diff --git a/model_compression/bert/bash_script/XTC/layer_reduction.sh b/model_compression/bert/bash_script/XTC/layer_reduction.sh
index b8af70200..dc68db445 100644
--- a/model_compression/bert/bash_script/XTC/layer_reduction.sh
+++ b/model_compression/bert/bash_script/XTC/layer_reduction.sh
@@ -18,7 +18,7 @@ mkdir -p ${SAVE_PATH}
 # %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% the following command will first download huggingface models and then compress %%%%%%%
 MODEL=yoshitomo-matsubara/bert-base-uncased-${TASK_NAME} ## for both student and teacher
 python -m torch.distributed.launch --nproc_per_node=1 \
-  --master_port 66665 \
+  --master_port 6665 \
   run_glue_no_trainer.py \
   --seed 42 \
   --distill_method ${STAGE} \
@@ -47,7 +47,7 @@ python -m torch.distributed.launch --nproc_per_node=1 \
 # MODEL=yoshitomo-matsubara/bert-base-uncased-${TASK_NAME} ## for both student and teacher
 # student_model=./out/XTC/layer_reduction/best/pytorch_model.bin # <====================================================================Need student model
 # python -m torch.distributed.launch --nproc_per_node=1 \
-#   --master_port 66665 \
+#   --master_port 6665 \
 #   run_glue_no_trainer.py \
 #   --pretrained_dir_student  ${student_model} \
 #   --seed 42 \
diff --git a/model_compression/bert/bash_script/XTC/layer_reduction_1bit.sh b/model_compression/bert/bash_script/XTC/layer_reduction_1bit.sh
index 12a40e05b..3ed8038b0 100644
--- a/model_compression/bert/bash_script/XTC/layer_reduction_1bit.sh
+++ b/model_compression/bert/bash_script/XTC/layer_reduction_1bit.sh
@@ -19,7 +19,7 @@ mkdir -p ${SAVE_PATH}
 MODEL=yoshitomo-matsubara/bert-base-uncased-${TASK_NAME} ## for both student and teacher
 student_model=./out/XTC/layer_reduction/best/pytorch_model.bin # <====================================================================Need student model
 python -m torch.distributed.launch --nproc_per_node=1 \
-  --master_port 66664 \
+  --master_port 6664 \
   run_glue_no_trainer.py \
   --pretrained_dir_student  ${student_model} \
   --seed 42 \
diff --git a/model_compression/bert/bash_script/XTC/quant_1bit.sh b/model_compression/bert/bash_script/XTC/quant_1bit.sh
index b6c17861e..3f0a89aab 100644
--- a/model_compression/bert/bash_script/XTC/quant_1bit.sh
+++ b/model_compression/bert/bash_script/XTC/quant_1bit.sh
@@ -18,7 +18,7 @@ mkdir -p ${SAVE_PATH}
 # %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% the following command will first download huggingface models and then compress %%%%%%%
 MODEL=yoshitomo-matsubara/bert-base-uncased-${TASK_NAME} ## for both student and teacher
 python -m torch.distributed.launch --nproc_per_node=1 \
-  --master_port 66664 \
+  --master_port 6664 \
   run_glue_no_trainer.py \
   --seed 42 \
   --distill_method ${STAGE} \
@@ -43,7 +43,7 @@ python -m torch.distributed.launch --nproc_per_node=1 \
 # TEACHER=/blob/users/xwu/compression/huggingface_models/bert-base-uncased-${TASK_NAME}/pytorch_model.bin
 # STUDENT=${TEACHER}
 # python -m torch.distributed.launch --nproc_per_node=1 \
-#   --master_port 66667 \
+#   --master_port 6667 \
 #   run_glue_no_trainer_clean.py \
 #   --seed 42 \
 #   --distill_method ${STAGE} \
diff --git a/model_compression/bert/bash_script/ZeroQuant/zero_quant.sh b/model_compression/bert/bash_script/ZeroQuant/zero_quant.sh
index 71c638e0f..539fb60d7 100644
--- a/model_compression/bert/bash_script/ZeroQuant/zero_quant.sh
+++ b/model_compression/bert/bash_script/ZeroQuant/zero_quant.sh
@@ -14,7 +14,7 @@ mkdir -p ${SAVE_PATH}
 # %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% the following command will first download huggingface models and then compress %%%%%%%
 MODEL=yoshitomo-matsubara/bert-base-uncased-${TASK_NAME} ## for both student and teacher
 python -m torch.distributed.launch --nproc_per_node=1 \
-  --master_port 66664 \
+  --master_port 6664 \
   run_glue_no_trainer.py \
   --seed 42 \
   --distill_method one_stage \
@@ -37,7 +37,7 @@ python -m torch.distributed.launch --nproc_per_node=1 \
 # TEACHER=/blob/users/xwu/compression/huggingface_models/bert-base-uncased-${TASK_NAME}/pytorch_model.bin
 # STUDENT=${TEACHER}
 # python -m torch.distributed.launch --nproc_per_node=1 \
-#   --master_port 66667 \
+#   --master_port 6667 \
 #   run_glue_no_trainer_clean.py \
 #   --seed 42 \
 #   --distill_method ${STAGE} \
diff --git a/model_compression/bert/bash_script/ZeroQuant/zero_quant_lkd.sh b/model_compression/bert/bash_script/ZeroQuant/zero_quant_lkd.sh
index 528226952..e9005260f 100644
--- a/model_compression/bert/bash_script/ZeroQuant/zero_quant_lkd.sh
+++ b/model_compression/bert/bash_script/ZeroQuant/zero_quant_lkd.sh
@@ -39,7 +39,7 @@ python -m torch.distributed.launch --nproc_per_node=1 \
 # TEACHER=/blob/users/xwu/compression/huggingface_models/bert-base-uncased-${TASK_NAME}/pytorch_model.bin
 # STUDENT=${TEACHER}
 # python -m torch.distributed.launch --nproc_per_node=1 \
-#   --master_port 66667 \
