使用bert时使用autokeras会引起文件系统方案'gs'错误

问题描述

我目前正在研究autokeras，并尝试检查它是否找到了比我在进行网格搜索后发现的模型更好的模型。

但是，我在使用以下代码时发现了

input_node = ak.Textinput()
output_node = ak.TextBlock(pretraining = 'random')(input_node)
output_node = ak.ClassificationHead()(output_node)
new_model = ak.AutoModel(
    inputs = input_node,outputs = output_node,overwrite = True,objective = 'val_accuracy',max_trials = 50)
history = new_model.fit(x_train,y_train)

我遇到以下错误：

Traceback (most recent call last):
  File "c:\users\maurits van roozenda\appdata\local\pypoetry\cache\virtualenvs\autokeras-project-dtjkqbuf-py3.8\src\keras-tuner\kerastuner\engine\hypermodel.py",line 104,in build
    model = self.hypermodel.build(hp)
  File "c:\users\maurits van roozenda\appdata\local\pypoetry\cache\virtualenvs\autokeras-project-dtjkqbuf-py3.8\src\keras-tuner\kerastuner\engine\hypermodel.py",line 64,in _build_wrapper
    return self._build(hp,*args,**kwargs)
  File "C:\Users\Maurits van Roozenda\AppData\Local\pypoetry\Cache\virtualenvs\autokeras-project-DtjkQbuf-py3.8\lib\site-packages\autokeras\graph.py",line 269,in build
    outputs = block.build(hp,inputs=temp_inputs)
  File "C:\Users\Maurits van Roozenda\AppData\Local\pypoetry\Cache\virtualenvs\autokeras-project-DtjkQbuf-py3.8\lib\site-packages\autokeras\engine\block.py",line 38,in _build_wrapper
    return super()._build_wrapper(hp,**kwargs)
  File "c:\users\maurits van roozenda\appdata\local\pypoetry\cache\virtualenvs\autokeras-project-dtjkqbuf-py3.8\src\keras-tuner\kerastuner\engine\hypermodel.py",**kwargs)
  File "C:\Users\Maurits van Roozenda\AppData\Local\pypoetry\Cache\virtualenvs\autokeras-project-DtjkQbuf-py3.8\lib\site-packages\autokeras\blocks\wrapper.py",line 148,in build
    output_node = self._build_block(hp,output_node,block_type)
  File "C:\Users\Maurits van Roozenda\AppData\Local\pypoetry\Cache\virtualenvs\autokeras-project-DtjkQbuf-py3.8\lib\site-packages\autokeras\blocks\wrapper.py",line 163,in _build_block
    output_node = basic.BertBlock().build(hp,output_node)
  File "C:\Users\Maurits van Roozenda\AppData\Local\pypoetry\Cache\virtualenvs\autokeras-project-DtjkQbuf-py3.8\lib\site-packages\autokeras\engine\block.py",**kwargs)
  File "C:\Users\Maurits van Roozenda\AppData\Local\pypoetry\Cache\virtualenvs\autokeras-project-DtjkQbuf-py3.8\lib\site-packages\autokeras\blocks\basic.py",line 777,in build
    tokenizer_layer = keras_layers.BertTokenizer(
  File "C:\Users\Maurits van Roozenda\AppData\Local\pypoetry\Cache\virtualenvs\autokeras-project-DtjkQbuf-py3.8\lib\site-packages\autokeras\keras_layers.py",line 105,in __init__
    self.tokenizer = tokenization.FullTokenizer(
  File "C:\Users\Maurits van Roozenda\AppData\Local\pypoetry\Cache\virtualenvs\autokeras-project-DtjkQbuf-py3.8\lib\site-packages\official\nlp\bert\tokenization.py",line 175,in __init__
    self.vocab = load_vocab(vocab_file)
  File "C:\Users\Maurits van Roozenda\AppData\Local\pypoetry\Cache\virtualenvs\autokeras-project-DtjkQbuf-py3.8\lib\site-packages\official\nlp\bert\tokenization.py",line 137,in load_vocab
    token = convert_to_unicode(reader.readline())
  File "C:\Users\Maurits van Roozenda\AppData\Local\pypoetry\Cache\virtualenvs\autokeras-project-DtjkQbuf-py3.8\lib\site-packages\tensorflow\python\lib\io\file_io.py",line 169,in readline
    self._preread_check()
  File "C:\Users\Maurits van Roozenda\AppData\Local\pypoetry\Cache\virtualenvs\autokeras-project-DtjkQbuf-py3.8\lib\site-packages\tensorflow\python\lib\io\file_io.py",line 78,in _preread_check
    self._read_buf = _pywrap_file_io.BufferedInputStream(
tensorflow.python.framework.errors_impl.UnimplementedError: File system scheme 'gs' not implemented (file: 'gs://cloud-tpu-checkpoints/bert/keras_bert/uncased_L-12_H-768_A-12\vocab.txt')
Traceback (most recent call last):
  File "c:\users\maurits van roozenda\appdata\local\pypoetry\cache\virtualenvs\autokeras-project-dtjkqbuf-py3.8\src\keras-tuner\kerastuner\engine\hypermodel.py",in _preread_check
    self._read_buf = _pywrap_file_io.BufferedInputStream(
tensorflow.python.framework.errors_impl.UnimplementedError: File system scheme 'gs' not implemented (file: 'gs://cloud-tpu-checkpoints/bert/keras_bert/uncased_L-12_H-768_A-12\vocab.txt')

在autokeras尝试使用bert模型之前，我遇到了这些错误。不幸的是，我不太了解这个“文件系统方案” gs”是什么，我尝试用谷歌搜索，似乎其他人在尝试将tensorboard与云一起使用时都遇到了错误，我正在本地使用但是，我认为这不是我应该尝试找到解决方案的方向。

任何帮助将不胜感激！

解决方法

暂无找到可以解决该程序问题的有效方法，小编努力寻找整理中！

如果你已经找到好的解决方法，欢迎将解决方案带上本链接一起发送给小编。

小编邮箱:dio#foxmail.com (将#修改为@）