diff --git a/examples/csmsc/tts3/local/synthesize_e2e.sh b/examples/csmsc/tts3/local/synthesize_e2e.sh
index 8130eff1487..512e062b779 100755
--- a/examples/csmsc/tts3/local/synthesize_e2e.sh
+++ b/examples/csmsc/tts3/local/synthesize_e2e.sh
@@ -109,6 +109,6 @@ if [ ${stage} -le 4 ] && [ ${stop_stage} -ge 4 ]; then
         --lang=zh \
         --text=${BIN_DIR}/../sentences.txt \
         --output_dir=${train_output_path}/test_e2e \
-        --phones_dict=dump/phone_id_map.txt #\
-        # --inference_dir=${train_output_path}/inference
+        --phones_dict=dump/phone_id_map.txt \
+        --inference_dir=${train_output_path}/inference
 fi
diff --git a/examples/csmsc/voc6/README.md b/examples/csmsc/voc6/README.md
index 26d4523d91c..7dcf133bdc8 100644
--- a/examples/csmsc/voc6/README.md
+++ b/examples/csmsc/voc6/README.md
@@ -114,6 +114,7 @@ The pretrained model can be downloaded here:
 
 The static model can be downloaded here:
 - [wavernn_csmsc_static_0.2.0.zip](https://paddlespeech.bj.bcebos.com/Parakeet/released_models/wavernn/wavernn_csmsc_static_0.2.0.zip)
+- [wavernn_csmsc_static_1.0.0.zip](https://paddlespeech.bj.bcebos.com/Parakeet/released_models/wavernn/wavernn_csmsc_static_1.0.0.zip) (fix bug for paddle 2.3)
 
 Model | Step | eval/loss
 :-------------:|:------------:| :------------:
diff --git a/paddlespeech/t2s/models/wavernn/wavernn.py b/paddlespeech/t2s/models/wavernn/wavernn.py
index b4b8b48091f..eb892eda56e 100644
--- a/paddlespeech/t2s/models/wavernn/wavernn.py
+++ b/paddlespeech/t2s/models/wavernn/wavernn.py
@@ -360,7 +360,9 @@ def generate(self,
                 x = sample.transpose([1, 0, 2])
 
             elif self.mode == 'RAW':
-                posterior = F.softmax(logits, axis=1)
+                # fix bug for paddle 2.3, see https://github.com/PaddlePaddle/Paddle/commit/01f606b4f1ca3e184a59111084ed460ee0798a5a
+                # posterior = F.softmax(logits, axis=1)
+                posterior = logits
                 distrib = paddle.distribution.Categorical(posterior)
                 # corresponding operate [np.floor((fx + 1) / 2 * mu + 0.5)] in enocde_mu_law
                 # distrib.sample([1])[0].cast('float32'): [0, 2**bits-1]