Created by: 13488862011

单机训练seq2seq模型时，样本为100w条数据，训练3轮已经成功，但在第四轮时报了如下的错误； 有两个问题：1. 这个错误通过设置可以避免吗？2.由于3轮已经有结果，是否可以增量训练呢？多谢 Fatal Python error: GC object already tracked Current Layer forward/backward stack is

* Aborted at 1479314862 (unix time) try "date -d @1479314862" if you are using GNU date *

Current Layer forward/backward stack is PC: @ 0x7f56ebf1b965 __GI_raise Current Layer forward/backward stack is

* SIGABRT (@0x1f400003870) received by PID 14448 (TID 0x7f55b7b82700) from PID 14448; stack trace: *

Current Layer forward/backward stack is @ 0x7f56ed6d4520 (unknown) Current Layer forward/backward stack is @ 0x7f56ebf1b965 __GI_raise Current Layer forward/backward stack is @ 0x7f56ebf1cd75 __GI_abort Current Layer forward/backward stack is @ 0x7f56ecfe9cf9 Py_FatalError Current Layer forward/backward stack is @ 0x7f56ecf860ce PyTuple_New Current Layer forward/backward stack is @ 0x7f56ecf86bb1 PyTuple_Pack Current Layer forward/backward stack is @ 0x7f56ecf4d098 enum_new Current Layer forward/backward stack is @ 0x7f56ecf89463 type_call Current Layer forward/backward stack is @ 0x7f56ecf34123 PyObject_Call Current Layer forward/backward stack is @ 0x7f56ecfcd8b7 PyEval_EvalFrameEx Current Layer forward/backward stack is @ 0x7f56ecf516e7 gen_send_ex Current Layer forward/backward stack is @ 0x7f56ecf3300b PyIter_Next Current Layer forward/backward stack is @ 0x6a3c0f paddle::PyDataProvider2::loadThread() Current Layer forward/backward stack is @ 0x7f56ec8162a8 execute_native_thread_routine Current Layer forward/backward stack is @ 0x7f56ed6cbd14 start_thread Current Layer forward/backward stack is @ 0x7f56ebfd3bfd clone /home/.jumbo/bin/paddle_local: line 47: 14448 Aborted ${DEBUGGER} $MYDIR/../opt/paddle/bin/paddle_trainer ${@:2}