YDOOK AI : Pytorch : tensor 与 numpy 的 ndarray 相互转化 pytorch 张量与 numpy 数组之间转化

YDOOK AI : Pytorch : tensor 与 numpy 的 ndarray 相互转化 pytorch 张量与 numpy 数组之间转化

1. 转换方法：

1. tensor => ndarray : tensor.numpy()

2. ndarray => tensor : tensor = torch.from_numpy(ndarray)

---

2. 源代码展示：

import torch
import numpy as np

# YDOOK JY Lin
print('torch.__version__ = ', torch.__version__)

# ndarray => tensor
np_a = np.array([1, 2, 3])
print('np_a = ', np_a)
print('type(np_a) = ', type(np_a))
print()

tensor_a = torch.from_numpy(np_a)
print('tensor_a = ', tensor_a)
print('type(tensor_a) = ', type(tensor_a))
print()

# tensor => ndarray
t = torch.tensor([4, 5, 6])
print('t = ', t)
print('type(t) = ', type(t))

np_t = t.numpy()
print('np_t = ', np_t)
print('type(np_t) = ', type(np_t))

---

3. 输出：

torch.__version__ =  1.7.1
np_a =  [1 2 3]
type(np_a) =  <class 'numpy.ndarray'>
tensor_a =  tensor([1, 2, 3], dtype=torch.int32)
type(tensor_a) =  <class 'torch.Tensor'>
t =  tensor([4, 5, 6])
type(t) =  <class 'torch.Tensor'>
np_t =  [4 5 6]
type(np_t) =  <class 'numpy.ndarray'>

---

---

---