set HCCL_CONNECT_TIMEOUT=600 for transformer distribute training

4 years ago · 0f610e11de
parent a492cc470b
commit 0f610e11de
1 changed files with 1 additions and 0 deletions
--- a/model_zoo/official/nlp/transformer/scripts/run_distribute_train_ascend.sh
+++ b/model_zoo/official/nlp/transformer/scripts/run_distribute_train_ascend.sh
@ -28,6 +28,7 @@ cd run_distribute_train || exit
 EPOCH_SIZE=$2
 DATA_PATH=$3

+export HCCL_CONNECT_TIMEOUT=600
 export RANK_TABLE_FILE=$4
 export RANK_SIZE=$1
 export HCCL_FLAG=1