model - { "error": "Serving signature name: "serving_default"not found in signature def"}

Question

我使用GCP(谷歌云平台)来训练我的模型，我可以导出导出的模型。我使用了该模型并使用了服务 1.8 CPU 的 Tensorflow 的本地 docker 镜像，我得到以下结果作为 REST 后调用的输出 { “错误”:“服务签名名称:\“serving_default\”在签名定义中找不到” }

Answer 1

使用 SavedModelCLI 命令查看模型的 SignatureDef，如下所示:

saved_model_cli show --dir /usr/local/google/home/abc/serving/tensorflow_serving/servables/tensorflow/testdata/saved_model_half_plus_three/00000123

它应该显示如下所示的内容:

signature_def['Classify']:
  The given SavedModel SignatureDef contains the following input(s):
    inputs['inputs'] tensor_info:
        dtype: DT_STRING
        shape: unknown_rank
        name: tf_example:0
  The given SavedModel SignatureDef contains the following output(s):
    outputs['outputs'] tensor_info:
        dtype: DT_FLOAT
        shape: (-1, 1)
        name: y2:0
  Method name is: tensorflow/serving/regress

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

检查您是否可以在上述签名定义中看到名为 serving_default 的签名。

根据我的理解，很可能该签名丢失，因此会导致错误。