diff --git a/examples/csmsc/tts2/local/inference_mlu.sh b/examples/csmsc/tts2/local/inference_mlu.sh
new file mode 100755
index 0000000000..d1bade84db
--- /dev/null
+++ b/examples/csmsc/tts2/local/inference_mlu.sh
@@ -0,0 +1,33 @@
+#!/bin/bash
+
+train_output_path=$1
+
+stage=0
+stop_stage=0
+
+# for more GAN Vocoders
+# multi band melgan
+if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
+    python3 ${BIN_DIR}/../inference.py \
+        --inference_dir=${train_output_path}/inference \
+        --am=speedyspeech_csmsc \
+        --voc=mb_melgan_csmsc \
+        --text=${BIN_DIR}/../../assets/sentences.txt \
+        --output_dir=${train_output_path}/pd_infer_out \
+        --phones_dict=dump/phone_id_map.txt \
+        --tones_dict=dump/tone_id_map.txt \
+        --device mlu
+fi
+
+# hifigan
+if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
+    python3 ${BIN_DIR}/../inference.py \
+        --inference_dir=${train_output_path}/inference \
+        --am=speedyspeech_csmsc \
+        --voc=hifigan_csmsc \
+        --text=${BIN_DIR}/../../assets/sentences.txt \
+        --output_dir=${train_output_path}/pd_infer_out \
+        --phones_dict=dump/phone_id_map.txt \
+        --tones_dict=dump/tone_id_map.txt \
+        --device mlu
+fi
diff --git a/examples/csmsc/tts2/local/synthesize_e2e_mlu.sh b/examples/csmsc/tts2/local/synthesize_e2e_mlu.sh
new file mode 100755
index 0000000000..7ad2024ff4
--- /dev/null
+++ b/examples/csmsc/tts2/local/synthesize_e2e_mlu.sh
@@ -0,0 +1,99 @@
+#!/bin/bash
+
+config_path=$1
+train_output_path=$2
+ckpt_name=$3
+
+stage=0
+stop_stage=0
+
+# for more GAN Vocoders
+# multi band melgan
+if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
+    FLAGS_allocator_strategy=naive_best_fit \
+    python3 ${BIN_DIR}/../synthesize_e2e.py \
+        --am=speedyspeech_csmsc \
+        --am_config=${config_path} \
+        --am_ckpt=${train_output_path}/checkpoints/${ckpt_name} \
+        --am_stat=dump/train/feats_stats.npy \
+        --voc=mb_melgan_csmsc \
+        --voc_config=mb_melgan_csmsc_ckpt_0.1.1/default.yaml \
+        --voc_ckpt=mb_melgan_csmsc_ckpt_0.1.1/snapshot_iter_1000000.pdz\
+        --voc_stat=mb_melgan_csmsc_ckpt_0.1.1/feats_stats.npy \
+        --lang=zh \
+        --text=${BIN_DIR}/../../assets/sentences.txt \
+        --output_dir=${train_output_path}/test_e2e \
+        --phones_dict=dump/phone_id_map.txt \
+        --tones_dict=dump/tone_id_map.txt \
+        --ngpu=0 \
+        --nmlu=1
+fi
+
+# the pretrained models haven't release now
+# style melgan
+# style melgan's Dygraph to Static Graph is not ready now
+if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
+    FLAGS_allocator_strategy=naive_best_fit \
+    python3 ${BIN_DIR}/../synthesize_e2e.py \
+        --am=speedyspeech_csmsc \
+        --am_config=${config_path} \
+        --am_ckpt=${train_output_path}/checkpoints/${ckpt_name} \
+        --am_stat=dump/train/feats_stats.npy \
+        --voc=style_melgan_csmsc \
+        --voc_config=style_melgan_csmsc_ckpt_0.1.1/default.yaml \
+        --voc_ckpt=style_melgan_csmsc_ckpt_0.1.1/snapshot_iter_1500000.pdz \
