python - 机器学习引擎 : Prediction Error while executing local predict command

Question

我已在 Google ML Engine 中上传了该模型的一个版本，其中包含 saved_model.pb 和一个变量文件夹。当我尝试执行命令时:

gcloud ml-engine local predict --model-dir=saved_model --json-instances=request.json

它显示以下错误:

ERROR: (gcloud.ml-engine.local.predict) 2018-09-11 19:06:39.770396: I tensorflow/core/platform/cpu_feature_guard.cc:141] 
Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA

Traceback (most recent call last):
  File "lib/googlecloudsdk/command_lib/ml_engine/local_predict.py", line 172, in <module>
    main()
  File "lib/googlecloudsdk/command_lib/ml_engine/local_predict.py", line 167, in main
    signature_name=args.signature_name)
  File "/usr/lib/google-cloud-sdk/lib/third_party/ml_sdk/cloud/ml/prediction/prediction_lib.py", line 106, in local_predict
    predictions = model.predict(instances, signature_name=signature_name)
  File "/usr/lib/google-cloud-sdk/lib/third_party/ml_sdk/cloud/ml/prediction/prediction_utils.py", line 230, in predict
    preprocessed = self.preprocess(instances, stats=stats, **kwargs)
  File "/usr/lib/google-cloud-sdk/lib/third_party/ml_sdk/cloud/ml/prediction/frameworks/tf_prediction_lib.py", line 436, in preprocess
    preprocessed = self._canonicalize_input(instances, signature)
  File "/usr/lib/google-cloud-sdk/lib/third_party/ml_sdk/cloud/ml/prediction/frameworks/tf_prediction_lib.py", line 453, in _canonicalize_input
    return canonicalize_single_tensor_input(instances, tensor_name)
  File "/usr/lib/google-cloud-sdk/lib/third_party/ml_sdk/cloud/ml/prediction/frameworks/tf_prediction_lib.py", line 166, in canonicalize_single_tensor_input
    instances = [parse_single_tensor(x, tensor_name) for x in instances]
  File "/usr/lib/google-cloud-sdk/lib/third_party/ml_sdk/cloud/ml/prediction/frameworks/tf_prediction_lib.py", line 162, in parse_single_tensor
    (tensor_name, list(x.keys())))


cloud.ml.prediction.prediction_utils.PredictionError: Invalid inputs: Expected tensor name: inputs, got tensor name: [u'inputs', u'key']. (Error code: 1)

我的request.json文件是

{"inputs": {"b64": "/9j/4AAQSkZJRgABAQAAAQABAAD/2wBDAAgGBgcGBQgHBwcJCQgKDBQNDAsLDBkSEw8UHRofHh0aHBwgJC4nICIsIxwcKDcpLDAxNDQ0Hyc5PTgyPC4zNDL/2wBDAQkJCQwLDBgNDRgyIRwhMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjL/wAARCAHVArwDASIAAhEBAxEB/8QAHwAAAQUBAQEBA....."}, "key": "841bananas.jpg"}

提前致谢。

Answer 1

您的模型似乎仅使用一个名为“inputs”的输入导出。在这种情况下，您不应该在 JSON 中发送“key”，即(滚动到末尾查看我已删除“keys”):

{"inputs": {"b64": "/9j/4AAQSkZJRgABAQAAAQABAAD/2wBDAAgGBgcGBQgHBwcJCQgKDBQNDAsLDBkSEw8UHRofHh0aHBwgJC4nICIsIxwcKDcpLDAxNDQ0Hyc5PTgyPC4zNDL/2wBDAQkJCQwLDBgNDRgyIRwhMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjL/wAARCAHVArwDASIAAhEBAxEB/8QAHwAAAQUBAQEBA....."}}