diff --git a/examples/aishell/asr0/local/train.sh b/examples/aishell/asr0/local/train.sh
index 7c0ad07565444dea96fcbd8bc822e585c0d352c2..102c051c164e9ca10c26e6baa4e7d40ca64a9292 100755
--- a/examples/aishell/asr0/local/train.sh
+++ b/examples/aishell/asr0/local/train.sh
@@ -29,6 +29,7 @@ python3 -u ${BIN_DIR}/train.py \
 --seed ${seed}
 else
 python3 -m paddle.distributed.launch --gpus=${CUDA_VISIBLE_DEVICES} ${BIN_DIR}/train.py \
+--ngpu ${ngpu} \
 --config ${config_path} \
 --output exp/${ckpt_name} \
 --model_type ${model_type} \
diff --git a/examples/aishell/asr0/run.sh b/examples/aishell/asr0/run.sh
index 0542e361e1b67e23d10cfe9ac99f4eb3046402c0..114af5a977edd78b300b5bed2bca05d837666bba 100755
--- a/examples/aishell/asr0/run.sh
+++ b/examples/aishell/asr0/run.sh
@@ -2,7 +2,7 @@
 set -e
 source path.sh
 
-gpus=1,2,3
+gpus=0,1,2,3
 stage=0
 stop_stage=100
 conf_path=conf/deepspeech2.yaml    #conf/deepspeech2.yaml or conf/deepspeech2_online.yaml
diff --git a/examples/aishell/asr1/local/train.sh b/examples/aishell/asr1/local/train.sh
index eb9f4f69e7007569ea441eaf3cb35960dfbb5a05..5617f7efe5cca2004bd6b6aae8fc8291edf22f34 100755
--- a/examples/aishell/asr1/local/train.sh
+++ b/examples/aishell/asr1/local/train.sh
@@ -38,6 +38,7 @@ python3 -u ${BIN_DIR}/train.py \
 --benchmark-max-step ${benchmark_max_step}
 else
 python3 -m paddle.distributed.launch --gpus=${CUDA_VISIBLE_DEVICES} ${BIN_DIR}/train.py \
+--ngpu ${ngpu} \
 --seed ${seed} \
 --config ${config_path} \
 --output exp/${ckpt_name} \
diff --git a/examples/callcenter/asr1/local/train.sh b/examples/callcenter/asr1/local/train.sh
index 3d7d35f11af33c1a24fae99dbe1b0823d2dd0a52..03b4588e30cb92c014f18d1509354f3caaa2b311 100755
--- a/examples/callcenter/asr1/local/train.sh
+++ b/examples/callcenter/asr1/local/train.sh
@@ -29,6 +29,7 @@ python3 -u ${BIN_DIR}/train.py \
 --seed ${seed}
 else
 python3 -m paddle.distributed.launch --gpus=${CUDA_VISIBLE_DEVICES} ${BIN_DIR}/train.py \
+--ngpu ${ngpu} \
 --config ${config_path} \
 --output exp/${ckpt_name} \
 --seed ${seed}
diff --git a/examples/librispeech/asr0/local/train.sh b/examples/librispeech/asr0/local/train.sh
index 788f9bf8945e9b9dea11ebcb32a9ddd4fbfa207a..50d1d1922b54005a80d3b90b17e1ea6def5a438a 100755
--- a/examples/librispeech/asr0/local/train.sh
+++ b/examples/librispeech/asr0/local/train.sh
@@ -29,6 +29,7 @@ python3 -u ${BIN_DIR}/train.py \
 --seed ${seed}
 else
 python3 -m paddle.distributed.launch --gpus=${CUDA_VISIBLE_DEVICES} ${BIN_DIR}/train.py \
+--ngpu ${ngpu} \
 --config ${config_path} \
 --output exp/${ckpt_name} \
 --model_type ${model_type} \
diff --git a/examples/librispeech/asr1/local/train.sh b/examples/librispeech/asr1/local/train.sh
index b21c101799126266e1cdfbe3039b394666939d0a..3860d85cf8c7a65810c2565a358aec2f537bde6b 100755
--- a/examples/librispeech/asr1/local/train.sh
+++ b/examples/librispeech/asr1/local/train.sh
@@ -30,6 +30,7 @@ python3 -u ${BIN_DIR}/train.py \
 --seed ${seed}
 else
 python3 -m paddle.distributed.launch --gpus=${CUDA_VISIBLE_DEVICES} ${BIN_DIR}/train.py \
+--ngpu ${ngpu} \
 --config ${config_path} \
 --output exp/${ckpt_name} \
 --seed ${seed}
diff --git a/examples/librispeech/asr2/local/train.sh b/examples/librispeech/asr2/local/train.sh
index a3ad5fd75c23bab60d2e9ba730aaa5f6dc290a7f..560424ea4fe5b79260d12b3678cae01a16df6226 100755
--- a/examples/librispeech/asr2/local/train.sh
+++ b/examples/librispeech/asr2/local/train.sh
@@ -28,6 +28,7 @@ python3 -u ${BIN_DIR}/train.py \
 --seed ${seed}
 else
 python3 -m paddle.distributed.launch --gpus=${CUDA_VISIBLE_DEVICES} ${BIN_DIR}/train.py \
+--ngpu ${ngpu} \
 --model-name u2_kaldi \
 --config ${config_path} \
 --output exp/${ckpt_name} \
diff --git a/examples/ted_en_zh/st0/local/train.sh b/examples/ted_en_zh/st0/local/train.sh
index e366376bb82da7dbefb2744cf1099949a5fadf06..ad00653b70b366addcab126f1e63fc64aec1b37c 100755
--- a/examples/ted_en_zh/st0/local/train.sh
+++ b/examples/ted_en_zh/st0/local/train.sh
@@ -19,11 +19,19 @@ if [ ${seed} != 0 ]; then
     export FLAGS_cudnn_deterministic=True
 fi
 