+        --voc_stat=style_melgan_csmsc_ckpt_0.1.1/feats_stats.npy \
+        --lang=zh \
+        --text=${BIN_DIR}/../../assets/sentences.txt \
+        --output_dir=${train_output_path}/test_e2e \
+        --phones_dict=dump/phone_id_map.txt \
+        --tones_dict=dump/tone_id_map.txt \
+        --ngpu=0 \
+        --nmlu=1
+        # --inference_dir=${train_output_path}/inference
+fi
+
+# hifigan
+if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
+    FLAGS_allocator_strategy=naive_best_fit \
+    python3 ${BIN_DIR}/../synthesize_e2e.py \
+        --am=speedyspeech_csmsc \
+        --am_config=${config_path} \
+        --am_ckpt=${train_output_path}/checkpoints/${ckpt_name} \
+        --am_stat=dump/train/feats_stats.npy \
+        --voc=hifigan_csmsc \
+        --voc_config=hifigan_csmsc_ckpt_0.1.1/default.yaml \
+        --voc_ckpt=hifigan_csmsc_ckpt_0.1.1/snapshot_iter_2500000.pdz \
+        --voc_stat=hifigan_csmsc_ckpt_0.1.1/feats_stats.npy \
+        --lang=zh \
+        --text=${BIN_DIR}/../../assets/sentences.txt \
+        --output_dir=${train_output_path}/test_e2e \
+        --phones_dict=dump/phone_id_map.txt \
+        --tones_dict=dump/tone_id_map.txt \
+        --inference_dir=${train_output_path}/inference \
+        --ngpu=0 \
+        --nmlu=1
+fi
+
+# wavernn
+if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
+    echo "in wavernn syn_e2e"
+    FLAGS_allocator_strategy=naive_best_fit \
+    python3 ${BIN_DIR}/../synthesize_e2e.py \
+        --am=speedyspeech_csmsc \
+        --am_config=${config_path} \
+        --am_ckpt=${train_output_path}/checkpoints/${ckpt_name} \
+        --am_stat=dump/train/feats_stats.npy \
+        --voc=wavernn_csmsc \
+        --voc_config=wavernn_csmsc_ckpt_0.2.0/default.yaml \
+        --voc_ckpt=wavernn_csmsc_ckpt_0.2.0/snapshot_iter_400000.pdz \
+        --voc_stat=wavernn_csmsc_ckpt_0.2.0/feats_stats.npy \
+        --lang=zh \
+        --text=${BIN_DIR}/../../assets/sentences.txt \
+        --output_dir=${train_output_path}/test_e2e \
+        --phones_dict=dump/phone_id_map.txt \
+        --tones_dict=dump/tone_id_map.txt \
+        --inference_dir=${train_output_path}/inference \
+        --ngpu=0 \
+        --nmlu=1
+fi
diff --git a/examples/csmsc/tts2/local/synthesize_mlu.sh b/examples/csmsc/tts2/local/synthesize_mlu.sh
new file mode 100755
index 0000000000..6c0b0b6503
--- /dev/null
+++ b/examples/csmsc/tts2/local/synthesize_mlu.sh
@@ -0,0 +1,90 @@
+#!/bin/bash
+
+config_path=$1
+train_output_path=$2
+ckpt_name=$3
+stage=0
+stop_stage=0
+
+# for more GAN Vocoders
+# multi band melgan
+if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
+    FLAGS_allocator_strategy=naive_best_fit \
+    python3 ${BIN_DIR}/../synthesize.py \
+        --am=speedyspeech_csmsc \
+        --am_config=${config_path} \
+        --am_ckpt=${train_output_path}/checkpoints/${ckpt_name} \
+        --am_stat=dump/train/feats_stats.npy \
+        --voc=mb_melgan_csmsc \
+        --voc_config=mb_melgan_csmsc_ckpt_0.1.1/default.yaml \
+        --voc_ckpt=mb_melgan_csmsc_ckpt_0.1.1/snapshot_iter_1000000.pdz\
