list、numpy.ndarray 和 pandas.dataframe的相互转换、torch.tensor和numpy.ndarray的相互转换

1、list 转 numpy.ndarray:   np.array()
2、numpy.ndarray 转 pandas.Dataframe： pandas.DataFrame()
3、pandas.Dataframe 转 numpy.ndarray： Dataframe.values
4、numpy.ndarray 转 list:  array.tolist()

示例：

import numpy as np
import pandas as pd

a = [[1, 2, 3],[4, 5, 6]] # list
b = np.array(a) # list to ndarray
c = pd.DataFrame(b) # ndarray to dataframe

print('type a:',type(a),'\ntype b:',type(b),'\ntype c:',type(c))

d = c.values #  dataframe to ndarray
e = d.tolist() # ndarray to list
print('type d:',type(d),'\ntype e:',type(e))

输出结果：

type a: <class 'list'>
type b: <class 'numpy.ndarray'>
type c: <class 'pandas.core.frame.DataFrame'>
type d: <class 'numpy.ndarray'>
type e: <class 'list'>

---

numpy.ndarray 与 torch.tensor 的相互转换：

import torch
import numpy as np

a = np.ones(10) # array

b = torch.from_numpy(a).type(torch.FloatTensor) # array to tensor

c = b.numpy()  # tensor to array 

print('type a:',type(a),'\ntype b:',type(b),'\ntype c:',type(c))

输出结果：

type a: <class 'numpy.ndarray'>
type b: <class 'torch.Tensor'>
type c: <class 'numpy.ndarray'>

CUDA中的tensor转numpy：

np_array = tensor.data.cpu().numpy()