+if [ ${ngpu} == 0 ]; then
 python3 -u ${BIN_DIR}/train.py \
 --ngpu ${ngpu} \
 --config ${config_path} \
 --output exp/${ckpt_name} \
 --seed ${seed}
+else
+python3 -m paddle.distributed.launch --gpus=${CUDA_VISIBLE_DEVICES} ${BIN_DIR}/train.py \
+--ngpu ${ngpu} \
+--config ${config_path} \
+--output exp/${ckpt_name} \
+--seed ${seed}
+fi
 
 if [ ${seed} != 0 ]; then
     unset FLAGS_cudnn_deterministic
diff --git a/examples/ted_en_zh/st1/local/train.sh b/examples/ted_en_zh/st1/local/train.sh
index a8e4acaa01428c13ab2ba36b73c706119160c3bf..5da64e99c95af8e23c9f027080cddcd72c0a4d17 100755
--- a/examples/ted_en_zh/st1/local/train.sh
+++ b/examples/ted_en_zh/st1/local/train.sh
@@ -20,12 +20,21 @@ if [ ${seed} != 0 ]; then
     export FLAGS_cudnn_deterministic=True
 fi
 
+if [ ${ngpu} == 0 ]; then
 python3 -u ${BIN_DIR}/train.py \
 --ngpu ${ngpu} \
 --config ${config_path} \
 --output exp/${ckpt_name} \
 --checkpoint_path "${ckpt_path}" \
 --seed ${seed}
+else
+python3 -m paddle.distributed.launch --gpus=${CUDA_VISIBLE_DEVICES} ${BIN_DIR}/train.py \
+--ngpu ${ngpu} \
+--config ${config_path} \
+--output exp/${ckpt_name} \
+--checkpoint_path "${ckpt_path}" \
+--seed ${seed}
+fi
 
 if [ ${seed} != 0 ]; then
     unset FLAGS_cudnn_deterministic
@@ -36,4 +45,4 @@ if [ $? -ne 0 ]; then
     exit 1
 fi
 
-exit 0
\ No newline at end of file
+exit 0
diff --git a/examples/timit/asr1/local/train.sh b/examples/timit/asr1/local/train.sh
index c4648c3edfb98220444b97c94cd0835e6cfbf183..661407582cc2776e1676249cbe002bcef8c3c2e7 100755
--- a/examples/timit/asr1/local/train.sh
+++ b/examples/timit/asr1/local/train.sh
@@ -19,10 +19,19 @@ if [ ${seed} != 0  ]; then
     export FLAGS_cudnn_deterministic=True
 fi
 
+if [ ${ngpu} == 0 ]; then
+python3 -u ${BIN_DIR}/train.py \
+--ngpu ${ngpu} \
+--config ${config_path} \
+--output exp/${ckpt_name} \
+--seed ${seed}
+else
 python3 -m paddle.distributed.launch --gpus=${CUDA_VISIBLE_DEVICES} ${BIN_DIR}/train.py \
+--ngpu ${ngpu} \
 --config ${config_path} \
 --output exp/${ckpt_name} \
 --seed ${seed}
+fi
 
 if [ ${seed} != 0 ]; then
     unset FLAGS_cudnn_deterministic
diff --git a/examples/tiny/asr0/local/train.sh b/examples/tiny/asr0/local/train.sh
index 5fecf3de9bcf95994e69274265e8343e8149b015..9060be674e5804005de6b23c2575bd4888efef00 100755
--- a/examples/tiny/asr0/local/train.sh
+++ b/examples/tiny/asr0/local/train.sh
@@ -36,6 +36,7 @@ python3 -u ${BIN_DIR}/train.py \
 --seed ${seed}
 else
 python3 -m paddle.distributed.launch --gpus=${CUDA_VISIBLE_DEVICES} ${BIN_DIR}/train.py \
+--ngpu ${ngpu} \
 --config ${config_path} \
 --output exp/${ckpt_name} \
 --model_type ${model_type} \
diff --git a/examples/tiny/asr1/local/train.sh b/examples/tiny/asr1/local/train.sh
index eb9f4f69e7007569ea441eaf3cb35960dfbb5a05..5617f7efe5cca2004bd6b6aae8fc8291edf22f34 100755
--- a/examples/tiny/asr1/local/train.sh
+++ b/examples/tiny/asr1/local/train.sh
@@ -38,6 +38,7 @@ python3 -u ${BIN_DIR}/train.py \
 --benchmark-max-step ${benchmark_max_step}
 else
 python3 -m paddle.distributed.launch --gpus=${CUDA_VISIBLE_DEVICES} ${BIN_DIR}/train.py \
+--ngpu ${ngpu} \
 --seed ${seed} \
 --config ${config_path} \
 --output exp/${ckpt_name} \