[英]What is the issue with my last dense keras layer?
I am working on a small NN in keras for multi-class classification problem.我正在研究 keras 中的一个小型 NN,用于解决多类分类问题。 I have 9 different labels and my features are also 9.
我有 9 个不同的标签,我的特征也是 9 个。
My train/test shapes are the following:我的火车/测试形状如下:
Sets shape:
x_train shape: (7079, 9)
y_train shape: (7079,)
x_test shape: (7079, 9)
y_test shape: (7079,)
But when I try to make them categorical:但是当我试图让它们分类时:
y_train = tf.keras.utils.to_categorical(y_train, num_classes=9)
y_test = tf.keras.utils.to_categorical(y_test, num_classes=9)
I get the following error:我收到以下错误:
IndexError: index 9 is out of bounds for axis 1 with size 9
Here is more info about the y_train
这是有关
y_train
的更多信息
print(np.unique(y_train)) # [1. 2. 3. 4. 5. 6. 7. 8. 9.]
print(len(np.unique(y_train))) # 9
Anyone would know what the problem is?任何人都会知道问题是什么?
The shape of the y_train
is 1D
. y_train
的形状是1D
。 You have to make it one-hot encoded.您必须对其进行一次热编码。 Something like
就像是
y_train = tf.keras.utils.to_categorical(y_train , num_classes=9)
And same goes for y_test
too. y_test
也是如此。
tf.keras.utils.to_categorical(y, num_classes=None, dtype="float32")
Here, y : class vector to be converted into a matrix (integers from 0
to num_classes
).这里, y : class 向量被转换成矩阵(从
0
到num_classes
的整数)。 As in your case, y_train
is something like [1,2,..]
.与您的情况一样,
y_train
类似于[1,2,..]
。 You need to do as follows:您需要执行以下操作:
y_train = tf.keras.utils.to_categorical(y_train - 1, num_classes=9)
Here is an example for reference.这是一个供参考的例子。 If we do
如果我们这样做
class_vector = np.array([1, 1, 2, 3, 5, 1, 4, 2])
print(class_vector)
output_matrix = tf.keras.utils.to_categorical(class_vector,
num_classes = 5, dtype ="float32")
print(output_matrix)
[1 1 2 3 5 1 4 2]
---------------------------------------------------------------------------
IndexError Traceback (most recent call last)
<ipython-input-15-69c8be7a0f1a> in <module>()
6 print(class_vector)
7
----> 8 output_matrix = tf.keras.utils.to_categorical(class_vector, num_classes = 5, dtype ="float32")
9 print(output_matrix)
/usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/utils/np_utils.py in to_categorical(y, num_classes, dtype)
76 n = y.shape[0]
77 categorical = np.zeros((n, num_classes), dtype=dtype)
---> 78 categorical[np.arange(n), y] = 1
79 output_shape = input_shape + (num_classes,)
80 categorical = np.reshape(categorical, output_shape)
IndexError: index 5 is out of bounds for axis 1 with size 5
To solve this, we convert the data to a zero-based format.为了解决这个问题,我们将数据转换为从零开始的格式。
output_matrix = tf.keras.utils.to_categorical(class_vector - 1,
num_classes = 5, dtype ="float32")
print(output_matrix)
[[1. 0. 0. 0. 0.]
[1. 0. 0. 0. 0.]
[0. 1. 0. 0. 0.]
[0. 0. 1. 0. 0.]
[0. 0. 0. 0. 1.]
[1. 0. 0. 0. 0.]
[0. 0. 0. 1. 0.]
[0. 1. 0. 0. 0.]]
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.