[INFO] exception_cb streamId:14 taskId:19 deviceId: 0 opName:/sample_func/Pad inputCnt:3 outputCnt:1
[ERROR] open file failed
[ERROR] save tensor memory to file: open file failed.
[WARN] exception_cb input_0 save failed
[ERROR] open file failed
[ERROR] save tensor memory to file: open file failed.
[WARN] exception_cb input_0 save failed
[ACL ERROR] EZ9999: Inner Error!
EZ9999: 2025-03-13-07:53:45.581.278 The error from device(0), serial number is 8, there is an aicore error, core id is 0, error code = 0, dump info: pc start: 0x1000124083ae6000, current: 0x124083ae61a4, vec error info: 0x1dedf57d, mte error info: 0x5d, ifu error info: 0x3f54772f5f100, ccu error info: 0, cube error info: 0xf4, biu error info: 0, aic error mask: 0x65000200d000288, para base: 0x124083ac00a8, errorStr: time out or trap error.[FUNC:PrintCoreErrorInfo][FILE:device_error_proc.cc][LINE:537]
TraceBack (most recent call last):
测试代码：
from ais_bench.infer.interface import InferSession
import numpy as np
if __name__ == '__main__':
step2_session = InferSession(device_id=0, model_path="xxxx.om"))
intensors_desc = step2_session.get_inputs()
for i, info in enumerate(intensors_desc):
print(
f"input info i:{i} shape:{info.shape} type:{info.datatype} val: \
{int(info.datatype)} realsize:{info.realsize} size:{info.size}"
)
outputs = step2_session.infer([np.zeros((1, 2, 1024), dtype=np.float32),
np.zeros((1, 3,512,512), dtype=np.float32),
np.zeros((1, 224), dtype=np.float32),
np.zeros((1, 107), dtype=np.float32),
np.zeros((1, 3, 3), dtype=np.float32),
np.zeros((1, 3, 1), dtype=np.float32),
np.zeros((1, 3, 3), dtype=np.float32),
])
print("推理结果形状:", [out.shape for out in outputs])
点赞
收藏
回复
分享
举报
已解决
以下完全同样的代码和步骤，在启智平台上的云脑任务，使用910A卡可以正常运行，但是另一台物理机上失败。
步骤：
环境：cann 8.0.RC1
1. 使用atc将onnx模型转换成om模型：成功
2. 查看om模型的输入输出结构：成功
3. 使用from ais_bench.infer.interface import InferSession进行推理，命令行卡死，过几分钟后报错，报错内容如下：
[INFO] exception_cb streamId:14 taskId:19 deviceId: 0 opName:/sample_func/Pad inputCnt:3 outputCnt:1 [ERROR] open file failed [ERROR] save tensor memory to file: open file failed. [WARN] exception_cb input_0 save failed [ERROR] open file failed [ERROR] save tensor memory to file: open file failed. [WARN] exception_cb input_0 save failed [ACL ERROR] EZ9999: Inner Error! EZ9999: 2025-03-13-07:53:45.581.278 The error from device(0), serial number is 8, there is an aicore error, core id is 0, error code = 0, dump info: pc start: 0x1000124083ae6000, current: 0x124083ae61a4, vec error info: 0x1dedf57d, mte error info: 0x5d, ifu error info: 0x3f54772f5f100, ccu error info: 0, cube error info: 0xf4, biu error info: 0, aic error mask: 0x65000200d000288, para base: 0x124083ac00a8, errorStr: time out or trap error.[FUNC:PrintCoreErrorInfo][FILE:device_error_proc.cc][LINE:537] TraceBack (most recent call last):测试代码：
from ais_bench.infer.interface import InferSession import numpy as np if __name__ == '__main__': step2_session = InferSession(device_id=0, model_path="xxxx.om")) intensors_desc = step2_session.get_inputs() for i, info in enumerate(intensors_desc): print( f"input info i:{i} shape:{info.shape} type:{info.datatype} val: \ {int(info.datatype)} realsize:{info.realsize} size:{info.size}" ) outputs = step2_session.infer([np.zeros((1, 2, 1024), dtype=np.float32), np.zeros((1, 3,512,512), dtype=np.float32), np.zeros((1, 224), dtype=np.float32), np.zeros((1, 107), dtype=np.float32), np.zeros((1, 3, 3), dtype=np.float32), np.zeros((1, 3, 1), dtype=np.float32), np.zeros((1, 3, 3), dtype=np.float32), ]) print("推理结果形状:", [out.shape for out in outputs])