“类型错误:‘张量’对象不可迭代"；张量流估计器的错误 [英] "TypeError: 'Tensor' object is not iterable" error with tensorflow Estimator

问题描述

我有一个程序生成的(无限)数据源，我正在尝试将其用作高级 Tensorflow Estimator 的输入，以训练基于图像的 3D 对象检测器.

I have a procedurally generated (infinite) data source and am trying to use this as input to the high-level Tensorflow Estimator to train a image-based 3D object detector.

我像在 Tensorflor Estimator 快速入门中一样设置了数据集，并且我的dataset_input_fn 返回一组特征和标签 Tensor 的，就像 Estimator.train 函数指定，以及这个 教程展示，但在尝试调用 train 函数时出现错误:

I set up the Dataset just as in the Tensorflor Estimator Quickstart, and my dataset_input_fn returns a tuple of features and labels Tensor's, just as the Estimator.train function specifies, and how this tutorial shows, but I am getting an error when trying to call the train function:

TypeError: 'Tensor' 对象不可迭代.

我做错了什么?

    def data_generator():
        """
        Generator for image (features) and ground truth object positions (labels)

        Sample an image and object positions from a procedurally generated data source
        """
        while True:
            source.step()  # generate next data point

            object_ground_truth = source.get_ground_truth() # list of 9 floats
            cam_img = source.get_cam_frame()  # image (224, 224, 3) 
            yield (cam_img, object_ground_truth)

    def dataset_input_fn():
        """
        Tensorflow `Dataset` object from generator
        """

        dataset = tf.data.Dataset.from_generator(data_generator, (tf.uint8, tf.float32), \
            (tf.TensorShape([224, 224, 3]), tf.TensorShape([9])))
        dataset = dataset.batch(16)

        iterator = dataset.make_one_shot_iterator()

        features, labels = iterator.get_next()
        return features, labels

    def main():
        """
        Estimator [from Keras model](https://www.tensorflow.org/programmers_guide/estimators#creating_estimators_from_keras_models) 

        Try to call `est_vgg.train()` leads to the error
        """
        ....
        est_vgg16 = tf.keras.estimator.model_to_estimator(keras_model=keras_vgg16)
        est_vgg16.train(input_fn=dataset_input_fn, steps=10)
        ....

这是完整代码

(注意:事物的命名与这个问题不同)

(note: things are named differently from this question)

这是堆栈跟踪:

Traceback (most recent call last):
  File "./rock_detector.py", line 155, in <module>
    main()
  File "./rock_detector.py", line 117, in main
    est_vgg16.train(input_fn=dataset_input_fn, steps=10)
  File "/usr/local/lib/python3.6/site-packages/tensorflow/python/estimator/estimator.py", line 302, in train
    loss = self._train_model(input_fn, hooks, saving_listeners)
  File "/usr/local/lib/python3.6/site-packages/tensorflow/python/estimator/estimator.py", line 711, in _train_model
    features, labels, model_fn_lib.ModeKeys.TRAIN, self.config)
  File "/usr/local/lib/python3.6/site-packages/tensorflow/python/estimator/estimator.py", line 694, in _call_model_fn
    model_fn_results = self._model_fn(features=features, **kwargs)
  File "/usr/local/lib/python3.6/site-packages/tensorflow/python/keras/_impl/keras/estimator.py", line 145, in model_fn
    labels)
  File "/usr/local/lib/python3.6/site-packages/tensorflow/python/keras/_impl/keras/estimator.py", line 92, in _clone_and_build_model
    keras_model, features)
  File "/usr/local/lib/python3.6/site-packages/tensorflow/python/keras/_impl/keras/estimator.py", line 58, in _create_ordered_io
    for key in estimator_io_dict:
  File "/usr/local/lib/python3.6/site-packages/tensorflow/python/framework/ops.py", line 505, in __iter__
    raise TypeError("'Tensor' object is not iterable.")
TypeError: 'Tensor' object is not iterable.

推荐答案

让你的输入函数返回一个特征字典，如下所示:

Make your input function return a dictionary of features like this:

def dataset_input_fn():
  ...
  features, labels = iterator.get_next()
  return {'image': features}, labels

这篇关于“类型错误:‘张量’对象不可迭代"；张量流估计器的错误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！