Pytochモデル変換、pth->onnx->trt(TensorRT engine)

pth->onnx    


##           
   torch.onnx  onnx   ，                     。


**    ：**
  ，           torch.onnx         。  ：
heads {'hm': 1, 'wh': 2, 'hps': 34, 'reg': 2, 'hm_hp': 17, 'hp_offset': 2}
Centerpose       onnx     ，           ，            。


##tensor.size(dim)        onnx     Gather[0]
          tensor.size(dim)  tensor.shape[1]      ，      onnx      Gather[0]  ，     ， TensorRT      。


**    ：**
         ，               tensor，                 。


##       onnx    
         ‘%’       ，        ATen   ，    torch/onnx/symbolic_opset9.py   。
/usr/local/lib/python3.6/dist-packages/torch/onnx/utils.py:562: UserWarning: ONNX export failed on ATen operator remainder because torch.onnx.symbolic_opset9.remainder does not exist
  .format(op_name, opset_version, op_name))
Traceback (most recent call last):
  File "transfer_centernet_dlav0_to_onnx2.py", line 65, in 
    onnx_module = torch.onnx.export(model,img,'multipose_dlav0_1x_modellast_process.onnx',verbose=True)
  File "/usr/local/lib/python3.6/dist-packages/torch/onnx/__init__.py", line 132, in export
    strip_doc_string, dynamic_axes)
  File "/usr/local/lib/python3.6/dist-packages/torch/onnx/utils.py", line 64, in export
    example_outputs=example_outputs, strip_doc_string=strip_doc_string, dynamic_axes=dynamic_axes)
  File "/usr/local/lib/python3.6/dist-packages/torch/onnx/utils.py", line 329, in _export
    _retain_param_name, do_constant_folding)
  File "/usr/local/lib/python3.6/dist-packages/torch/onnx/utils.py", line 225, in _model_to_graph
    _disable_torch_constant_prop=_disable_torch_constant_prop)
  File "/usr/local/lib/python3.6/dist-packages/torch/onnx/utils.py", line 127, in _optimize_graph
    graph = torch._C._jit_pass_onnx(graph, operator_export_type)
  File "/usr/local/lib/python3.6/dist-packages/torch/onnx/__init__.py", line 163, in _run_symbolic_function
    return utils._run_symbolic_function(*args, **kwargs)
  File "/usr/local/lib/python3.6/dist-packages/torch/onnx/utils.py", line 563, in _run_symbolic_function
    op_fn = sym_registry.get_registered_op(op_name, '', opset_version)
  File "/usr/local/lib/python3.6/dist-packages/torch/onnx/symbolic_registry.py", line 91, in get_registered_op
    return _registry[(domain, version)][opname]
KeyError: 'remainder'


**    ：**
 torch/onnx/symbolic_opset9.py      remainder   ，       

@parse_args( 'v', 'v')
def remainder(g,input,division):
    return g.op("Remainder",input,division)