In [ ]:
import numpy as np
import matplotlib.pyplot as plt
純量(0階tensor)¶
In [ ]:
x = np.array([[9], [4], [8], [7]]) #一維陣列建立
In [ ]:
x
Out[ ]:
array([[9],
[4],
[8],
[7]])
In [ ]:
x.shape
Out[ ]:
(4, 1)
向量(1階tensor)¶
In [ ]:
x = np.array([[9, 4, 8 , 7], [8, 7, 4, 5]]) #二維陣列建立
In [ ]:
x
Out[ ]:
array([[9, 4, 8, 7],
[8, 7, 4, 5]])
In [ ]:
x.shape
Out[ ]:
(2, 4)
矩陣(2階tensor)¶
In [ ]:
# 矩陣
a = x[0]
b = x[1]
In [ ]:
a
Out[ ]:
array([9, 4, 8, 7])
In [ ]:
a.reshape(2,2)
Out[ ]:
array([[9, 4],
[8, 7]])
In [ ]:
a = a.reshape(2,2)
b = b.reshape(2,2)
In [ ]:
x = np.array([a,b])
In [ ]:
x
Out[ ]:
array([[[9, 4],
[8, 7]],
[[8, 7],
[4, 5]]])
In [ ]:
x.shape
Out[ ]:
(2, 2, 2)
one hot encoding¶
In [ ]:
#one hot encoding
from tensorflow.keras.utils import to_categorical
In [ ]:
x=[0,1,2,3,4,5]
In [ ]:
# 類別個數:6
x_1 = to_categorical(x,6)
x_1
Out[ ]:
array([[1., 0., 0., 0., 0., 0.],
[0., 1., 0., 0., 0., 0.],
[0., 0., 1., 0., 0., 0.],
[0., 0., 0., 1., 0., 0.],
[0., 0., 0., 0., 1., 0.],
[0., 0., 0., 0., 0., 1.]], dtype=float32)