Error Code 9: Internal Error (node_of_bevpool_out: could not find any supported formats consistent with input/output data types)

log:

 ----> debug <---- call [Bevpoolv2Trt2IPluginV2DynamicExt.cpp:225][getOutputDimensions]
 ----> debug <---- call [Bevpoolv2Trt2IPluginV2DynamicExt.cpp:214][clone]
 ----> debug <---- call [Bevpoolv2Trt2IPluginV2DynamicExt.cpp:82][Bevpoolv2Trt2IPluginV2DynamicExt]
 ----> debug <---- call [Bevpoolv2Trt2IPluginV2DynamicExt.cpp:174][setPluginNamespace]
 ----> debug <---- call [Bevpoolv2Trt2IPluginV2DynamicExt.cpp:225][getOutputDimensions]
[10/09/2023-18:07:26] [V] [TRT] Applying generic optimizations to the graph for inference.
[10/09/2023-18:07:26] [V] [TRT] Original: 6 layers
[10/09/2023-18:07:26] [V] [TRT] After dead-layer removal: 6 layers
 ----> debug <---- call [Bevpoolv2Trt2IPluginV2DynamicExt.cpp:103][getPluginType]
[10/09/2023-18:07:26] [V] [TRT] After Myelin optimization: 6 layers
[10/09/2023-18:07:26] [V] [TRT] Applying ScaleNodes fusions.
[10/09/2023-18:07:26] [V] [TRT] After scale fusion: 6 layers
[10/09/2023-18:07:26] [V] [TRT] After dupe layer removal: 6 layers
[10/09/2023-18:07:26] [V] [TRT] After final dead-layer removal: 6 layers
[10/09/2023-18:07:26] [V] [TRT] After tensor merging: 6 layers
[10/09/2023-18:07:26] [V] [TRT] After vertical fusions: 6 layers
[10/09/2023-18:07:26] [V] [TRT] After dupe layer removal: 6 layers
[10/09/2023-18:07:26] [V] [TRT] After final dead-layer removal: 6 layers
[10/09/2023-18:07:26] [V] [TRT] After tensor merging: 6 layers
[10/09/2023-18:07:26] [V] [TRT] After slice removal: 6 layers
[10/09/2023-18:07:26] [V] [TRT] After concat removal: 6 layers
[10/09/2023-18:07:26] [V] [TRT] Trying to split Reshape and strided tensor
[10/09/2023-18:07:26] [V] [TRT] Graph construction and optimization completed in 0.00248367 seconds.
[10/09/2023-18:07:26] [V] [TRT] Using cublas as a tactic source
[10/09/2023-18:07:26] [I] [TRT] [MemUsageChange] Init cuBLAS/cuBLASLt: CPU +115, GPU +46, now: CPU 293, GPU 275 (MiB)
[10/09/2023-18:07:26] [V] [TRT] Using cuDNN as a tactic source
[10/09/2023-18:07:26] [I] [TRT] [MemUsageChange] Init cuDNN: CPU +95, GPU +40, now: CPU 388, GPU 315 (MiB)
[10/09/2023-18:07:26] [I] [TRT] Local timing cache in use. Profiling results in this builder pass will not be stored.
[10/09/2023-18:07:26] [V] [TRT] Constructing optimization profile number 0 [1/1].
 ----> debug <---- call [Bevpoolv2Trt2IPluginV2DynamicExt.cpp:214][clone]
 ----> debug <---- call [Bevpoolv2Trt2IPluginV2DynamicExt.cpp:82][Bevpoolv2Trt2IPluginV2DynamicExt]
 ----> debug <---- call [Bevpoolv2Trt2IPluginV2DynamicExt.cpp:174][setPluginNamespace]
 ----> debug <---- call [Bevpoolv2Trt2IPluginV2DynamicExt.cpp:244][supportsFormatCombination]
 ----> debug <---- call [Bevpoolv2Trt2IPluginV2DynamicExt.cpp:244][supportsFormatCombination]
 ----> debug <---- call [Bevpoolv2Trt2IPluginV2DynamicExt.cpp:244][supportsFormatCombination]
 ----> debug <---- call [Bevpoolv2Trt2IPluginV2DynamicExt.cpp:244][supportsFormatCombination]
 ----> debug <---- call [Bevpoolv2Trt2IPluginV2DynamicExt.cpp:244][supportsFormatCombination]
 ----> debug <---- call [Bevpoolv2Trt2IPluginV2DynamicExt.cpp:167][destroy]
 ----> debug <---- call [Bevpoolv2Trt2IPluginV2DynamicExt.cpp:167][destroy]
[10/09/2023-18:07:26] [E] Error[9]: [pluginV2Builder.cpp::reportPluginError::23] Error Code 9: Internal Error (node_of_bevpool_out: could not find any supported formats consistent with input/output data types)
[10/09/2023-18:07:26] [E] Error[2]: [builder.cpp::buildSerializedNetwork::636] Error Code 2: Internal Error (Assertion engine != nullptr failed. )
[10/09/2023-18:07:26] [E] Engine could not be created from network
[10/09/2023-18:07:26] [E] Building engine failed
[10/09/2023-18:07:26] [E] Failed to create engine from model or file.
[10/09/2023-18:07:26] [E] Engine set up failed
&&&& FAILED TensorRT.trtexec [TensorRT v8401] # /Data2/zengsheng/TensorRT-8.4.1.5/bin/trtexec --onnx=bevpoolv2.onnx --plugins=/Data1/zengsheng/works/dl_works/tensorrt_plugin_generator/bevpoolv2_plugin_codes/Bevpoolv2Trt2IPluginV2DynamicExt/libBevpoolv2Trt2IPluginV2DynamicExt.so --verbose
 ----> debug <---- call [Bevpoolv2Trt2IPluginV2DynamicExt.cpp:167][destroy]

onnx: bevpoolv2.zip

onnx

yml:

BEVPoolV2TRT2:
  attributes:
    out_height:
      datatype: int32
    out_width:
      datatype: int32
  inputs:
    tpg_input_0:
      shape: 6x59x16x44
    tpg_input_1:
      shape: 6x16x44x64
    tpg_input_2:
      shape: 179324
    tpg_input_3:
      shape: 179324
    tpg_input_4:
      shape: 179324
    tpg_input_5:
      shape: 11431
    tpg_input_6:
      shape: 11431
  outputs:
    tpg_output_0:
      shape: 1x128x128x64
  plugin_type: IPluginV2DynamicExt
  support_format_combination:
  - "float32+float32+float32+float32+float32+float32+float32+float32"

Why is this error reported, can the author help answer this question, thank you!

NVIDIA-AI-IOT / tensorrt_plugin_generator

Error Code 9: Internal Error (node_of_bevpool_out: could not find any supported formats consistent with input/output data types) #5