TensorFlow ExportOutputs、PredictOuput 和指定 signature_constants.DEFAULT_SERVING_SIGNATURE_DEF_KEY [英] TensorFlow ExportOutputs, PredictOuput, and specifying signature_constants.DEFAULT_SERVING_SIGNATURE_DEF_KEY

问题描述

我有一个 colab 和一个非常简单的演示 Estimator 用于学习/理解 Estimator API 旨在为即插即用模型，具有有用的交易技巧(例如，如果验证集停止改进，则提前停止，导出模型等).

I have a colab with a very simple demo Estimator for the purpose of learning / understanding the Estimator API with the goal of making a convention for a plug-and-play model with useful bells and whistles of the trade in tack (e.g. early stopping if the validation set stops improving, exporting the model, etc).

三个 Estimator 模式(TRAIN、EVAL 和 PREDICT)返回一个 EstimatorSpec.

Each of the three Estimator modes (TRAIN, EVAL, and PREDICT) return an EstimatorSpec.

根据文档:

__new__(
    cls,
    mode,
    predictions=None,          # required by PREDICT
    loss=None,                 # required by TRAIN and EVAL
    train_op=None,             # required by TRAIN
    eval_metric_ops=None,
    export_outputs=None,
    training_chief_hooks=None,
    training_hooks=None,
    scaffold=None,
    evaluation_hooks=None,
    prediction_hooks=None.     
)

在这些命名参数中，我想提请注意 predictions 和 export_outputs，它们在 docs 为:

Of these named arguments I would like to bring attention to predictions and export_outputs, which are described in the docs as:

• predictions:预测张量或张量字典.
• export_outputs:描述要导出到 SavedModel 并在服务期间使用的输出签名.一个 dict {name: output} 其中:
  • name:此输出的任意名称.
  • output:一个ExportOutput 对象，例如 ClassificationOutput、RegressionOutput 或 PredictOutput.单头模型只需要在这个字典中指定一个条目.多头模型应为每个头指定一个条目，其中之一必须使用 signature_constants.DEFAULT_SERVING_SIGNATURE_DEF_KEY 命名.如果未提供条目，则将创建一个默认的 PredictOutput 映射到预测.

  • predictions: Predictions Tensor or dict of Tensor.
  • export_outputs: Describes the output signatures to be exported to SavedModel and used during serving. A dict {name: output} where:
    • name: An arbitrary name for this output.
    • output: an ExportOutput object such as ClassificationOutput, RegressionOutput, or PredictOutput. Single-headed models only need to specify one entry in this dictionary. Multi-headed models should specify one entry for each head, one of which must be named using signature_constants.DEFAULT_SERVING_SIGNATURE_DEF_KEY. If no entry is provided, a default PredictOutput mapping to predictions will be created.

  因此应该很清楚为什么我提出了export_outputs；也就是说，因为人们最有可能在未来使用他们训练的模型(通过从 SavedModel 加载它).

  Thus it should be clear why I bring up export_outputs; namely, as one would most likely like to use the model they trained in the future (by loading it from a SavedModel).

  为了让这个问题更容易理解/增加一些清晰度:

  To make this question a bit more accessible / add some clarity:

  • 单头"模型是人们遇到的最常见的模型，其中 input_fn features 被转换为单个(批处理)output

  • "single-headed" models are the most common model one encounters where the input_fn features are transformed to a singular (batched) output

  多头"模型是具有多个输出的模型

  "multi-headed" models are models where there is more than one output

  例如这个多头模型的input_fn(根据Estimator api) 返回一个元组 (features, labels) 即这个模型有两个头.

  e.g. this multi-headed model's input_fn (in accordance with the Estimator api) returns a tuple (features, labels) i.e. this model has two heads).

  def input_fn():
    features = ...
    labels1 = ...
    labels2 = ...
    return features, {'head1': labels1, 'head2': labels2}
  

  如何指定signature_constants.DEFAULT_SERVING_SIGNATURE_DEF_KEY是这个问题的核心.即，如何指定它?(例如，它应该是一个 dict {signature_constants.DEFAULT_SERVING_SIGNATURE_DEF_KEY: head})

