tensorflow线性或增强树训练失败并显示“错误解析消息”

如何解决tensorflow线性或增强树训练失败并显示“错误解析消息”

我正在尝试通过TF 2.0运行一个示例，该示例与以下示例基本相同：https://www.tensorflow.org/tutorials/estimator/boosted_trees，除了以下事实：

我的数组更大，为63319行x 7330列，而不是627行和9列
我没有分类列，只有数字列
我的许多列名都相当长（可能是50个字符，而不是那个长！）

在该示例中，数据来自pandas数据框，具有int标签，等等。我的数据中没有nan或inf。

如果我尝试在数据上训练线性分类器或增强树分类器，最终我会得到

google.protobuf.message.DecodeError: Error parsing message

最后提到的文件是

File "/home/ginsberg/.local/lib/python3.6/site-packages/tensorflow/python/framework/ops.py",line 3211,in _as_graph_def
    graph.ParseFromString(compat.as_bytes(data))

有什么想法吗？我在Ubuntu 18.04和MacOS 10.15.6上都看到了相同的行为。 Ubuntu是Python 3.6.9，TF 2.3.0，protobuf 3.13.0。 Mac是Python 3.8.3，TF 2.3.0，protobuf 3.12.3

我可以采取一些措施来跟踪这种情况，例如逐渐缩小数据集直到错误消失，更改列标题等，但是它们都比较痛苦，似乎不太可能产生有用的信息。

非常感谢！

完全追溯错误（格式有些令人不愉快；我希望可以）：

File "<stdin>",line 1,in <module>  
  File "/home/ginsberg/.local/lib/python3.6/site-packages/tensorflow_estimator/python/estimator/estimator.py",line 349,in train  
    loss = self._train_model(input_fn,hooks,saving_listeners)  
  File "/home/ginsberg/.local/lib/python3.6/site-packages/tensorflow_estimator/python/estimator/estimator.py",line 1175,in _train_model  
    return self._train_model_default(input_fn,line 1208,in _train_model_default  
    saving_listeners)  
  File "/home/ginsberg/.local/lib/python3.6/site-packages/tensorflow_estimator/python/estimator/estimator.py",line 1507,in _train_with_estimator_spec  
    log_step_count_steps=log_step_count_steps) as mon_sess:  
  File "/home/ginsberg/.local/lib/python3.6/site-packages/tensorflow/python/training/monitored_session.py",line 604,in MonitoredTrainingSession  
    stop_grace_period_secs=stop_grace_period_secs)  
  File "/home/ginsberg/.local/lib/python3.6/site-packages/tensorflow/python/training/monitored_session.py",line 1038,in __init__  
    stop_grace_period_secs=stop_grace_period_secs)  
  File "/home/ginsberg/.local/lib/python3.6/site-packages/tensorflow/python/training/monitored_session.py",line 737,in __init__  
    h.begin()  
  File "/home/ginsberg/.local/lib/python3.6/site-packages/tensorflow/python/training/basic_session_run_hooks.py",line 563,in begin  
    self._summary_writer = SummaryWriterCache.get(self._checkpoint_dir)  
  File "/home/ginsberg/.local/lib/python3.6/site-packages/tensorflow/python/summary/writer/writer_cache.py",line 63,in get  
    logdir,graph=ops.get_default_graph())  
  File "/home/ginsberg/.local/lib/python3.6/site-packages/tensorflow/python/summary/writer/writer.py",line 371,in __init__  
    super(FileWriter,self).__init__(event_writer,graph,graph_def)  
  File "/home/ginsberg/.local/lib/python3.6/site-packages/tensorflow/python/summary/writer/writer.py",line 84,in __init__  
    self.add_graph(graph=graph,graph_def=graph_def)  
  File "/home/ginsberg/.local/lib/python3.6/site-packages/tensorflow/python/summary/writer/writer.py",line 194,in add_graph  
    true_graph_def = graph.as_graph_def(add_shapes=True)  
  File "/home/ginsberg/.local/lib/python3.6/site-packages/tensorflow/python/framework/ops.py",line 3294,in as_graph_def  
    result,_ = self._as_graph_def(from_version,add_shapes)  
  File "/home/ginsberg/.local/lib/python3.6/site-packages/tensorflow/python/framework/ops.py",in _as_graph_def  
    graph.ParseFromString(compat.as_bytes(data))  
google.protobuf.message.DecodeError: Error parsing message

和产生错误的代码只是est.train（train_input_fn，max_steps = 100），就像在示例中一样（train_input_fn也是示例的副本）。数组大小如上所述； X是63319 x 7330，Y是63319 x 1。