+        --voc_stat=mb_melgan_csmsc_ckpt_0.1.1/feats_stats.npy \
+        --test_metadata=dump/test/norm/metadata.jsonl \
+        --output_dir=${train_output_path}/test \
+        --phones_dict=dump/phone_id_map.txt \
+        --tones_dict=dump/tone_id_map.txt \
+        --ngpu=0 \
+        --nmlu=1
+fi
+
+# style melgan
+if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
+    FLAGS_allocator_strategy=naive_best_fit \
+    python3 ${BIN_DIR}/../synthesize.py \
+        --am=speedyspeech_csmsc \
+        --am_config=${config_path} \
+        --am_ckpt=${train_output_path}/checkpoints/${ckpt_name} \
+        --am_stat=dump/train/feats_stats.npy \
+        --voc=style_melgan_csmsc \
+        --voc_config=style_melgan_csmsc_ckpt_0.1.1/default.yaml \
+        --voc_ckpt=style_melgan_csmsc_ckpt_0.1.1/snapshot_iter_1500000.pdz \
+        --voc_stat=style_melgan_csmsc_ckpt_0.1.1/feats_stats.npy \
+        --test_metadata=dump/test/norm/metadata.jsonl \
+        --output_dir=${train_output_path}/test \
+        --phones_dict=dump/phone_id_map.txt \
+        --tones_dict=dump/tone_id_map.txt \
+        --ngpu=0 \
+        --nmlu=1
+fi
+
+# hifigan
+if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
+    echo "in hifigan syn"
+    FLAGS_allocator_strategy=naive_best_fit \
+    python3 ${BIN_DIR}/../synthesize.py \
+        --am=speedyspeech_csmsc \
+        --am_config=${config_path} \
+        --am_ckpt=${train_output_path}/checkpoints/${ckpt_name} \
+        --am_stat=dump/train/feats_stats.npy \
+        --voc=hifigan_csmsc \
+        --voc_config=hifigan_csmsc_ckpt_0.1.1/default.yaml \
+        --voc_ckpt=hifigan_csmsc_ckpt_0.1.1/snapshot_iter_2500000.pdz \
+        --voc_stat=hifigan_csmsc_ckpt_0.1.1/feats_stats.npy \
+        --test_metadata=dump/test/norm/metadata.jsonl \
+        --output_dir=${train_output_path}/test \
+        --phones_dict=dump/phone_id_map.txt \
+        --tones_dict=dump/tone_id_map.txt \
+        --ngpu=0 \
+        --nmlu=1
+fi
+
+# wavernn
+if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
+    echo "in wavernn syn"
+    FLAGS_allocator_strategy=naive_best_fit \
+    python3 ${BIN_DIR}/../synthesize.py \
+        --am=speedyspeech_csmsc \
+        --am_config=${config_path} \
+        --am_ckpt=${train_output_path}/checkpoints/${ckpt_name} \
+        --am_stat=dump/train/feats_stats.npy \
+        --voc=wavernn_csmsc \
+        --voc_config=wavernn_csmsc_ckpt_0.2.0/default.yaml \
+        --voc_ckpt=wavernn_csmsc_ckpt_0.2.0/snapshot_iter_400000.pdz \
+        --voc_stat=wavernn_csmsc_ckpt_0.2.0/feats_stats.npy \
+        --test_metadata=dump/test/norm/metadata.jsonl \
+        --output_dir=${train_output_path}/test \
+        --tones_dict=dump/tone_id_map.txt \
+        --phones_dict=dump/phone_id_map.txt \
+        --ngpu=0 \
+        --nmlu=1
+fi
diff --git a/examples/csmsc/tts2/local/train_mlu.sh b/examples/csmsc/tts2/local/train_mlu.sh
new file mode 100755
index 0000000000..4c14864343
--- /dev/null
+++ b/examples/csmsc/tts2/local/train_mlu.sh
@@ -0,0 +1,16 @@
+
+#!/bin/bash
+
+config_path=$1
+train_output_path=$2
