qwen2-audio-7B的MindIE服务化部署正常，curl调用的时候报错

一、问题现象（附报错日志上下文）：
qwen2-audio-7B的MindIE服务化部署正常 ，curl调用的时候报错：

[2025-10-19 20:14:55.432+08:00] [139] [cb370d57eb97] [system] [server] [Start mindie server] [success]
[2025-10-19 20:14:55.432+08:00] [139] [141] [server] [INFO] [endpoint.cpp:111] : [endpoint] Start endpoint success
Daemon start success!
[2025-10-19 20:14:56.620+08:00] [139] [cb370d57eb97] [system] [server] [start endpoint] [success]
[2025-10-19 20:14:56.621+08:00] [139] [141] [server] [INFO] [llm_daemon.cpp:165] : [daemon] Daemon start success!
[2025-10-20 10:18:56.728+08:00] [139] [cb370d57eb97] [127.0.0.1:59962] [server] [Request mindie server, method is POST, uri is /v1/chat/completions] [success]
[2025-10-20 10:18:56.728+08:00] [139] [1211] [server] [INFO] [http_server.cpp:54] : [endpoint] Receive request from 127.0.0.1:59962 ,method is POST, uri is /v1/chat/completions
[2025-10-20 10:18:56.729+08:00] [139] [1211] [server] [ERROR] [common_wrapper.cpp:50] : [MIE04E02040E] [endpoint] Convert string to json object exception, cbId is 1
[2025-10-20 10:18:56.729+08:00] [139] [1211] [server] [ERROR] [common_infer.cpp:444] : [MIE04E02040E] [endpoint] Failed to parse request context to json. The logId is endpoint_common_0
[2025-10-20 10:18:56.729+08:00] [139] [1211] [server] [ERROR] [common_infer.cpp:328] : [MIE04E02040E] [endpoint] Failed to parse context to json body. The logId is endpoint_common_0
[2025-10-20 10:18:56.729+08:00] [139] [cb370d57eb97] [127.0.0.1:18081] [server] [Response mindie server, method is POST, uri is /v1/chat/completions] [status code: 422]
[2025-10-20 10:18:56.729+08:00] [139] [1211] [server] [INFO] [http_server.cpp:68] : [endpoint] Http receive response from 127.0.0.1:18081, status code is 422, body len 89
[2025-10-20 10:19:47.798+08:00] [139] [cb370d57eb97] [127.0.0.1:51290] [server] [Request mindie server, method is POST, uri is /v1/chat/completions] [success]
[2025-10-20 10:19:47.798+08:00] [139] [1214] [server] [INFO] [http_server.cpp:54] : [endpoint] Receive request from 127.0.0.1:51290 ,method is POST, uri is /v1/chat/completions
[2025-10-20 10:19:48.172] [139] [281472934276736] [llm] [INFO] [llm_manager_impl.cpp:956] Get a new inferRequest in ProcessRequests, request id is endpoint_common_1
[2025-10-20 10:19:48.173] [139] [281472934276736] [llm] [INFO] [llm_manager_impl.cpp:1018] Insert a new inferRequest, string id is endpoint_common_1, int id is 0, ibis id is 0
[2025-10-20 10:19:48.173+08:00] [139] [1171] [batchscheduler] [INFO] [ibis_request.cpp:103] : [batch_scheduler] Create IBIS request: 0 successfully with batchSize: , isSync: 0, maxOutputLen: 32768, priority: 5
[2025-10-20 10:19:48.173+08:00] [139] [1214] [server] [INFO] [llm_infer_model_instance.cpp:523] : [llm_backend] LLM backend enqueue request finished, request id: endpoint_common_1
[2025-10-20 10:19:48.173+08:00] [139] [1171] [batchscheduler] [INFO] [ibis_request.cpp:69] : [batch_scheduler] Create IBIS request: 0 successfully with batchSize: , isSync: 0, maxOutputLen: 32768, priority: 5
[2025-10-20 10:19:48.173+08:00] [139] [1171] [batchscheduler] [INFO] [ibis_config.cpp:183] : [batch_scheduler] Init Block Manager with dpsize1
[2025-10-20 10:19:48.173+08:00] [139] [1171] [batchscheduler] [INFO] [ibis_batch_scheduler.cpp:217] : [batch_scheduler] IBIS Pipeline [0] of WorkFlow Standard_LLM: enqueue sequence: 0
[2025-10-20 10:19:48.178+08:00] [139] [1170] [batchscheduler] [INFO] [ibis_model_instance_block_manager_v2.cpp:62] : [batch_scheduler] [IBISModelInstanceBlockManagerV2::CanAllocate] KVCacheLog  RequestID: 0 CanAllocate { requiredBlockNum: 4, freeNpuBlockNum: 475 }
