import random
import torch
from torch import nn, optim
import math
from IPython import display
from plot_lib import plot_data, plot_model, set_default
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
print('device: ', device)
# 初始化随机数种子
seed = 12345
random.seed(seed)
torch.manual_seed(seed)
N = 1000 # 每类样本的数量
D = 2 # 每个样本的特征维度
C = 3 # 样本的类别
H = 100 # 神经网络里隐层单元的数量
X = torch.zeros(N * C, D).to(device)
Y = torch.zeros(N * C, dtype=torch.long).to(device)
for c in range(C):
index = 0
t = torch.linspace(0, 1, N) # 在[0,1]间均匀的取10000个数,赋给t
inner_var = torch.linspace( (2*math.pi/C)*c, (2*math.pi/C)*(2+c), N) + torch.randn(N) * 0.2
# 每个样本的(x,y)坐标都保存在 X 里
# Y 里存储的是样本的类别,分别为 [0, 1, 2]
for ix in range(N * c, N * (c + 1)):
X[ix] = t[index] * torch.FloatTensor((math.sin(inner_var[index]), math.cos(inner_var[index])))
Y[ix] = c
index += 1
print("Shapes:")
print("X:", X.size())
print("Y:", Y.size())