+# export MLU_VISIBLE_DEVICES=8
+python ${BIN_DIR}/train.py \
+    --train-metadata=dump/train/norm/metadata.jsonl \
+    --dev-metadata=dump/dev/norm/metadata.jsonl \
+    --config=${config_path} \
+    --output-dir=${train_output_path} \
+    --ngpu=0 \
+    --nmlu=2 \
+    --phones-dict=dump/phone_id_map.txt \
+    --tones-dict=dump/tone_id_map.txt \
+    --use-relative-path=True
diff --git a/examples/csmsc/tts2/run_mlu.sh b/examples/csmsc/tts2/run_mlu.sh
new file mode 100755
index 0000000000..848e540774
--- /dev/null
+++ b/examples/csmsc/tts2/run_mlu.sh
@@ -0,0 +1,76 @@
+#!/bin/bash
+
+set -e
+source path.sh
+export CUSTOM_DEVICE_BLACK_LIST=elementwise_max
+mlus=0
+stage=0
+stop_stage=100
+
+conf_path=conf/default.yaml
+train_output_path=exp/default
+ckpt_name=snapshot_iter_30600.pdz
+
+# with the following command, you can choose the stage range you want to run
+# such as `./run.sh --stage 0 --stop-stage 0`
+# this can not be mixed use with `$1`, `$2` ...
+source ${MAIN_ROOT}/utils/parse_options.sh || exit 1
+
+if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
+    # prepare data
+    ./local/preprocess.sh ${conf_path} || exit -1
+fi
+
+if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
+    # train model, all `ckpt` under `train_output_path/checkpoints/` dir
+    FLAGS_selected_mlus=${mlus} ./local/train_mlu.sh ${conf_path} ${train_output_path} || exit -1
+fi
+
+if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
+    # synthesize, vocoder is pwgan by default
+    FLAGS_selected_mlus=${mlus} ./local/synthesize_mlu.sh ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
+fi
+
+if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
+    # synthesize_e2e, vocoder is pwgan by default
+    FLAGS_selected_mlus=${mlus} ./local/synthesize_e2e_mlu.sh ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
+fi
+
+if [ ${stage} -le 4 ] && [ ${stop_stage} -ge 4 ]; then
+    # inference with static model
+    FLAGS_selected_mlus=${mlus} ./local/inference_mlu.sh ${train_output_path} || exit -1
+fi
+
+# paddle2onnx, please make sure the static models are in ${train_output_path}/inference first
+# we have only tested the following models so far
+if [ ${stage} -le 5 ] && [ ${stop_stage} -ge 5 ]; then
+    # install paddle2onnx
+    pip install paddle2onnx --upgrade
+    ./local/paddle2onnx.sh ${train_output_path} inference inference_onnx speedyspeech_csmsc
+    # considering the balance between speed and quality, we recommend that you use hifigan as vocoder
+    ./local/paddle2onnx.sh ${train_output_path} inference inference_onnx pwgan_csmsc
+    # ./local/paddle2onnx.sh ${train_output_path} inference inference_onnx mb_melgan_csmsc
+    # ./local/paddle2onnx.sh ${train_output_path} inference inference_onnx hifigan_csmsc
+fi
+
+# inference with onnxruntime
+if [ ${stage} -le 6 ] && [ ${stop_stage} -ge 6 ]; then
+    ./local/ort_predict.sh ${train_output_path}
+fi
+
+# must run after stage 3 (which stage generated static models)
+if [ ${stage} -le 7 ] && [ ${stop_stage} -ge 7 ]; then
+    ./local/export2lite.sh ${train_output_path} inference pdlite speedyspeech_csmsc x86
