MPI train exit: check_return 'hadoop_ls_file: Generate filelist for train meta failed'
Created by: Jarlonyan
MPI训练异常退出,信息如下,请问这是什么原因呢?
Job Status Series :
epilogue finished
epilogue started
user_main failed (Status:kTaskStatusExited)
user_main startedprologue finishedprologue startedReady
train.log中提示:
check_return 'hadoop_ls_file: Generate filelist for train meta failed'