[2025-10-20 10:19:48.178+08:00] [139] [1170] [batchscheduler] [INFO] [ibis_model_instance_block_manager_v2.cpp:88] : [batch_scheduler] [IBISModelInstanceBlockManagerV2::Allocate] KVCacheLog  Request:0 SeqId:100 Allocate. New Block Table:blockTable(reqId=100, type=DEVICE){0, 1, 2, 3}
[2025-10-20 10:19:48.178+08:00] [405] [411] [batchscheduler] [INFO] [slave_IPC_communicator.cpp:524] : [model_backend] slave receive msg len: 6837
[2025-10-20 10:20:00.938+08:00] [139] [1163] [batchscheduler] [ERROR] [master_IPC_communicator.cpp:264] : [MIE04E120407] [model_backend] The req batchSize from ibis is 1, but receive batchSize from sub-process is 0
[2025-10-20 10:20:00.938] [139] [281470749830784] [llm] [ERROR] [llm_infer_engine_ibis_api.cpp:458] The model instance executed requests failed.
[2025-10-20 10:20:00.938+08:00] [139] [1163] [server] [ERROR] [base_wrapper.cpp:46] : [MIE04E020414] [endpoint] Failed to get ibis seqs id by output 'IBIS_SEQS_ID' not found in response
[2025-10-20 10:20:00.938+08:00] [139] [1163] [batchscheduler] [ERROR] [master_IPC_communicator.cpp:638] : [MIE04E120407] [model_backend] Failed to deSerialize execute response
[2025-10-20 10:20:00.938+08:00] [139] [1163] [batchscheduler] [ERROR] [ibis_model_instance.cpp:77] : [MIE04E110408] [batch_scheduler] GMIS executes modelInstance failed, ExecuteType: 1
[2025-10-20 10:20:00.938+08:00] [139] [1163] [server] [ERROR] [base_wrapper.cpp:319] : [MIE04E020414] [endpoint] Failed to generate result for parse seq id error.
[2025-10-20 10:20:00.939+08:00] [139] [1214] [server] [ERROR] [base_wrapper.cpp:46] : [MIE04E020414] [endpoint] Failed to get ibis seqs id by output 'IBIS_SEQS_ID' not found in response
[2025-10-20 10:20:00.938+08:00] [139] [1163] [batchscheduler] [ERROR] [model_backend.cpp:784] : [MIE04E120414] [model_backend] Failed to get execute response message from local progress.
[2025-10-20 10:20:00.939+08:00] [139] [1214] [server] [ERROR] [base_wrapper.cpp:319] : [MIE04E020414] [endpoint] Failed to generate result for parse seq id error.
[2025-10-20 10:20:00.939+08:00] [139] [1214] [server] [ERROR] [common_wrapper.cpp:321] : [MIE04E020414] [endpoint] Failed to get engine response
[2025-10-20 10:20:00.939+08:00] [139] [1163] [batchscheduler] [ERROR] [gmis_agent_base_plugin.cpp:51] : [MIE04E110E07] [batch_scheduler] Model instance executes failed, executeType: 1
[2025-10-20 10:20:00.939+08:00] [139] [cb370d57eb97] [127.0.0.1:18081] [server] [Response mindie server, method is POST, uri is /v1/chat/completions] [status code: 500]
[2025-10-20 10:20:00.939+08:00] [139] [1163] [batchscheduler] [INFO] [ibis_request.h:37] : [batch_scheduler] Update AsyncRequest 0 to SECOND_END
[2025-10-20 10:20:00.939+08:00] [139] [1214] [server] [INFO] [http_server.cpp:68] : [endpoint] Http receive response from 127.0.0.1:18081, status code is 500, body len 79
[2025-10-20 10:20:00.939+08:00] [139] [1163] [batchscheduler] [INFO] [gmis_agent_base_plugin.cpp:162] : [batch_scheduler] Request end, reqId: 0, endState: 2
[2025-10-20 10:20:00.939+08:00] [139] [1169] [batchscheduler] [INFO] [ibis_request_manager.cpp:175] : [batch_scheduler] IBIS requestManager release request: 0
[2025-10-20 10:20:03.001+08:00] [139] [1169] [server] [WARN] [llm_daemon.cpp:74] : [MIE04W01011A] [daemon] Received exit signal[17]
[mindie] run mindie service ok

![输入图片说明](https://foruda.gitee.com/images/1760943298800423800/d5b6d7d3_14847670.png "屏幕截图")

二、测试步骤：
接口调用

curl http://localhost:18081/v1/chat/completions \

-H "Content-Type: application/json" \

-d '{

"model": "qwen-audio-instruct",

"messages": [   {

"role": "user",

"content": [

{"type": "text", "text": "请分析下这段音频的内容"},

{"type": "audio_url", "audio_url": "file:///home/HwHiAiUser/model/guess_age_gender.wav"}

]

}]

}'

Ascend/ModelZoo-PyTorch

内容风险标识

评论 (1)

Ascend/ModelZoo-PyTorch .gitee-modal { width: 500px !important; }

内容风险标识