+    ./local/export2lite.sh ${train_output_path} inference pdlite pwgan_csmsc x86
+    # ./local/export2lite.sh ${train_output_path} inference pdlite mb_melgan_csmsc x86
+    # ./local/export2lite.sh ${train_output_path} inference pdlite hifigan_csmsc x86
+fi
+
+if [ ${stage} -le 8 ] && [ ${stop_stage} -ge 8 ]; then
+    CUDA_VISIBLE_DEVICES=${gpus} ./local/lite_predict.sh ${train_output_path} || exit -1
+fi
+
+# PTQ_static
+if [ ${stage} -le 9 ] && [ ${stop_stage} -ge 9 ]; then
+    CUDA_VISIBLE_DEVICES=${gpus} ./local/PTQ_static.sh  ${train_output_path} speedyspeech_csmsc || exit -1
+fi
diff --git a/paddlespeech/t2s/exps/inference.py b/paddlespeech/t2s/exps/inference.py
index 21d105adeb..e8ddd3befc 100644
--- a/paddlespeech/t2s/exps/inference.py
+++ b/paddlespeech/t2s/exps/inference.py
@@ -112,7 +112,7 @@ def parse_args():
     parser.add_argument(
         "--device",
         default="gpu",
-        choices=["gpu", "cpu", "xpu", "npu"],
+        choices=["gpu", "cpu", "xpu", "npu", "mlu"],
         help="Device selected for inference.", )
     parser.add_argument('--cpu_threads', type=int, default=1)
 
diff --git a/paddlespeech/t2s/exps/speedyspeech/train.py b/paddlespeech/t2s/exps/speedyspeech/train.py
index b82d688022..b1916fbc48 100644
--- a/paddlespeech/t2s/exps/speedyspeech/train.py
+++ b/paddlespeech/t2s/exps/speedyspeech/train.py
@@ -55,6 +55,8 @@ def train_sp(args, config):
         paddle.device.set_device("npu")
         if world_size > 1:
             paddle.distributed.init_parallel_env()
+    elif args.nmlu > 0:
+        paddle.device.set_device("mlu")
     else:
         paddle.set_device("cpu")
 
@@ -194,13 +196,19 @@ def main():
         "--nxpu",
         type=int,
         default=0,
-        help="if wish to use xpu, set ngpu == 0 and nxpu > 0, otherwise use gpu, npu or cpu."
+        help="if wish to use xpu, set ngpu == 0 and nxpu > 0, otherwise use gpu, npu, mlu or cpu."
     )
     parser.add_argument(
         "--nnpu",
         type=int,
         default=0,
-        help="if wish to use npu, set ngpu == 0 and nnpu > 0, otherwise use gpu, xpu or cpu."
+        help="if wish to use npu, set ngpu == 0 and nnpu > 0, otherwise use gpu, xpu, mlu or cpu."
+    )
+    parser.add_argument(
+        "--nmlu",
+        type=int,
+        default=1,
+        help="if wish to use npu, set ngpu == 0 and nmlu > 0, otherwise use gpu, xpu, npu or cpu."
     )
     parser.add_argument(
         "--ngpu",
diff --git a/paddlespeech/t2s/exps/synthesize.py b/paddlespeech/t2s/exps/synthesize.py
index 9eb459894e..b159725e26 100644
--- a/paddlespeech/t2s/exps/synthesize.py
+++ b/paddlespeech/t2s/exps/synthesize.py
@@ -222,18 +222,25 @@ def parse_args():
         "--ngpu",
         type=int,
         default=1,
-        help="if wish to use gpu, set ngpu > 0, otherwise use xpu, npu or cpu.")
+        help="if wish to use gpu, set ngpu > 0, otherwise use xpu, npu, mlu or cpu."
+    )
     parser.add_argument(
         "--nxpu",
         type=int,
         default=0,
-        help="if wish to use xpu, set ngpu == 0 and nxpu > 0, otherwise use gpu, npu or cpu."
