在神经网络中,我有一些 2D 特征图,其值介于 0 和 1 之间。对于这些图,我想根据每个坐标的值计算协方差矩阵。不幸的是,pytorch 没有.cov()
像 numpy 那样的功能。所以我改写了以下函数:
def get_covariance(tensor):
bn, nk, w, h = tensor.shape
tensor_reshape = tensor.reshape(bn, nk, 2, -1)
x = tensor_reshape[:, :, 0, :]
y = tensor_reshape[:, :, 1, :]
mean_x = torch.mean(x, dim=2).unsqueeze(-1)
mean_y = torch.mean(y, dim=2).unsqueeze(-1)
xx = torch.sum((x - mean_x) * (x - mean_x), dim=2).unsqueeze(-1) / (h * w - 1)
xy = torch.sum((x - mean_x) * (y - mean_y), dim=2).unsqueeze(-1) / (h * w - 1)
yx = xy
yy = torch.sum((y - mean_y) * (y - mean_y), dim=2).unsqueeze(-1) / (h * w - 1)
cov = torch.cat((xx, xy, yx, yy), dim=2)
cov = cov.reshape(bn, nk, 2, 2)
return cov
这是正确的方法吗?
编辑:
这是与numpy函数的比较:
a = torch.randn(1, 1, 64, 64)
a_numpy = a.reshape(1, 1, 2, -1).numpy()
torch_cov = get_covariance(a)
numpy_cov = np.cov(a_numpy[0][0])
torch_cov
tensor([[[[ 0.4964, -0.0053],
[-0.0053, 0.4926]]]])
numpy_cov
array([[ 0.99295635, -0.01069122],
[-0.01069122, 0.98539236]])
显然,我的值太小了 2 倍。为什么会这样?
Edit2:啊,我想通了。它必须除以(h*w/2 - 1)
:) 然后值匹配。