def shuffle_sets(self, samples, targets):
rng_state = np.random.get_state()
copy_samples = np.empty((np.shape(samples)))
copy_samples[:] = samples
np.random.shuffle(copy_samples)
np.random.set_state(rng_state)
copy_targets = np.empty((np.shape(targets)))
copy_targets[:] = targets
np.random.shuffle(copy_targets)
return copy_samples[:self.SAMPLES,:], copy_targets[:self.SAMPLES,:]
def cpd_G(x, y, beta):
'''
Constructs gaussian affinity matrix
:param x: goal positions
:param y: current positions
:param beta: std of G
:return: Gaussian affinity matrix G
'''
k = -2 * beta**2
[n, d] = cp.shape(x)
[m, d] = cp.shape(y)
G = cp.tile(x, [m, 1, 1]) - cp.reshape(
cp.repeat(y[:, :, cp.newaxis], n, axis=0), [m, n, d])
G = cp.sum(cp.power(G, 2), 2)
G = G / k
return cp.exp(G)
def __pca(self, XMat, k):
average = np.mean(XMat, axis=0)
m, n = np.shape(XMat) # m: samples_nums, n: features
data_adjust = XMat - average
cov_x = self.__cov(data_adjust.T) # 计算协方差矩阵
feat_value, feat_vec = npy.linalg.eig(cov_x.get()) # 求解协方差矩阵的特征值和特征向量
index = np.argsort(-feat_value) # 依照featValue进行从大到小排序
if k > n:
print("k must lower than feature number")
return
else:
selectVec = feat_vec[:, index[:k]] # 所以这里须要进行转置
finalData = np.matmul(data_adjust, np.asarray(selectVec)) # cupy
return finalData, np.asarray(selectVec)
def _truncnorm_ppf(self,q, N):
out = cp.zeros(cp.shape(q))
delta = self._truncnorm_get_delta(N)
cond1 = delta > 0
cond2 = (delta > 0) & (self.a > 0)
cond21 = (delta > 0) & (self.a<=0)
if cp.any(cond1) == True:
sa = self.norm_sf(a[cond2])
out[:,cond2] = -self.ndtri((1 - q[:,cond2]) * sa)
if cp.any(cond21) == True:
na = norm_cdf(self.a[cond21])
out[:,cond21] = self._ndtri(q[:,cond21] + na * (1.0 - q[:,cond21]))
cond3 = ~cond1 & cp.isinf(self.b)
cond4 = ~cond1 & cp.isinf(self.a)
if cp.any(cond3) == True:
out[:,cond3] = -self._norm_ilogcdf(cp.log1p(-q[:,cond3]) + self.norm_logsf(self.a[cond3]))
if cp.any(cond4) == True:
out[:,cond4] = self._norm_ilogcdf(cp.log(q) + self.norm_logcdf(self.b))
cond5 = out < self.a
if cp.any(cond5) == True:
out[cond5] = ((cond5) * self.a)[cond5]
return out
def test_shape_list(self):
shape = self.shape
a = testing.shaped_arange(shape, numpy)
a = a.tolist()
assert cupy.shape(a) == shape
def test_shape(self):
shape = self.shape
for xp in (numpy, cupy):
a = testing.shaped_arange(shape, xp)
assert cupy.shape(a) == shape