Keras和pytorch报错集合整理

1、IndexError: invalid index of a 0-dim tensor. Use tensor.item() to convert a 0-dim tensor to a Python number

torch版本的不同造成的。
解决：将loss.data[0] 改成loss.item()

2、AttributeError: 'module' object has no attribute 'Cmd'

 完整报错：

Using TensorFlow backend.
Traceback (most recent call last):
  File "../create_pairs_dataset.py", line 3, in <module>
    import pose_transform
  File "/home/hyw/zp/keras/pose-gan-master/pose_transform.py", line 18, in <module>
    from tensorflow.contrib.image import transform as tf_affine_transform
  File "/home/hyw/zp/Anaconda2_pytorch_keras/lib/python2.7/site-packages/tensorflow/contrib/__init__.py", line 37, in <module>
    from tensorflow.contrib import cudnn_rnn
  File "/home/hyw/zp/Anaconda2_pytorch_keras/lib/python2.7/site-packages/tensorflow/contrib/cudnn_rnn/__init__.py", line 38, in <module>
    from tensorflow.contrib.cudnn_rnn.python.layers import *
  File "/home/hyw/zp/Anaconda2_pytorch_keras/lib/python2.7/site-packages/tensorflow/contrib/cudnn_rnn/python/layers/__init__.py", line 23, in <module>
    from tensorflow.contrib.cudnn_rnn.python.layers.cudnn_rnn import *
  File "/home/hyw/zp/Anaconda2_pytorch_keras/lib/python2.7/site-packages/tensorflow/contrib/cudnn_rnn/python/layers/cudnn_rnn.py", line 20, in <module>
    from tensorflow.contrib.cudnn_rnn.python.ops import cudnn_rnn_ops
  File "/home/hyw/zp/Anaconda2_pytorch_keras/lib/python2.7/site-packages/tensorflow/contrib/cudnn_rnn/python/ops/cudnn_rnn_ops.py", line 22, in <module>
    from tensorflow.contrib.rnn.python.ops import lstm_ops
  File "/home/hyw/zp/Anaconda2_pytorch_keras/lib/python2.7/site-packages/tensorflow/contrib/rnn/__init__.py", line 93, in <module>
    from tensorflow.contrib.rnn.python.ops.rnn_cell import *
  File "/home/hyw/zp/Anaconda2_pytorch_keras/lib/python2.7/site-packages/tensorflow/contrib/rnn/python/ops/rnn_cell.py", line 24, in <module>
    from tensorflow.contrib.layers.python.layers import layers
  File "/home/hyw/zp/Anaconda2_pytorch_keras/lib/python2.7/site-packages/tensorflow/contrib/layers/__init__.py", line 116, in <module>
    from tensorflow.contrib.layers.python.layers import *
  File "/home/hyw/zp/Anaconda2_pytorch_keras/lib/python2.7/site-packages/tensorflow/contrib/layers/python/layers/__init__.py", line 33, in <module>
    from tensorflow.contrib.layers.python.layers.target_column import *
  File "/home/hyw/zp/Anaconda2_pytorch_keras/lib/python2.7/site-packages/tensorflow/contrib/layers/python/layers/target_column.py", line 24, in <module>
    from tensorflow.contrib.losses.python.losses import loss_ops
  File "/home/hyw/zp/Anaconda2_pytorch_keras/lib/python2.7/site-packages/tensorflow/contrib/losses/__init__.py", line 25, in <module>
    from tensorflow.contrib.losses.python import metric_learning
  File "/home/hyw/zp/Anaconda2_pytorch_keras/lib/python2.7/site-packages/tensorflow/contrib/losses/python/metric_learning/__init__.py", line 25, in <module>
    from tensorflow.contrib.losses.python.metric_learning.metric_loss_ops import *
  File "/home/hyw/zp/Anaconda2_pytorch_keras/lib/python2.7/site-packages/tensorflow/contrib/losses/python/metric_learning/metric_loss_ops.py", line 34, in <module>
    from sklearn import metrics
  File "/home/hyw/zp/Anaconda2_pytorch_keras/lib/python2.7/site-packages/sklearn/metrics/__init__.py", line 7, in <module>
    from .ranking import auc
  File "/home/hyw/zp/Anaconda2_pytorch_keras/lib/python2.7/site-packages/sklearn/metrics/ranking.py", line 36, in <module>
    from ..preprocessing import label_binarize
  File "/home/hyw/zp/Anaconda2_pytorch_keras/lib/python2.7/site-packages/sklearn/preprocessing/__init__.py", line 6, in <module>
    from ._function_transformer import FunctionTransformer
  File "/home/hyw/zp/Anaconda2_pytorch_keras/lib/python2.7/site-packages/sklearn/preprocessing/_function_transformer.py", line 5, in <module>
    from ..utils.testing import assert_allclose_dense_sparse
  File "/home/hyw/zp/Anaconda2_pytorch_keras/lib/python2.7/site-packages/sklearn/utils/testing.py", line 751, in <module>
    import pytest
  File "/home/hyw/zp/Anaconda2_pytorch_keras/lib/python2.7/site-packages/pytest.py", line 13, in <module>
    from _pytest.debugging import pytestPDB as __pytestPDB
  File "/home/hyw/zp/Anaconda2_pytorch_keras/lib/python2.7/site-packages/_pytest/debugging.py", line 6, in <module>
    import pdb
  File "/home/hyw/zp/Anaconda2_pytorch_keras/lib/python2.7/pdb.py", line 59, in <module>
    class Pdb(bdb.Bdb, cmd.Cmd):
AttributeError: 'module' object has no attribute 'Cmd'

