npu上适配，请问这个语言识别模型要怎么用

[Yanceye]

ERROR - ❌ Error loading models: capture_end:build/CMakeFiles/torch_npu.dir/compiler_depend.ts:171 NPU function error: c10_npu::acl::AclmdlRICaptureEnd(capture_stream_, &model_ri), error code is 107025
1120 [ERROR] 2025-11-20-20:37:06 (PID:374, Device:0, RankID:-1) ERR00100 PTA call acl api failed.
1121 EH9999: Inner Error!
1122 rtStreamEndCapture execute failed, reason=[capture model contains a stream that was not joined to the original stream][FUNC:FuncErrorReason][FILE:error_message_manage.cc][LINE:53]
1123 EH9999: [PID: 374] 2025-11-20-20:37:06.119.103 end capture stream failed, runtime result = 107025[FUNC:ReportCallError][FILE:log_inner.cpp][LINE:161]
1124 TraceBack (most recent call last):
1125
1126 2025-11-20 20:37:06,120 - main - ERROR - Please check your model paths and source configuration.
1127 ckpt: ./Step-Audio-Tokenizer/dengcunqin/speech_paraformer-large_asr_nat-zh-cantonese-en-16k-vocab8501-online/model.pt

在npu上跑了一下，发现导入这个语言识别模型会出错，可能要转onnx

用funasr命令转了一下
funasr-export ++model=./dengcunqin/speech_paraformer-large_asr_nat-zh-cantonese-en-16k-vocab8501-online ++quantize=false
我用这个命令转了onnx,，它变成两个文件了model.onnx和decoder.onnx，这两个要怎么使用？