cov = [[0.02, 0], [0, 0.02]]
K = 50
x, y = np.random.multivariate_normal(mean, cov, K).T
x = np.stack((x, y), axis=1).clip(-0.99, 0.99)
use_gpu = False
if use_gpu and torch.cuda.is_available():
device = 'cuda:0'
else:
device = 'cpu'
x1 = torch.from_numpy(x1)
x2 = torch.from_numpy(x2)
x = torch.from_numpy(x)
vwb = VWB(x, [x1, x2], device=device, verbose=False)
output = vwb.cluster(max_iter_h=5000, max_iter_p=1)
idx = output['idx']
xmin, xmax, ymin, ymax = -1.0, 1.0, -0.5, 0.5
for k in [33]:
plt.figure(figsize=(8, 4))
for i in range(2):
ce = np.array(plt.get_cmap('viridis')(idx[i].cpu().numpy() / (K - 1)))
utils.scatter_otsamples(vwb.data_p,
vwb.data_e[i],
size_p=30,
marker_p='o',
color_x=ce,
xmin=xmin,
fig6.savefig("x3_histogram_kmeans1.png", dpi=300, bbox_inches='tight')
# -------------- VWB ------------------ #
use_gpu = False
if use_gpu and torch.cuda.is_available():
device = 'cuda:0'
else:
device = 'cpu'
x1 = torch.from_numpy(x1.clip(-0.99, 0.99))
x2 = torch.from_numpy(x2.clip(-0.99, 0.99))
x3 = torch.from_numpy(x3.clip(-0.99, 0.99))
x = torch.from_numpy(C1_16.clip(-0.99, 0.99))
vwb = VWB(x, [x1], device=device, verbose=False)
output = vwb.cluster(lr=1, max_iter_h=5000, max_iter_p=1)
e_idx = output['idx']
x1_vwb_16 = vwb.data_p[e_idx[0], :].cpu().numpy()
x1_vwb_16 = np.transpose(np.reshape(x1_vwb_16 * 255, (128, 128, 3)), (1, 0, 2))
imageio.imwrite("x1_vwb_16.png", x1_vwb_16)
fig1 = plt.figure(figsize=(8, 8))
ax1 = fig1.add_subplot(111, projection='3d')
ce2 = vwb.data_p[e_idx].numpy()
p11 = vwb.data_p.numpy()
ax1.scatter(x1[:, 0], x1[:, 1], x1[:, 2], s=dot_size, color=ce2, alpha=alpha)
ax1.xaxis.pane.fill = False
ax1.yaxis.pane.fill = False
ax1.zaxis.pane.fill = False
ax1.scatter(p11[:, 0],
p11[:, 1],
x1 = torch.from_numpy(x1).double().to(device)
x2 = torch.from_numpy(x2).double().to(device)
wd = np.zeros((50, 50))
k = 0
for K in [50, 125, 250, 500, 750, 1000, 1500, 2500]:
for i in range(10):
mean = [0.0, -0.5]
cov = [[0.02, 0], [0, 0.02]]
x, y = np.random.multivariate_normal(mean, cov, K).T
x = np.stack((x, y), axis=1).clip(-0.99, 0.99)
x = torch.from_numpy(x).double().to(device)
vwb = VWB(x, [x1, x2], device=device, verbose=False)
output = vwb.cluster(lr=1,
max_iter_h=10000,
max_iter_p=1,
beta=0.9,
lr_decay=500)
wd[k, i] = output['wd']
print(output['wd'])
# np.savetxt(f, output['wd'], delimiter=',')
del x
del vwb
k += 1
# np.savetxt('ship_error/K_total.csv', wd, delimiter=',')
import ot
use_gpu = False
if use_gpu and torch.cuda.is_available():
device = 'cuda:0'
else:
device = 'cpu'
# ---------------VWB---------------
x = newx
for reg in [0.5, 2, 1e9]:
x_copy = torch.from_numpy(x)
x0_copy = torch.from_numpy(x0)
vwb = VWB(x_copy, [x0_copy], device=device, verbose=False)
output = vwb.cluster(lr=0.5,
max_iter_h=1000,
max_iter_p=1,
beta=0.5,
reg=reg)
e_idx = output['idx']
fig = plt.figure(figsize=(4, 4))
ce = color_map[e_idx[0]]
utils.scatter_otsamples(vwb.data_p,
vwb.data_e[0],
size_p=30,
marker_p='o',
color_e=ce,
cur_axes = plt.gca()
cur_axes.axes.get_xaxis().set_ticks([])
cur_axes.axes.get_yaxis().set_ticks([])
# plt.savefig("euclidean_center.svg", bbox_inches='tight')
plt.savefig("euclidean_center.png", dpi=300, bbox_inches='tight')
# VWB
use_gpu = False
if use_gpu and torch.cuda.is_available():
device = 'cuda:0'
else:
device = 'cpu'
ell_all = [torch.from_numpy(ell).t() for ell in ell_all]
vwb = VWB(ell_all[0][0:-1:5, :], ell_all, device=device, verbose=False)
output = vwb.cluster(lr=0.5, max_iter_h=3000, max_iter_p=1)
e_idx = output['idx']
fig = plt.figure(4, figsize=(2, 2))
ax = fig.add_subplot(111)
plt.grid(True)
for i in range(N):
plt.plot(vwb.data_e[i][:, 0],
vwb.data_e[i][:, 1],
linewidth=5,
c='lightgray')
size = vwb.data_p.shape[0] // 2
ax.set_prop_cycle(