Αi SSKrSSKrSSKJrJr SSKJr 0SS\4_SS\4_SS \4_S S \4_S S \4_SS\4_SS\4_SS\4_SS\4_SS\4_SS\4_SS\4_SS\4_SS\4_S S!\4_S"S#\4_S$S%\4_S&\4S'\4S(\4S)\4S*\4S+\4S,\4S-\4S..Er S/r S0r g)1N) REMAINDERArgumentParser) strtoboolPOD_IPhost PADDLE_MASTERmasterPADDLE_DEVICESdevices PADDLE_NNODESnnodesPADDLE_RUN_MODErun_modePADDLE_LOG_LEVEL log_levelPADDLE_LOG_OVERWRITE log_overwritePADDLE_SORT_IPsort_ipPADDLE_NPROC_PER_NODEnproc_per_node PADDLE_JOB_IDjob_id PADDLE_RANKrankPADDLE_LOG_DIRlog_dirPADDLE_MAX_RESTART max_restartPADDLE_ELASTIC_LEVEL elastic_levelPADDLE_ELASTIC_TIMEOUTelastic_timeoutPADDLE_SERVER_NUM server_numPADDLE_TRAINER_NUM trainer_numserverstrainers gloo_port with_gloo start_portipsauto_parallel_configauto_cluster_config)PADDLE_SERVERS_ENDPOINTSPADDLE_TRAINERS_ENDPOINTSPADDLE_GLOO_PORTPADDLE_WITH_GLOOPADDLE_START_PORT PADDLE_IPSPADDLE_AUTO_PARALLEL_CONFIGPADDLE_AUTO_CLUSTERc:SHxn[RRU5cM%[RRU5[RUS3'[R "SUSUS3[ S9 Mz [RR5$)N) http_proxy https_proxy _originalzUnset 'zM' to ensure stable NCCL communication in distributed training (backed up as 'z _original').)category)osenvirongetpopwarningswarn UserWarningcopy) proxy_keys k/var/www/html/banglarbhumi/venv/lib/python3.13/site-packages/paddle/distributed/launch/context/args_envs.py fetch_envsrG2s{2 ::>>) $ 024**..2KBJJ)I. / MM)%""+L:$ 3 ::?? c R[5nURS5nURS[SSS9 URS[SSS9 URS[ S S S9 URS [S S S9 URS[SSS9 URS[SSS9 URS[SSS9 URS[SSS9 URS[ SSS9 URS[SSS9 URS[SSS9 URS[S S!S9 URS"S#S$S%[SS&S9 URS'[SS(S9 URS)[SS*S9 URS+[ S,S-S9 URS.[SS/S9 URS0[S1S2S9 URS3[S4S59 URS6[SS7S9 URS8[ S99 URS:5nURS;[S[Srjs ." vsmh_y#&h_   C(  c* _i8y),.4h_FC=y#&=#._c20#6 ,,!"=#.#$"+C 0",c!2$c*$c*&,#,$:C#@19=3: wrH