报错原因分析： cmd.py是python的一个模块，而在我的代码树目录同样自定义了一个同名的模块cmd.py,该自定义的模块中无Cmd属性，python调用自身的cmd.py导致报错

解决办法：将自定义的cmd.py重命名为 cmdd.py,同时删除同目录下产生的 cmd.pyc文件，重新运行代码，问题解决

参考：  The kernel has died, and the automatic restart has failed

3、ImportError: No module named keras_contrib.layers.normalization

报错原因： Keras未安装 keras_contrib 包

解决办法： 安装即可： github安装链接     或者用这个链接：keras实现Bi-LSTM+CRF

使用该方式导入：from keras_contrib.layers.normalization.instancenormalization import InstanceNormalization

参考：ImportError: No module named 'keras_contrib'

tf.keras 和 keras有什么区别？

4、ValueError: Tensor._shape cannot be assigned, use Tensor.set_shape instead. 

解决办法：将 o._shape = tf.TensorShape(new_shape)  换为 ：   o.set_shape = tf.TensorShape(new_shape)

 参考： https://github.com/bioinf-jku/TTUR/issues/6

5、ValueError: Shape must be rank 2 but is rank 1 for 'MatMul' (op: 'MatMul') with input shapes: [2048], [2048,1008].

解决办法：使用 ： logits = tf.matmul(tf.squeeze(pool3, [1, 2]), w)   代替： logits = tf.matmul(tf.squeeze(pool3), w)

参考：ValueError: Shape must be rank 2 but is rank 1 for 'MatMul' (op: 'MatMul') with input shapes: [2048]

或者： Shape must be rank 2 but is rank 1 for 'MatMul'

6. opencv为4.1版本，img = cv2.imread(img_path)，module cv2 has no imread member

分析： 其实代码是没有错误的,,可以直接运行的,就是cscode的插件检测不到模块而已, 主要原因是cv2模块下还有cv2模块,简单的让vscode不报错的方法是在添加包的时候添加

解决办法：  from cv2 import cv2

参考：解决vscode报错: Module 'cv2' has no 'imshow' member

7、 img = cv2.imread(img_path)     h, w, c = img.shape

Error：    h, w, c = img.shape
AttributeError: 'NoneType' object has no attribute 'shape'

分析 ：经过检查发现是图像路径中图像不存在

解决办法： 把图像路径填写正确就 OK了 

 参考： python图像处理问题AttributeError: 'NoneType' object has no attribute 'shape'

 8、Ubuntu18.04的系统，使用nvidia-smi就出来详细显卡信息了，但是过了段时间，再使用指令就出现这样的提示：NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.

问题分析：
NVIDIA驱动是在低版本的时候安装，由于系统更新，内核版本升级导致NVIDIA无法应用在高内核版本，所以，说明是内核版本的问题，选择低版本的内核即可。

 参考：https://blog.csdn.net/sinat_23619409/article/details/85220561