segment error(core dumped

[Note-Liu]

using cude using tfboard segment error(core dumped)

[GOATmessi8]

please provide more information about your error.....I have no idea what happened on your machine

[Note-Liu]

TRAIN: LR: 0.001 MOMENTUM: 0.9 DECAY: 0.0005 BURN_IN: 5 MAXEPOCH: 300 COS: True SYBN: False#True MIX: True NO_MIXUP_EPOCHS: 30 LABAL_SMOOTH: True BATCHSIZE: 1 IMGSIZE: 608 IGNORETHRE: 0.7

train script: python main.py --cfg config/yolov3_baseline.cfg -d VOC --tfboard --ngpu 1 --checkpoint weights/darknet53_feature_mx.pth --start_epoch 0 --half --log_dir log/VOC -s 608

[GOATmessi8]

The version of you cuda, pytorch, apex and so on. And the details of your error information.... I know you use the default script, but obviously the error is not in the script.

[Note-Liu]

If I use only one GPU , Is my train script right? thanks

[GOATmessi8]

I didn't test the code without distributed training, so even with single gpu, I suggest you keep the distributed training. And your batchsize is only one, which could significantly depress your performance.

[lxyyang]

@Note-Liu have you solverd it？？

[Note-Liu]

@Note-Liu have you solverd it？？

no.[cry][cry]

[eternalgogi92]

I think it may be caused by the wrong gcc version error and the DCN module part. When I compile the CenterNet code, it happened once! You can update the gcc to 5.0 or above and try the lateset DCN module from the original github.