这是一段神经网络代码示例:
def forward_step(X, W, b, W2, b2):
hidden_layer = np.maximum(0, np.dot(X, W) + b)
scores = np.dot(hidden_layer, W2) + b2
exp_scores = np.exp(scores)
probs = exp_scores / np.sum(exp_scores, axis=1, keepdims=True)
...
上面显示的代码的最后一行引发了错误:
<ipython-input-49-d97cff51c360> in forward_step(X, W, b, W2, b2)
14 scores = np.dot(hidden_layer, W2) + b2
15 exp_scores = np.exp(scores)
---> 16 probs = exp_scores / np.sum(exp_scores, axis=1, keepdims=True)
17 corect_logprobs = -np.log(probs[range(X.shape[0]), y])
/Users/###/anaconda/lib/python3.6/site-packages/numpy/core/fromnumeric.py in sum(a, axis, dtype, out, keepdims)
1810 pass
1811 else:
-> 1812 return sum(axis=axis, dtype=dtype, out=out, **kwargs)
1813 return _methods._sum(a, axis=axis, dtype=dtype,
1814 out=out, **kwargs)
TypeError: sum() got an unexpected keyword argument 'keepdims'
有一个类似的问题Numpy sum keepdims error说 numpy 的版本应该大于 1.7。我检查了我的 numpy 版本:
import numpy
numpy.version.version
>> 1.12.1
现在我对这个错误是如何发生的感到困惑。