+        help="if wish to use xpu, set ngpu == 0 and nxpu > 0, otherwise use gpu, npu, mlu or cpu."
     )
     parser.add_argument(
         "--nnpu",
         type=int,
         default=0,
-        help="if wish to use npu, set ngpu == 0 and nnpu > 0, otherwise use gpu, xpu or cpu."
+        help="if wish to use npu, set ngpu == 0 and nnpu > 0, otherwise use gpu, xpu, mlu or cpu."
+    )
+    parser.add_argument(
+        "--nmlu",
+        type=int,
+        default=0,
+        help="if wish to use xpu, set ngpu == 0 and nmlu > 0, otherwise use gpu, xpu, npu or cpu."
     )
     parser.add_argument("--test_metadata", type=str, help="test metadata.")
     parser.add_argument("--output_dir", type=str, help="output dir.")
@@ -256,10 +263,14 @@ def main():
         paddle.set_device("xpu")
     elif args.nnpu > 0:
         paddle.set_device("npu")
-    elif args.ngpu == 0 and args.nxpu == 0 and args.nnpu == 0:
+    elif args.nmlu > 0:
+        paddle.set_device("mlu")
+    elif args.ngpu == 0 and args.nxpu == 0 and args.nnpu == 0 and args.nmlu == 0:
         paddle.set_device("cpu")
     else:
-        print("ngpu, nxpu and nnpu should be >= 0")
+        print(
+            "one of ngpu, nxpu, nnpu or nmlu should be greater than 0 or all of them equal to 0"
+        )
 
     evaluate(args)
 
diff --git a/paddlespeech/t2s/exps/synthesize_e2e.py b/paddlespeech/t2s/exps/synthesize_e2e.py
index b9073124bd..08a14b3154 100644
--- a/paddlespeech/t2s/exps/synthesize_e2e.py
+++ b/paddlespeech/t2s/exps/synthesize_e2e.py
@@ -302,18 +302,25 @@ def parse_args():
         "--ngpu",
         type=int,
         default=1,
-        help="if wish to use gpu, set ngpu > 0, otherwise use xpu, npu or cpu.")
+        help="if wish to use gpu, set ngpu > 0, otherwise use xpu, npu, mlu or cpu."
+    )
     parser.add_argument(
         "--nxpu",
         type=int,
         default=0,
-        help="if wish to use xpu, set ngpu == 0 and nxpu > 0, otherwise use gpu, npu or cpu."
+        help="if wish to use xpu, set ngpu == 0 and nxpu > 0, otherwise use gpu, npu, mlu or cpu."
     )
     parser.add_argument(
         "--nnpu",
         type=int,
         default=0,
-        help="if wish to use npu, set ngpu == 0 and nnpu > 0, otherwise use gpu, xpu or cpu."
+        help="if wish to use npu, set ngpu == 0 and nnpu > 0, otherwise use gpu, xpu, mlu or cpu."
+    )
+    parser.add_argument(
+        "--nmlu",
+        type=int,
+        default=0,
+        help="if wish to use xpu, set ngpu == 0 and nmlu > 0, otherwise use gpu, xpu, npu or cpu."
     )
     parser.add_argument(
         "--text",
@@ -350,10 +357,14 @@ def main():
         paddle.set_device("xpu")
     elif args.nnpu > 0:
         paddle.set_device("npu")
-    elif args.ngpu == 0 and args.nxpu == 0 or args.nnpu == 0:
+    elif args.nmlu > 0:
+        paddle.set_device("mlu")
+    elif args.ngpu == 0 and args.nxpu == 0 and args.nnpu == 0 and args.nmlu == 0:
         paddle.set_device("cpu")
     else:
-        print("ngpu, nxpu and nnpu should be >= 0")
+        print(
+            "one of ngpu, nxpu, nnpu or nmlu should be greater than 0 or all of them equal to 0"
+        )
 
     evaluate(args)