diff --git a/python/examples/imdb/benchmark_batch.py b/python/examples/imdb/benchmark_batch.py
index 57ee6816989d4a807a328342f188a7298b7772de..107008f5a3b42afe94c94f8dc66da81f2e0cebcf 100644
--- a/python/examples/imdb/benchmark_batch.py
+++ b/python/examples/imdb/benchmark_batch.py
@@ -40,21 +40,28 @@ def single_func(idx, resource):
             if args.batch_size >= 1:
                 feed_batch = []
                 for bi in range(args.batch_size):
-                    word_ids, label = imdb_dataset.get_words_and_label(line)
+                    word_ids, label = imdb_dataset.get_words_and_label(dataset[
+                        bi])
                     feed_batch.append({"words": word_ids})
                 result = client.predict(feed=feed_batch, fetch=["prediction"])
+                if result is None:
+                    raise ("predict failed.")
             else:
                 print("unsupport batch size {}".format(args.batch_size))
 
     elif args.request == "http":
-        for fn in filelist:
-            fin = open(fn)
-            for line in fin:
-                word_ids, label = imdb_dataset.get_words_and_label(line)
-                r = requests.post(
-                    "http://{}/imdb/prediction".format(args.endpoint),
-                    data={"words": word_ids,
-                          "fetch": ["prediction"]})
+        #TODO: not support yet
+        raise ("no batch predict for http")
+        if args.batch_size >= 1:
+            feed_batch = []
+            for bi in range(args.batch_size):
+                word_ids, label = imdb_dataset.get_words_and_label(dataset[bi])
+                feed_batch.append(word_ids)
+            r = requests.post(
+                "http://{}/imdb/prediction".format(args.endpoint),
+                data={"words": feed_batch,
+                      "fetch": ["prediction"]})
+            print(r)
     end = time.time()
     return [[end - start]]
 
diff --git a/python/examples/imdb/benchmark_batch.sh b/python/examples/imdb/benchmark_batch.sh
index 322aaafffb9b252a307fcb63dc7b910e33e5e002..15b65338b21675fd89056cf32f9a247b385a6a36 100644
--- a/python/examples/imdb/benchmark_batch.sh
+++ b/python/examples/imdb/benchmark_batch.sh
@@ -3,7 +3,7 @@ for thread_num in 1 2 4 8 16
 do
 for batch_size in 1 2 4 8 16 32 64 128 256 512
 do
-    $PYTHONROOT/bin/python benchmark_batch.py --thread $thread_num --batch_size $batch_size --model imdbo_bow_client_conf/serving_client_conf.prototxt --request rpc > profile 2>&1
+    $PYTHONROOT/bin/python benchmark_batch.py --thread $thread_num --batch_size $batch_size --model imdb_bow_client_conf/serving_client_conf.prototxt --request rpc > profile 2>&1
     echo "========================================"
     echo "batch size : $batch_size" >> profile_log
     $PYTHONROOT/bin/python ../util/show_profile.py profile $thread_num >> profile_log
diff --git a/tools/serving_build.sh b/tools/serving_build.sh
index e4bf6ece3a9df1808b9190e9e77d8d2e8aba62c0..920134a904e02c147eb271070e6abed99c87b606 100644
--- a/tools/serving_build.sh
+++ b/tools/serving_build.sh
@@ -288,15 +288,6 @@ function python_test_bert() {
             pip install paddle_serving_app
             check_cmd "head -n 10 data-c.txt | python bert_client.py --model bert_chinese_L-12_H-768_A-12_client/serving_client_conf.prototxt"
             kill_server_process
-            # python prepare_model.py 20
-            # sh get_data.sh
-            # check_cmd "python -m paddle_serving_server.serve --model bert_seq20_model/ --port 9292 &"
-            # sleep 5
-            # pip install paddle_serving_app
-            # check_cmd "head -n 10 data-c.txt | python bert_client.py --model bert_seq20_client/serving_client_conf.prototxt"
-            # kill_server_process
-            # ps -ef | grep "paddle_serving_server" | grep -v grep | awk '{print $2}' | xargs kill
-            # ps -ef | grep "serving" | grep -v grep | awk '{print $2}' | xargs kill
             echo "bert RPC inference pass" 
             ;;
         GPU)
@@ -312,14 +303,6 @@ function python_test_bert() {
             pip install paddle_serving_app
             check_cmd "head -n 10 data-c.txt | python bert_client.py --model bert_chinese_L-12_H-768_A-12_client/serving_client_conf.prototxt"
             kill_server_process
-            # python prepare_model.py 20
-            # sh get_data.sh
-            # check_cmd "python -m paddle_serving_server_gpu.serve --model bert_seq20_model/ --port 9292 --gpu_ids 0 &"
-            # sleep 5
-            # pip install paddle_serving_app
-            # check_cmd "head -n 10 data-c.txt | python bert_client.py --model bert_seq20_client/serving_client_conf.prototxt"
-            # kill_server_process
-            # ps -ef | grep "paddle_serving_server" | grep -v grep | awk '{print $2}' | xargs kill
             echo "bert RPC inference pass"
             ;;
         *)
@@ -337,11 +320,11 @@ function python_test_imdb() {
     case $TYPE in
         CPU)
             sh get_data.sh
-            sleep 5
             check_cmd "python -m paddle_serving_server.serve --model imdb_cnn_model/ --port 9292 &"
+            sleep 5
             check_cmd "head test_data/part-0 | python test_client.py imdb_cnn_client_conf/serving_client_conf.prototxt imdb.vocab"
             echo "imdb CPU RPC inference pass"
-            ps -ef | grep "paddle_serving_server" | grep -v grep | awk '{print $2}' | xargs kill
+            kill_server_process
             rm -rf work_dir1
             sleep 5
 
@@ -352,6 +335,13 @@ function python_test_imdb() {
             ps -ef | grep "paddle_serving_server" | grep -v grep | awk '{print $2}' | xargs kill
             ps -ef | grep "text_classify_service.py" | grep -v grep | awk '{print $2}' | xargs kill
             echo "imdb CPU HTTP inference pass"
+
+            # test batch predict
+            check_cmd "python -m paddle_serving_server.serve --model imdb_bow_model --thread 4 --port 9292 &"
+            sleep 5
+            check_cmd "python benchmark_batch.py --thread 4 --batch_size 8 --model imdb_bow_client_conf/serving_client_conf.prototxt --request rpc --endpoint 127.0.0.1:9292"
+            kill_server_process
+            echo "imdb CPU rpc batch inference pass"
             ;;
         GPU)
             echo "imdb ignore GPU test"