Does saving the model with 'tf.saved_model.save' match this?

Version

tensorflow==2.0.0
Keras==2.2.4
tensorrt==7.0.0

ubuntu18.04
cuda10.2
Python 3.6.9

SavedModel

I don’t know which one to use for the SavedModel format, because the saved model formats are different.

In Case 1, since the layer information is not complete, it is unclear whether the model conversion is done properly with ‘trt.TrtGraphConverterV2’.

In Case 2, the layer information is completely written out, but it cannot be converted with ‘trt.TrtGraphConverterV2’.

It is a numerical image of the data MNIST used.
Please help someone.

[Case1]

tf.saved_model.save(model,saved_model_path)
$ saved_model_cli show --dir ./saved_models --all


MetaGraphDef with tag-set: 'serve' contains the following SignatureDefs:

signature_def['serving_default']:
  The given SavedModel SignatureDef contains the following input(s):
    inputs['input'] tensor_info:
        dtype: DT_FLOAT
        shape: (-1, 28, 28, 1)
        name: conv2d_input:0
  The given SavedModel SignatureDef contains the following output(s):
    outputs['dense_1/Identity:0'] tensor_info:
        dtype: DT_FLOAT
        shape: (-1, 10)
        name: dense_1/Identity:0
  Method name is: tensorflow/serving/predict

[Case2]

tf.keras.experimental.export_saved_model(model, saved_model_path)
$ saved_model_cli show --dir ./saved_models --all

MetaGraphDef with tag-set: 'eval' contains the following SignatureDefs:

signature_def['__saved_model_init_op']:
  The given SavedModel SignatureDef contains the following input(s):
  The given SavedModel SignatureDef contains the following output(s):
    outputs['__saved_model_init_op'] tensor_info:
        dtype: DT_INVALID
        shape: unknown_rank
        name: init_1
  Method name is: 

signature_def['eval']:
  The given SavedModel SignatureDef contains the following input(s):
    inputs['conv2d_input'] tensor_info:
        dtype: DT_FLOAT
        shape: (-1, 28, 28, 1)
        name: conv2d_input:0
    inputs['dense_1_target'] tensor_info:
        dtype: DT_FLOAT
        shape: (-1, -1)
        name: dense_1_target:0
  The given SavedModel SignatureDef contains the following output(s):
    outputs['loss'] tensor_info:
        dtype: DT_FLOAT
        shape: ()
        name: loss/mul:0
    outputs['metrics/accuracy/update_op'] tensor_info:
        dtype: DT_FLOAT
        shape: (0)
        name: metric_op_wrapper:0
    outputs['metrics/accuracy/value'] tensor_info:
        dtype: DT_FLOAT
        shape: ()
        name: Identity_13:0
    outputs['predictions/dense_1'] tensor_info:
        dtype: DT_FLOAT
        shape: (-1, 10)
        name: dense_1/Softmax:0
  Method name is: tensorflow/supervised/eval

MetaGraphDef with tag-set: 'serve' contains the following SignatureDefs:

signature_def['__saved_model_init_op']:
  The given SavedModel SignatureDef contains the following input(s):
  The given SavedModel SignatureDef contains the following output(s):
    outputs['__saved_model_init_op'] tensor_info:
        dtype: DT_INVALID
        shape: unknown_rank
        name: init_1
  Method name is: 

signature_def['serving_default']:
  The given SavedModel SignatureDef contains the following input(s):
    inputs['conv2d_input'] tensor_info:
        dtype: DT_FLOAT
        shape: (-1, 28, 28, 1)
        name: conv2d_input:0
  The given SavedModel SignatureDef contains the following output(s):
    outputs['dense_1'] tensor_info:
        dtype: DT_FLOAT
        shape: (-1, 10)
        name: dense_1/Softmax:0
  Method name is: tensorflow/serving/predict

MetaGraphDef with tag-set: 'train' contains the following SignatureDefs:

signature_def['__saved_model_init_op']:
  The given SavedModel SignatureDef contains the following input(s):
  The given SavedModel SignatureDef contains the following output(s):
    outputs['__saved_model_init_op'] tensor_info:
        dtype: DT_INVALID
        shape: unknown_rank
        name: init_1
  Method name is: 

signature_def['__saved_model_train_op']:
  The given SavedModel SignatureDef contains the following input(s):
  The given SavedModel SignatureDef contains the following output(s):
    outputs['__saved_model_train_op'] tensor_info:
        dtype: DT_INVALID
        shape: unknown_rank
        name: training_1/group_deps
  Method name is: 

signature_def['train']:
  The given SavedModel SignatureDef contains the following input(s):
    inputs['conv2d_input'] tensor_info:
        dtype: DT_FLOAT
        shape: (-1, 28, 28, 1)
        name: conv2d_input:0
    inputs['dense_1_target'] tensor_info:
        dtype: DT_FLOAT
        shape: (-1, -1)
        name: dense_1_target:0
  The given SavedModel SignatureDef contains the following output(s):
    outputs['loss'] tensor_info:
        dtype: DT_FLOAT
        shape: ()
        name: loss/mul:0
    outputs['metrics/accuracy/update_op'] tensor_info:
        dtype: DT_FLOAT
        shape: (0)
        name: metric_op_wrapper:0
    outputs['metrics/accuracy/value'] tensor_info:
        dtype: DT_FLOAT
        shape: ()
        name: Identity_38:0
    outputs['predictions/dense_1'] tensor_info:
        dtype: DT_FLOAT
        shape: (-1, 10)
        name: dense_1/Softmax:0
  Method name is: tensorflow/supervised/training