UbiquitousLearning
QNN模型初始化失败
how to do solve it? device: 8775 QNN SDK ：2.33 hexagon sdk： 5.4.2
initialize Backend Returned Status = 0 0.0ms [ ERROR ] QnnDsp <E> DspTransport.openSession qnn_open failed, 0xffffffff, prio 100
0.0ms [ ERROR ] QnnDsp <E> IDspTransport: Unknown rpc status 0x000003ff
0.0ms [ ERROR ] QnnDsp <E> DspTransport failed,cannot open session, error 0xffffffff
0.0ms [ ERROR ] QnnDsp <E> Error from rpc transport. transportStatus = -1
0.0ms [ ERROR ] QnnDsp <E> Failed to retrieve skel build id: err: 1003
0.0ms [ ERROR ] QnnDsp <E> Error in verify skel version
0.0ms [ ERROR ] QnnDsp <E> Failed to create transport for device, error: 4000
0.0ms [ ERROR ] QnnDsp <E> Failed to load skel, error: 4000
0.0ms [ ERROR ] QnnDsp <E> Transport layer setup failed: 14001
0.0ms [ ERROR ] QnnDsp <E> Failed to parse default platform info: 14001
0.0ms [ ERROR ] QnnDsp <E> Failed to load default platform info: 14001
0.0ms [ ERROR ] QnnDsp <E> Failed to parse platform config: 14001
[ERROR] Thu Jan 23 04:27:37 2025 [/home/bicv/worlspace/mllm/src/backends/qnn/QNNBackend.cpp:750] Failed to create device [ERROR] Thu Jan 23 04:27:37 2025 [/home/bicv/worlspace/mllm/src/backends/qnn/QNNBackend.cpp:590] Device Creation failure
The 8775 device seems like the chip for the automotive. And the QNN Library for automotive and mobile is different as far as I know. Can you write the full version of the QNN SDK of yours, after 2.33, like 2.33.####.
The 8775 device seems like the chip for the automotive. And the QNN Library for automotive and mobile is different as far as I know. Can you write the full version of the QNN SDK of yours, after 2.33, like 2.33.####.
QNN SDK: v2.33.0.250327.
The 8775 chip doesn't have 16GB of memory—could this be related to the issue?
The QNN initialization failed, which is not related to the runtime memory size. We only tested on 8gen3 and not sure whether it can be run on 8775. Maybe you need to check the QNN document for automotive.
