same code. run ok on version 19.04 but cannot run in version 19.11 or 19.12

tensorrtserver.api.InferenceServerException: [ 0] input 'pred_sbbox/concat_2' values must be specified as a list of numpy arrays or as a single c_void_p (representing the shared memory handle) or as a tuple of c_void_p (representing the shared memory handle) and list (representing the shape of the input tensor)

my config file

input [
    {
        name: "input/input_data"
        data_type: TYPE_FP32
        dims: [ 416, 416, 3 ]
    }
]
output [
   {
      name: "pred_sbbox/concat_2"
      data_type: TYPE_FP32
      dims: [ 52, 52, 3, 85 ]
      label_filename: "labels.txt"
   },
   {
      name: "pred_mbbox/concat_2"
      data_type: TYPE_FP32
      dims: [26, 26, 3, 85 ]
      label_filename: "labels.txt"
   },
   {                              
      name: "pred_lbbox/concat_2"               
      data_type: TYPE_FP32        
      dims: [ 13, 13, 3, 85 ]         
      label_filename: "labels.txt"
   }                              

]

[deadeyegoodwin]

image_data is an np array. [image_data] is a list of np array. Why do you wrap it again in list()? That gives you a list of list of np.array.

[HoangTienDuc]

same error. how to solve it

image_data = np.random.rand(3,112,112).astype("float32")

# print(np.array([image_data]).shape)


#TODO run inference`
request_ids = []
request_ids.append(ctx.async_run(
        { input_name : list([image_data]) },
        { output_name : (InferContext.ResultFormat.RAW) }, 1)
        )

tensorrtserver.api.InferenceServerException: [ 0] input 'fc1' values must be specified as a list of numpy arrays or as a single c_void_p(representing the shared memory handle) or as a tuple of c_void_p (representing the shared memory handle) and list (representing the shape of the input tensor)