  How one specifies the signature_constants.DEFAULT_SERVING_SIGNATURE_DEF_KEY is the core of this question. Namely, how does one specify it? (e.g. should it be a dict {signature_constants.DEFAULT_SERVING_SIGNATURE_DEF_KEY: head})

  是的，所以在 colab 中，您会看到我们模型的 export_outputs 实际上是以多头方式定义的(虽然不应该如此):

  Right, so in the colab you see that our model's export_outputs is actually defined in a multi-head manner (although it shouldn't be):

  来自 estimator 函数 >colab 的模型_fn:

  From estimator functions > model_fn of the colab:

  def model_fn(...):
  
      # ...
  
      # send the features through the graph
      MODEL = build_fn(MODEL)
  
      # prediction
      MODEL['predictions'] = {'labels': MODEL['net_logits']} # <--- net_logits added in the build_fn
  
      MODEL['export_outputs'] = {
          k: tf.estimator.export.PredictOutput(v) for k, v in MODEL['predictions'].items()
      }
  
      # ...
  

  在这个特定的例子中，如果我们扩展字典理解，我们的功能相当于:

  in this particular instance, if we expand the dictionary comprehension, we have the functional equivalent of:

  MODEL['export_outputs'] = {
      'labels': tf.estimator.export.PredictOutput(MODEL['net_logits'])
  }
  

  它在这种情况下有效，因为我们的字典有一个键，因此有一个 PredictOutput，其中在 colab我们的 model_fn 只有一个头部，格式更合适:

  It works in this instance as our dictionary has one key and hence one PredictOutput, where in the colab our model_fn has only a single head and would be more properly formatted as:

  MODEL['export_outputs'] = {
      'predictions': tf.estimator.export.PredictOutput(MODEL['predictions'])
  }
  

  正如 PredictOutput:

  __init__(outputs)
  

  哪里

  • outputs:一个 Tensor 或一个表示预测的 Tensor 字符串的字典.
  • outputs: A Tensor or a dict of string to Tensor representing the predictions.

  因此我的问题如下:

  1. 如果 PredictOutput 可以是一本字典，当/为什么要多个 PredictOutputs 作为 export_outputs="nofollow noreferrer">EstimatorSpec?

  1. if PredictOutput can be a dictionary, when / why would one want multiple PredictOutputs as their export_outputs for the EstimatorSpec?

  如果有一个多头模型，(比如有多个 PredictOutputs) 如何实际指定 signature_constants.DEFAULT_SERVING_SIGNATURE_DEF_KEY

  If one has a multi-headed model, (say with multiple PredictOutputs) how does one actually specify the signature_constants.DEFAULT_SERVING_SIGNATURE_DEF_KEY

  EstimatorSpec 当它在 export_outputs 中也是必需的"(对于关心使用 SavedModels 的任何人)时?

  what is the point of predictions in the EstimatorSpec when it is also "required" (for anyone who cares about using SavedModels) in export_outputs?

  推荐答案

  感谢您的详细提问；你显然在这里挖得很深.

  Thanks for your detailed question; you have clearly dug deep here.

  1. 还有一些不能作为字典的 RegressionOutput 和 ClassificationOutput 类.使用 export_outputs dict 允许对这些用例进行泛化.

  1. There are also classes for RegressionOutput and ClassificationOutput which cannot be dictionaries. The use of an export_outputs dict allows for generalizations over those use cases.

  您希望从保存的模型中默认使用的头部应采用默认签名密钥.例如:

  The head you want to be served by default from the saved model should take the default signature key. For example:

  export_outputs = {
    signature_constants.DEFAULT_SERVING_SIGNATURE_DEF_KEY:
      PredictOutput(outputs={'some_output_1': output_1}),
    'head-2': PredictOutput(outputs={'some_output_2': output_2}),
    'head-3': PredictOutput(outputs={'some_output_3': output_3})
  }
  

  1. 原因 1:很多人使用默认的 export_outputs(这又是预测的值)，或者不导出到保存的模型.原因2:历史.预测是第一位的，随着时间的推移，越来越多的功能被添加进来.这些功能需要灵活性和额外信息，因此被独立打包到 EstimatorSpec 中.

  这篇关于TensorFlow ExportOutputs、PredictOuput 和指定 signature_constants.DEFAULT_SERVING_SIGNATURE_DEF_KEY的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！