diff --git a/README.md b/README.md
index 4f03f574..f9ae1ead 100644
--- a/README.md
+++ b/README.md
@@ -178,6 +178,7 @@ Via the easy-to-use, efficient, flexible and scalable implementation, our vision
   - 🧩  *Cascaded models application*: as an extension of the typical traditional audio tasks, we combine the workflows of the aforementioned tasks with other fields like Natural language processing (NLP) and Computer Vision (CV).
 
 ### Recent Update
+- ⚡ 2023.04.28: Fix [0-d tensor](https://github.com/PaddlePaddle/PaddleSpeech/pull/3214), with the upgrade of paddlepaddle==2.5, the problem of modifying 0-d tensor has been solved.
 - 👑 2023.04.25: Add [AMP for U2 conformer](https://github.com/PaddlePaddle/PaddleSpeech/pull/3167).
 - 🔥 2023.04.06: Add [subtitle file (.srt format) generation example](./demos/streaming_asr_server).
 - 🔥 2023.03.14: Add SVS(Singing Voice Synthesis) examples with Opencpop dataset, including [DiffSinger](./examples/opencpop/svs1)、[PWGAN](./examples/opencpop/voc1) and [HiFiGAN](./examples/opencpop/voc5), the effect is continuously optimized.
diff --git a/README_cn.md b/README_cn.md
index e5e18f0c..25a716d2 100644
--- a/README_cn.md
+++ b/README_cn.md
@@ -183,6 +183,7 @@
   - 🧩 级联模型应用: 作为传统语音任务的扩展，我们结合了自然语言处理、计算机视觉等任务，实现更接近实际需求的产业级应用。
 
 ### 近期更新
+- ⚡ 2023.04.28: 修正 [0-d tensor](https://github.com/PaddlePaddle/PaddleSpeech/pull/3214), 配合PaddlePaddle2.5升级修改了0-d tensor的问题。
 - 👑 2023.04.25: 新增 [U2 conformer 的 AMP 训练](https://github.com/PaddlePaddle/PaddleSpeech/pull/3167).
 - 👑 2023.04.06: 新增 [srt格式字幕生成功能](./demos/streaming_asr_server)。
 - 🔥 2023.03.14: 新增基于 Opencpop 数据集的 SVS (歌唱合成) 示例，包含 [DiffSinger](./examples/opencpop/svs1)、[PWGAN](./examples/opencpop/voc1) 和 [HiFiGAN](./examples/opencpop/voc5)，效果持续优化中。
diff --git a/paddlespeech/cli/asr/infer.py b/paddlespeech/cli/asr/infer.py
index 525eb9cb..231a00f4 100644
--- a/paddlespeech/cli/asr/infer.py
+++ b/paddlespeech/cli/asr/infer.py
@@ -274,7 +274,7 @@ class ASRExecutor(BaseExecutor):
             # fbank
             audio = preprocessing(audio, **preprocess_args)
 
-            audio_len = paddle.to_tensor(audio.shape[0])
+            audio_len = paddle.to_tensor(audio.shape[0]).unsqueeze(axis=0)
             audio = paddle.to_tensor(audio, dtype='float32').unsqueeze(axis=0)
 
             self._inputs["audio"] = audio
diff --git a/paddlespeech/cli/ssl/infer.py b/paddlespeech/cli/ssl/infer.py
index 49fdbf39..bc3c632d 100644
--- a/paddlespeech/cli/ssl/infer.py
+++ b/paddlespeech/cli/ssl/infer.py
@@ -247,7 +247,7 @@ class SSLExecutor(BaseExecutor):
         # fbank
         audio = preprocessing(audio, **preprocess_args)
 
-        audio_len = paddle.to_tensor(audio.shape[0])
+        audio_len = paddle.to_tensor(audio.shape[0]).unsqueeze(axis=0)
         audio = paddle.to_tensor(audio, dtype='float32').unsqueeze(axis=0)
 
         self._inputs["audio"] = audio
diff --git a/paddlespeech/cli/whisper/infer.py b/paddlespeech/cli/whisper/infer.py
index ebcca890..17e8c0b8 100644
--- a/paddlespeech/cli/whisper/infer.py
+++ b/paddlespeech/cli/whisper/infer.py
@@ -253,7 +253,7 @@ class WhisperExecutor(BaseExecutor):
         # fbank
         audio = log_mel_spectrogram(audio, resource_path=self.resource_path)
 
-        audio_len = paddle.to_tensor(audio.shape[0])
+        audio_len = paddle.to_tensor(audio.shape[0]).unsqueeze(axis=0)
 
         self._inputs["audio"] = audio
         self._inputs["audio_len"] = audio_len