一、问题现象：
在310P1上转换SuperGlue模型正常，使用ais_banch推理时报错
二、软件版本:
-- CANN 版本：8.0.RC2:
--Python 3.7:
--操作系统版本：openEuler 22.03:
--型号：Ascend310P1
三、测试步骤：
按https://gitee.com/ascend/ModelZoo-PyTorch/tree/master/ACL_PyTorch/built-in/cv/SuperGlue_with_SuperPoint_for_Pytorch 中步骤与参数转换模型成功，由于使用的芯片是310P1，将以下命令中的soc_version修改为Ascend310P1：
模型转换正常：
根据SuperGlue_with_SuperPoint_for_Pytorch提供的推理脚本infer.py，对图像进行匹配推理，Superpoint正常，经测试其推理得到的关键点和描述子精度正常，但是在对Superglue模型进行推理时报错：
四、日志信息:
Inference of SuperGlue:
[INFO] create new context
[WARN] exception_cb deviceId:0 streamId:14 taskId:1 failed:500002
[ACL ERROR] EZ9999: Inner Error!
EZ9999: 2025-07-03-08:52:48.978.924 Kernel task happen error, retCode=0x2a, [aicpu exception].[FUNC:PreCheckTaskErr][FILE:davinic_kernel_task.cc][LINE:1220]
TraceBack (most recent call last):
Aicpu kernel execute failed, device_id=0, stream_id=14, task_id=1, errorCode=2a.[FUNC:PrintAicpuErrorInfo][FILE:davinic_kernel_task.cc][LINE:1023]
get opdesc info failed, device_id:0, stream_id:14, task_id:1.[FUNC:GetOpDescInfo][FILE:ge_executor.cc][LINE:1310]
[Get][OpDescInfo]get op desc faild, ge result[-1], deviceId[0], streamId[14], taskId[1][FUNC:ReportCallError][FILE:log_inner.cpp][LINE:161]
Aicpu kernel execute failed, device_id=0, stream_id=14, task_id=1, flip_num=0, fault so_name=, fault kernel_name=, fault op_name=, extend_info=.[FUNC:GetError][FILE:stream.cc][LINE:1082]
rtStreamSynchronize execute failed, reason=[aicpu exception][FUNC:FuncErrorReason][FILE:error_message_manage.cc][LINE:53]
Call rtStreamSynchronize(stream) fail, ret: 0x7BC8A[FUNC:KernelLaunchEx][FILE:model_manager.cc][LINE:171]
Failed to execute init graph[FUNC:Load][FILE:model_v2_executor.cc][LINE:171]
[Model][FromData]call load executorV2 failed, ge result[1343225857][FUNC:ReportCallError][FILE:log_inner.cpp][LINE:161]
DEVICE[0] PID[37723]:
EXCEPTION STREAM:
Exception info:TGID=37723, model id=65535, stream id=14, stream phase=3
Message info[0]:RTS_HWTS: Aicpu exception, slot_id=21, stream_id=14
Other info[0]:time=2025-07-03-08:52:48.935.550, function=process_hwts_error_exception, line=646, error code=0x2a
[ERROR] load model from file failed, model file is superglue.om
[WARN] Check failed:processModel->LoadModelFromFile(modelPath), ret:1
[WARN] no model had been loaded, unload failed
Traceback (most recent call last):
File "infer.py", line 229, in <module>
device_id=option.device_id,
File "infer.py", line 177, in infer_dymShape
session = InferSession(device_id, om_path)
File "/root/miniconda3/envs/ais_bench/lib/python3.7/site-packages/ais_bench/infer/interface.py", line 82, in __init__
self.session = aclruntime.InferenceSession(self.model_path, self.device_id, self.options)
RuntimeError: [1][ACL: invalid parameter]
点赞
收藏
回复
分享
举报
已解决
一、问题现象：
在310P1上转换SuperGlue模型正常，使用ais_banch推理时报错
二、软件版本:
-- CANN 版本：8.0.RC2:
--Python 3.7:
--操作系统版本：openEuler 22.03:
--型号：Ascend310P1
三、测试步骤：
按https://gitee.com/ascend/ModelZoo-PyTorch/tree/master/ACL_PyTorch/built-in/cv/SuperGlue_with_SuperPoint_for_Pytorch 中步骤与参数转换模型成功，由于使用的芯片是310P1，将以下命令中的soc_version修改为Ascend310P1：
模型转换正常：
根据SuperGlue_with_SuperPoint_for_Pytorch提供的推理脚本infer.py，对图像进行匹配推理，Superpoint正常，经测试其推理得到的关键点和描述子精度正常，但是在对Superglue模型进行推理时报错：
四、日志信息:
Inference of SuperGlue:
[INFO] create new context
[WARN] exception_cb deviceId:0 streamId:14 taskId:1 failed:500002
[ACL ERROR] EZ9999: Inner Error!
EZ9999: 2025-07-03-08:52:48.978.924 Kernel task happen error, retCode=0x2a, [aicpu exception].[FUNC:PreCheckTaskErr][FILE:davinic_kernel_task.cc][LINE:1220]
TraceBack (most recent call last):
Aicpu kernel execute failed, device_id=0, stream_id=14, task_id=1, errorCode=2a.[FUNC:PrintAicpuErrorInfo][FILE:davinic_kernel_task.cc][LINE:1023]
get opdesc info failed, device_id:0, stream_id:14, task_id:1.[FUNC:GetOpDescInfo][FILE:ge_executor.cc][LINE:1310]
[Get][OpDescInfo]get op desc faild, ge result[-1], deviceId[0], streamId[14], taskId[1][FUNC:ReportCallError][FILE:log_inner.cpp][LINE:161]
Aicpu kernel execute failed, device_id=0, stream_id=14, task_id=1, flip_num=0, fault so_name=, fault kernel_name=, fault op_name=, extend_info=.[FUNC:GetError][FILE:stream.cc][LINE:1082]
rtStreamSynchronize execute failed, reason=[aicpu exception][FUNC:FuncErrorReason][FILE:error_message_manage.cc][LINE:53]
Call rtStreamSynchronize(stream) fail, ret: 0x7BC8A[FUNC:KernelLaunchEx][FILE:model_manager.cc][LINE:171]
Failed to execute init graph[FUNC:Load][FILE:model_v2_executor.cc][LINE:171]
[Model][FromData]call load executorV2 failed, ge result[1343225857][FUNC:ReportCallError][FILE:log_inner.cpp][LINE:161]
DEVICE[0] PID[37723]:
EXCEPTION STREAM:
Exception info:TGID=37723, model id=65535, stream id=14, stream phase=3
Message info[0]:RTS_HWTS: Aicpu exception, slot_id=21, stream_id=14
Other info[0]:time=2025-07-03-08:52:48.935.550, function=process_hwts_error_exception, line=646, error code=0x2a
[ERROR] load model from file failed, model file is superglue.om
[WARN] Check failed:processModel->LoadModelFromFile(modelPath), ret:1
[WARN] no model had been loaded, unload failed
Traceback (most recent call last):
File "infer.py", line 229, in <module>
device_id=option.device_id,
File "infer.py", line 177, in infer_dymShape
session = InferSession(device_id, om_path)
File "/root/miniconda3/envs/ais_bench/lib/python3.7/site-packages/ais_bench/infer/interface.py", line 82, in __init__
self.session = aclruntime.InferenceSession(self.model_path, self.device_id, self.options)
RuntimeError: [1][ACL: invalid parameter]