tensorflow 2.7 load saved model error with java

Question

我用 tf2.7 保存了一個 model 格式的 model 並成功加載：

       model = tf.saved_model.load(export_dir=export_dir)

當我使用 saved_model_cli 檢查這個 model 時，它重新返回：

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

signature_def['serving_default']:
  The given SavedModel SignatureDef contains the following input(s):
    inputs['examples'] tensor_info:
        dtype: DT_STRING
        shape: (-1)
        name: input_example_tensor:0
  The given SavedModel SignatureDef contains the following output(s):
    outputs['cf_1'] tensor_info:
        dtype: DT_INT64
        shape: (-1, 1)
        name: ParseExample/ParseExampleV2:0
    outputs['cf_2'] tensor_info:
        dtype: DT_INT64
        shape: (-1, 1)
        name: ParseExample/ParseExampleV2:1
    outputs['cf_label'] tensor_info:
        dtype: DT_INT64
        shape: (-1, 1)
        name: ParseExample/ParseExampleV2:2
    outputs['score'] tensor_info:
        dtype: DT_FLOAT
        shape: (-1, 1)
        name: score:0
  Method name is: tensorflow/serving/predict

但是當我嘗試使用 java 來加載這個 model 時，它出錯了：

    val bundle = tf.SavedModelBundle.load("/Users/jiyue/git/dl_flow/model_dir_cfck/1646106315","serve")

  <dependency>
            <groupId>org.tensorflow</groupId>
            <artifactId>tensorflow</artifactId>
            <version>1.15.0</version>
        </dependency>

錯誤是：

Exception in thread "main" org.tensorflow.TensorFlowException: Op type not registered 'ParseExampleV2' in binary running on xxxx. Make sure the Op and Kernel are registered in the binary running in this process. Note that if you are loading a saved graph which used ops from tf.contrib, accessing (e.g.) `tf.contrib.resampler` should be done before importing the graph, as contrib ops are lazily registered when the module is first accessed.

我需要有人幫助我解決它，謝謝！

Answer 1

我寫了一個庫來導入和管理保存的模型。 如果你想看看： JavaTF

如果您願意，可以按照我在自述文件中提供的說明進行操作，然后測試功能。

您可以運行 Object 檢測（推理），並獲得檢測結果。您還可以繪制邊界框。

讓我知道是否有幫助:)