+#   --master_port 6667 \
 #   run_glue_no_trainer_clean.py \
 #   --seed 42 \
 #   --distill_method ${STAGE} \
diff --git a/model_compression/bert/bash_script/layer_reduction.sh b/model_compression/bert/bash_script/layer_reduction.sh
index 31e01b210..c5df728d9 100644
--- a/model_compression/bert/bash_script/layer_reduction.sh
+++ b/model_compression/bert/bash_script/layer_reduction.sh
@@ -65,7 +65,7 @@ CONFIG=${config_json}
 # %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% the following command will first download huggingface models and then compress %%%%%%%
 MODEL=yoshitomo-matsubara/bert-base-uncased-${TASK_NAME} ## for both student and teacher
 run_cmd="python -m torch.distributed.launch --nproc_per_node=1 \
-  --master_port 66664 \
+  --master_port 6664 \
   run_glue_no_trainer.py \
   --seed 42 \
   --distill_method ${STAGE} \
diff --git a/model_compression/bert/bash_script/pruning_row.sh b/model_compression/bert/bash_script/pruning_row.sh
index 8836b201b..4114f3d09 100644
--- a/model_compression/bert/bash_script/pruning_row.sh
+++ b/model_compression/bert/bash_script/pruning_row.sh
@@ -67,7 +67,7 @@ CONFIG=${config_json}
 # %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% the following command will first download huggingface models and then compress %%%%%%%
 MODEL=yoshitomo-matsubara/bert-base-uncased-${TASK_NAME} ## for both student and teacher
 run_cmd="python -m torch.distributed.launch --nproc_per_node=1 \
-  --master_port 66669 \
+  --master_port 6669 \
   run_glue_no_trainer.py \
   --seed 42 \
   --distill_method ${STAGE} \
diff --git a/model_compression/bert/bash_script/pruning_sparse.sh b/model_compression/bert/bash_script/pruning_sparse.sh
index e6b916f52..a1813fe67 100644
--- a/model_compression/bert/bash_script/pruning_sparse.sh
+++ b/model_compression/bert/bash_script/pruning_sparse.sh
@@ -67,7 +67,7 @@ CONFIG=${config_json}
 # %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% the following command will first download huggingface models and then compress %%%%%%%
 MODEL=yoshitomo-matsubara/bert-base-uncased-${TASK_NAME} ## for both student and teacher
 run_cmd="python -m torch.distributed.launch --nproc_per_node=1 \
-  --master_port 66668 \
+  --master_port 6668 \
   run_glue_no_trainer.py \
   --seed 42 \
   --distill_method ${STAGE} \
diff --git a/model_compression/bert/bash_script/quant_activation.sh b/model_compression/bert/bash_script/quant_activation.sh
index 9f308cd4c..4330d3e29 100644
--- a/model_compression/bert/bash_script/quant_activation.sh
+++ b/model_compression/bert/bash_script/quant_activation.sh
@@ -67,7 +67,7 @@ CONFIG=${config_json}
 # %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% the following command will first download huggingface models and then compress %%%%%%%
 MODEL=yoshitomo-matsubara/bert-base-uncased-${TASK_NAME} ## for both student and teacher
 run_cmd="python -m torch.distributed.launch --nproc_per_node=1 \
-  --master_port 66666 \
+  --master_port 6666 \
   run_glue_no_trainer.py \
   --seed 42 \
   --distill_method ${STAGE} \
diff --git a/model_compression/bert/bash_script/quant_weight.sh b/model_compression/bert/bash_script/quant_weight.sh
index e08311d26..de2afe9ff 100644
--- a/model_compression/bert/bash_script/quant_weight.sh
+++ b/model_compression/bert/bash_script/quant_weight.sh
@@ -66,7 +66,7 @@ CONFIG=${config_json}
 # %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% the following command will first download huggingface models and then compress %%%%%%%
 MODEL=yoshitomo-matsubara/bert-base-uncased-${TASK_NAME} ## for both student and teacher
 run_cmd="python -m torch.distributed.launch --nproc_per_node=1 \
-  --master_port 66665 \
+  --master_port 6665 \
   run_glue_no_trainer.py \
   --seed 42 \
   --distill_method ${STAGE} \
diff --git a/model_compression/cifar/run_compress.sh b/model_compression/cifar/run_compress.sh
index 8a4a7a71b..11dac36ed 100644
--- a/model_compression/cifar/run_compress.sh
+++ b/model_compression/cifar/run_compress.sh
@@ -11,7 +11,7 @@ python -m torch.distributed.launch --nproc_per_node=1 \
 ### Step 2: compress: channel pruning 
 ### you may enbale other compression methods, see ds_config.json or our compression tutorial 
 python -m torch.distributed.launch --nproc_per_node=1 \
-  --master_port 66665 \
+  --master_port 6665 \
   train.py \
   --deepspeed_config config/ds_config_channel_prune.json \
   --deepspeed \
@@ -32,7 +32,7 @@ python -m torch.distributed.launch --nproc_per_node=1 \
 #### Step 2: compress: channel pruning 
 #### you may enbale other compression methods, see ds_config.json or our compression tutorial 
 # python -m torch.distributed.launch --nproc_per_node=1 \
-#   --master_port 66665 \
+#   --master_port 6665 \
 #   train.py \
 #   --deepspeed_config config/ds_config_channel_prune.json \
 #   --deepspeed \