TensorFlow报错TypeError: Fetch argument None has invalid type class 'NoneType'

最新推荐文章于 2026-06-17 12:20:09 发布

转载最新推荐文章于 2026-06-17 12:20:09 发布 · 1.4k 阅读

标签

#tensorflow #cnn #python

tensorflow 同时被 2 个专栏收录

2 篇文章

订阅专栏

cnn

2 篇文章

订阅专栏

博客主要讲述了TensorFlow运行CNN卷积神经网络时出现的问题。循环第一步顺畅，第二步报错，问题出在第88行，因有两个相同变量train_op，第一次迭代时train_op被赋值为None，导致第二次出错。解决办法是将第二个变量改为“_”或其他变量名。

tensorflow运行cnn卷积神经网络时出现问题具体如下：

Traceback (most recent call last):
  File "cnn_mnist.py", line 88, in <module>
    train_loss, train_op = sess.run([loss, train_op], {input_x: batch[0], output_y:batch[1]})
  File "/home/wmm/wmmf/Envs/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 929, in run
    run_metadata_ptr)
  File "/home/wmm/wmmf/Envs/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 1137, in _run
    self._graph, fetches, feed_dict_tensor, feed_handles=feed_handles)
  File "/home/wmm/wmmf/Envs/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 471, in __init__
    self._fetch_mapper = _FetchMapper.for_fetch(fetches)
  File "/home/wmm/wmmf/Envs/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 261, in for_fetch
    return _ListFetchMapper(fetch)
  File "/home/wmm/wmmf/Envs/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 370, in __init__
    self._mappers = [_FetchMapper.for_fetch(fetch) for fetch in fetches]
  File "/home/wmm/wmmf/Envs/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 258, in for_fetch
    type(fetch)))
TypeError: Fetch argument None has invalid type <type 'NoneType'>

而且循环的第一步进行得很顺畅，从第二步开始报这个错误。

可以根据报错看到问题出现在第88行：

train_loss, train_op = sess.run([loss, train_op], {input_x: batch[0], output_y:batch[1]})

为什么第一次没有错，而第二次出错呢？
可以看到这一行有两个相同变量train_op，在第一次迭代时，train_op变量被赋值了sess.run()结果的第二个元素，恰好是None，导致第二次迭代中sess.run()中的train_op变成了None，就出错了。
解决的方法很简单，就是把两个变量的第二个变量改为“_”（或者是任意的变量名如“hhh”）：

train_loss, _ = sess.run([loss, train_op], {input_x: batch[0], output_y:batch[1]})

参考（转载）自：https://blog.csdn.net/qq_41000891/article/details/84555225