c++的GPU版本 无法识别 寻寻觅觅，冷冷清清，凄凄惨惨戚戚

[zane5566]

Notice: In order to resolve issues more efficiently, please raise issue following the template.
（注意：为了更加高效率解决您遇到的问题，请按照模板提问，补充细节）

🐛 Bug

To Reproduce

Steps to reproduce the behavior (always include the command you ran):

1. 使用镜像registry.cn-hangzhou.aliyuncs.com/funasr_repo/funasr:funasr-runtime-sdk-gpu-0.1.1部署的服务
2. iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch 此模型导出torchscript
3. 读这段话：寻寻觅觅，冷冷清清，凄凄惨惨戚戚，无法识别。
4. 用python版本可以正常识别

Code sample

Expected behavior

Environment

• OS (e.g., Linux):
• FunASR Version (e.g., 1.0.28):
• ModelScope Version (e.g., 1.11.0):
• PyTorch Version (e.g., 2.0.0):
• How you installed funasr (pip, source):
• Python version:
• GPU (e.g., V100M32)
• CUDA/cuDNN version (e.g., cuda11.7):
• Docker version (e.g., funasr-runtime-sdk-cpu-0.4.1)
• Any other relevant information:

Additional context

使用镜像registry.cn-hangzhou.aliyuncs.com/funasr_repo/funasr:funasr-runtime-sdk-gpu-0.1.1部署的服务