X3588. fastdeploy 编译OK，运行是提示 failed to open rknpu module

1. <blockquote><blockquote>No LSB modules are available.

复制代码

RK3588/test_demo/pp_yolo/build# make -j8                     

Scanning dependencies of target infer_ppyoloe_demo

Scanning dependencies of target infer_yolov8_demo

[ 50%] Building CXX object CMakeFiles/infer_yolov8_demo.dir/infer_yolov8.cc.o

[ 50%] Building CXX object CMakeFiles/infer_ppyoloe_demo.dir/infer_ppyoloe.cc.o

[ 75%] Linking CXX executable infer_yolov8_demo

[100%] Linking CXX executable infer_ppyoloe_demo

[100%] Built target infer_yolov8_demo

[100%] Built target infer_ppyoloe_demo

编译正常
但是运行时 错误

1. /yolov8_n_500e_coco/000000014439.jpg 1                        
   
2. [INFO] fastdeploy/vision/common/processors/transform.cc(45)::FuseNormalizeCast        Normalize and Cast are fused to Normalize in preprocessing pipeline.
   
3. [INFO] fastdeploy/vision/common/processors/transform.cc(93)::FuseNormalizeHWC2CHW        Normalize and HWC2CHW are fused to NormalizeAndPermute  in preprocessing pipeline.
   
4. [INFO] fastdeploy/vision/common/processors/transform.cc(159)::FuseNormalizeColorConvert        BGR2RGB and NormalizeAndPermute are fused to NormalizeAndPermute with swap_rb=1
   
5. E RKNN: [10:32:16.099] failed to open rknpu module, need to insmod rknpu dirver!
   
6. E RKNN: [10:32:16.099] failed to open rknn device!
   
7. [ERROR] fastdeploy/runtime/backends/rknpu2/rknpu2_backend.cc(146)::LoadModel        rknn_init fail! ret=-1
   
8. [ERROR] fastdeploy/runtime/backends/rknpu2/rknpu2_backend.cc(89)::InitFromRKNN        load model failed
   
9. [ERROR] fastdeploy/runtime/runtime.cc(432)::CreateRKNPU2Backend        Load model from nb file failed while initializing LiteBackend.
   
10. Aborted
    

复制代码

重新更换 固件OK  

1. [INFO] fastdeploy/vision/common/processors/transform.cc(45)::FuseNormalizeCast        Normalize and Cast are fused to Normalize in preprocessing pipeline.
   
2. [INFO] fastdeploy/vision/common/processors/transform.cc(93)::FuseNormalizeHWC2CHW        Normalize and HWC2CHW are fused to NormalizeAndPermute  in preprocessing pipeline.
   
3. [INFO] fastdeploy/vision/common/processors/transform.cc(159)::FuseNormalizeColorConvert        BGR2RGB and NormalizeAndPermute are fused to NormalizeAndPermute with swap_rb=1
   
4. [INFO] fastdeploy/runtime/backends/rknpu2/rknpu2_backend.cc(57)::GetSDKAndDeviceVersion        rknn_api/rknnrt version: 1.4.1b22 (e8b31839e@2023-02-02T08:52:53), driver version: 0.7.2
   
5. index=0, name=image, n_dims=4, dims=[1, 640, 640, 3], n_elems=1228800, size=2457600, fmt=NHWC, type=FP16, qnt_type=AFFINE, zp=0, scale=1.000000, pass_through=0
   
6. index=0, name=p2o.Mul.117, n_dims=4, dims=[1, 8400, 4, 1], n_elems=33600, size=67200, fmt=NCHW, type=FP16, qnt_type=AFFINE, zp=0, scale=1.000000, pass_through=0
   
7. index=1, name=p2o.Concat.43, n_dims=4, dims=[1, 80, 8400, 1], n_elems=672000, size=1344000, fmt=NCHW, type=FP16, qnt_type=AFFINE, zp=0, scale=1.000000, pass_through=0
   
8. [INFO] fastdeploy/runtime/runtime.cc(438)::CreateRKNPU2Backend        Runtime initialized with Backend::RKNPU2 in Device::RKNPU.
   
9. [INFO] fastdeploy/vision/common/processors/transform.cc(159)::FuseNormalizeColorConvert        BGR2RGB and Normalize are fused to Normalize with swap_rb=1
   
10. [WARNING] fastdeploy/runtime/backends/rknpu2/rknpu2_backend.cc(315)::Infer        The input tensor type != model's inputs type.The input_type need FP16,but inputs[0].type is UINT8
    
11. [FastDeploy] PPDet in RKNPU2 duration = 0.127267s.
    
12. DetectionResult: [xmin, ymin, xmax, ymax, score, label_id]
    
13. 504.750000,112.812500, 597.000000, 277.000000, 0.882324, 0
    
14. 414.250000,89.187500, 506.000000, 286.000000, 0.828613, 0
    
15. 104.937500,45.718750, 127.250000, 93.375000, 0.827148, 0
    
16. 162.500000,82.812500, 199.250000, 167.125000, 0.763184, 0
    
17. 268.000000,80.875000, 299.250000, 169.625000, 0.754883, 0
    
18. 328.500000,39.937500, 346.500000, 80.437500, 0.707520, 0
    
19. 364.500000,61.000000, 382.250000, 114.125000, 0.619629, 0
    
20. 380.000000,41.000000, 395.000000, 84.250000, 0.612305, 0
    
21. 352.750000,44.843750, 368.250000, 98.000000, 0.547852, 0
    
22. 581.500000,112.875000, 613.000000, 199.125000, 0.486816, 0

复制代码