAB 6d5addd176 Fixed new_coords=1 scale_x_y = 2.0 for yolov4x-mish.cfg and yolov4-csp.cfg 2020-12-01 17:28:28 +03:00
AlexeyAB eb208de417 Commented in the Makefile: -gencode arch=compute_86,code=[sm_86,compute_86] 2020-12-01 16:45:28 +03:00
AlexeyAB e87042a49a Fixed Makefile and CMakeLists.txt for RTX 3090 2020-12-01 05:14:02 +03:00
AlexeyAB 80e1da22ee Added [net] ema_alpha=0.998 2020-11-29 21:30:41 +03:00
AlexeyAB 332ec6496e [net] equidistant_point=200000 num_sigmas_reject_badlabels=3 badlabels_rejection_percentage=0.1 2020-11-28 17:35:54 +03:00
AlexeyAB bf5084934f Added yolov4-csp.cfg 2020-11-28 02:57:12 +03:00
AlexeyAB 181967937d Minor fix 2020-11-24 14:24:35 +03:00
AlexeyAB 333cc14a06 [net] equidistant_point=1 instead of LR-scheduler 2020-11-24 01:41:00 +03:00
AlexeyAB f056fc3b6a Speedup [yolo] layer 2020-11-23 23:32:55 +03:00
AlexeyAB d2703b0ecf optimized_memory=1, learning_rate=0.001, max_delta=20, 5, 2 for yolov4x-mish.cfg 2020-11-23 19:55:26 +03:00
AlexeyAB be906dfa0e Minor fix 2020-11-21 05:51:31 +03:00
AlexeyAB d18e22ae1b fix new_coords for CPU 2020-11-20 21:14:31 +03:00