[E ProcessGroupNCCL.cpp:828] [Rank 3] Watchdog caught collective operation timeout: WorkNCCL...
报错信息[EProcessGroupNCCL.cpp:828][Rank3]Watchdogcaughtcollectiveoperationtimeout:WorkNCCL(SeqNum=5,OpType=BROADCAST,Timeout(ms)=1800000)ranfor1802615millisecondsbeforetimingout可能的原因可能的解决办法bash中输入更改环境变量e