Open the-cat-crying opened 3 years ago
Try the latest commit, I fixed a bug recently.
I found that if you directly compile opencv without compiling cuda, cudnn and other options, it will not appear‘Segmentation fault (core dumped) ’.
Finally able to run
I'm on the latest commit and still seeing this segfault while following the VOC tutorial.
When I get to training...
896 x 896
Create 64 permanent cpu-threads
mosaic=1 - compile Darknet with OpenCV for using mosaic=1
mosaic=1 - compile Darknet with OpenCV for using mosaic=1
mosaic=1 - compile Darknet with OpenCV for using mosaic=1
mosaic=1 - compile Darknet with OpenCV for using mosaic=1
mosaic=1 - compile Darknet with OpenCV for using mosaic=1
free(): corrupted unsorted chunks
Segmentation fault (core dumped)
I built with the following options (and have tried a few variants of them). Always with OPENCV=0 so I don't know why it's compiling with opencv.
GPU=1
CUDNN=1 # I have tried with 0
CUDNN_HALF=0
OPENCV=0
AVX=0
OPENMP=0
LIBSO=1 # I have tried with 0
ZED_CAMERA=0
ZED_CAMERA_v2_8=0
compile opencv and OPENCV=1
158 conv 512 1 x 1/ 1 19 x 19 x1024 -> 19 x 19 x 512 0.379 BF 159 conv 1024 3 x 3/ 1 19 x 19 x 512 -> 19 x 19 x1024 3.407 BF 160 conv 18 1 x 1/ 1 19 x 19 x1024 -> 19 x 19 x 18 0.013 BF 161 yolo [yolo] params: iou loss: ciou (4), iou_norm: 0.07, obj_norm: 1.00, cls_norm: 1.00, delta_norm: 1.00, scale_x_y: 1.05 nms_kind: greedynms (1), beta = 0.600000 Total BFLOPS 127.232 avg_outputs = 1046213 Allocate additional workspace_size = 189.86 MB Loading weights from ../yolov4.conv.137... seen 64, trained: 0 K-images (0 Kilo-batches_64) Done! Loaded 137 layers from weights-file Learning Rate: 0.0013, Momentum: 0.949, Decay: 0.0005 Detection layer: 139 - type = 28 Detection layer: 150 - type = 28 Detection layer: 161 - type = 28 If error occurs - run training with flag: -dont_show Create 6 permanent cpu-threads Segmentation fault (core dumped)