tensorflow 各种报错总结

最新推荐文章于 2024-08-17 21:07:01 发布

AICVer

最新推荐文章于 2024-08-17 21:07:01 发布

阅读量6.9k

点赞数

CC 4.0 BY-SA版权

分类专栏：编码 Python

本文链接：https://blog.youkuaiyun.com/u011489887/article/details/79587136

Python 同时被 2 个专栏收录

40 篇文章

订阅专栏

编码

11 篇文章

订阅专栏

本文主要解析了在使用TensorFlow进行深度学习模型开发时遇到的一些典型错误，包括属性错误、版本不一致导致的问题以及常见的ValueError等，并给出了相应的解决办法。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

1. 'list' object has no attribute 'lower' 报错如下

Traceback (most recent call last):
  File "H:/FasionAI/MyNet/train.py", line 27, in <module>
    train_logits = model.inference(train_batch, BATCH_SIZE, N_CLASSES)
  File "H:\FasionAI\MyNet\CnnNet.py", line 60, in inference
    pool2 = tf.nn.pool(conv2, [1, 2, 2, 1], [1, 2, 2, 1], padding='SAME', name='pooling2')
  File "E:\softinstall\Anaconda\lib\site-packages\tensorflow\python\ops\nn_ops.py", line 931, in pool
    (pooling_type.lower()), [input]) as scope:
AttributeError: 'list' object has no attribute 'lower'

问题原因：是tf.nn.max_pool()而不是tf.nn.pool()

2.tensorflow版本不一致

1）AttributeError: module 'tensorflow' has no attribute 'scalar_summary'

tf.image_summary('images', images)改为：tf.summary.image('images', images)

2）AttributeError: 'module' object has no attribute 'scalar_summary'

tf.scalar_summary('images', images)改为：tf.summary.scalar('images', images)

3）AttributeError: module 'tensorflow' has no attribute 'merge_all_summaries'

tf.merge_all_summaries()改为：summary_op = tf.summary.merge_all()

4）AttributeError: 'module' object has no attribute 'SummaryWriter'

tf.train.SummaryWritter改为tf.summary.FileWriter

3.pycharm程序无法识别自己写的程序import时显示红线

1）打开File--》Setting—》打开 Console下的Python Console，把选项（Add source roots to PYTHONPAT）点击勾选上

2）右键点击自己的工作空间，找下面的Mark Directory as 选择Source Root，就可以解决上面的问题了

4.ValueError: Both labels and logits must be provided.

  File "H:\FasionAI\MyNet\resnetmodel\resnet_train.py", line 35, in train
    loss_ = loss(logits=logits, labels=labels)
  File "H:\FasionAI\MyNet\resnetmodel\resnet.py", line 148, in loss
    cross_entropy = tf.nn.sparse_softmax_cross_entropy_with_logits(labels=labels,logits=logits)
  File "E:\softinstall\Anaconda\lib\site-packages\tensorflow\python\ops\nn_ops.py", line 1935, in sparse_softmax_cross_entropy_with_logits
    labels, logits)
  File "E:\softinstall\Anaconda\lib\site-packages\tensorflow\python\ops\nn_ops.py", line 1715, in _ensure_xent_args
    raise ValueError("Both labels and logits must be provided.")
ValueError: Both labels and logits must be provided.

查看源码后发现是：

  if labels is None or logits is None:
    raise ValueError("Both labels and logits must be provided.")

也就是说有一个输入为空，后检查发现调用函数 inference_small_config(x, c)时没有return 正确应该是

return inference_small_config(x, c)