caffe2用detectron预训练模型训练自己数据集，训练半小时左右中断报错

[mstong68]

json_stats: {"accuracy_cls": 0.998047, "eta": "1:46:24", "iter": 9360, "loss": 0.007533, "loss_bbox": 0.001410, "loss_cls": 0.006326, "loss_rpn_bbox_fpn2": 0.000000, "loss_rpn_bbox_fpn3": 0.000003, "loss_rpn_bbox_fpn4": 0.000035, "loss_rpn_bbox_fpn5": 0.000000, "loss_rpn_bbox_fpn6": 0.000000, "loss_rpn_cls_fpn2": 0.000001, "loss_rpn_cls_fpn3": 0.000019, "loss_rpn_cls_fpn4": 0.000013, "loss_rpn_cls_fpn5": 0.000000, "loss_rpn_cls_fpn6": 0.000000, "lr": 0.002500, "mb_qsize": 64, "mem": 2444, "time": 0.126069}
*** Aborted at 1523366002 (unix time) try "date -d @1523366002" if you are using GNU date ***
PC: @ 0x7f58f127b545 (unknown)
*** SIGSEGV (@0x0) received by PID 5930 (TID 0x7f58355fd700) from PID 0; stack trace: ***
@ 0x7f58f15d4390 (unknown)
@ 0x7f58f127b545 (unknown)
@ 0x7f58f127d184 __libc_malloc
@ 0x7f58e8345c51 (unknown)
@ 0x7f58e83ad061 (unknown)
@ 0x7f58e83ad29e (unknown)
@ 0x7f58e843f1b9 (unknown)
@ 0x4bc3fa PyEval_EvalFrameEx
@ 0x4b9ab6 PyEval_EvalCodeEx
@ 0x4d54b9 (unknown)
@ 0x4eebee (unknown)
@ 0x4a577e PyObject_Call
@ 0x4c5e10 PyEval_CallObjectWithKeywords
@ 0x7f58db00e3b0 pybind11::detail::object_api<>::operator()<>()
@ 0x7f58db00fb31 caffe2::python::PythonOpBase<>::RunOnDevice()
@ 0x7f58dafd72ab caffe2::Operator<>::Run()
@ 0x7f58db020e1a _ZN6caffe213GPUFallbackOpINS_6python8PythonOpINS_10CPUContextELb0EEENS_11SkipIndicesIJEEEE11RunOnDeviceEv
@ 0x7f58db01d5b5 caffe2::Operator<>::Run()
@ 0x7f58d6e1dbfa caffe2::DAGNet::RunAt()
@ 0x7f58d6e1c8ae caffe2::DAGNetBase::WorkerFunction()
@ 0x7f58ea80fc80 (unknown)
@ 0x7f58f15ca6ba start_thread
@ 0x7f58f130041d clone
@ 0x0 (unknown)

[zj09130913]

请问您这个问题解决了吗？我是在infer_simple测试的时候出现同样的问题了

[sinat_41811362]

请问博主怎么制作自己的能应用于detectron训练的coco数据集呀，有资料的话请把教程给我参考一下可以吗，[e-mail]。谢谢博主了