def network_initialization(
self, in_channel, out_channel, channels, embedding_channel, kernel_size, D=2,
):
self.conv1_1 = self.get_conv_block(in_channel, channels[0], kernel_size=kernel_size, stride=1)
self.conv1_2 = self.get_conv_block(channels[0], channels[1], kernel_size=kernel_size, stride=1)
self.pool1 = ME.MinkowskiMaxPooling(kernel_size=3, stride=2, dimension=D)
self.conv2_1 = self.get_conv_block(channels[1], channels[2], kernel_size=kernel_size, stride=1)
self.conv2_2 = self.get_conv_block(channels[2], channels[3], kernel_size=kernel_size, stride=1)
self.pool2 = ME.MinkowskiMaxPooling(kernel_size=3, stride=2, dimension=D)
self.conv3_1 = self.get_conv_block(channels[3], channels[4], kernel_size=kernel_size, stride=1)
self.conv3_2 = self.get_conv_block(channels[4], channels[5], kernel_size=kernel_size, stride=1)
self.conv3_3 = self.get_conv_block(channels[5], channels[6], kernel_size=kernel_size, stride=1)
self.pool3 = ME.MinkowskiMaxPooling(kernel_size=3, stride=2, dimension=D)
self.conv4_1 = self.get_conv_block(channels[6], channels[7], kernel_size=kernel_size, stride=1)
self.conv4_2 = self.get_conv_block(channels[7], channels[8], kernel_size=kernel_size, stride=1)
self.conv4_3 = self.get_conv_block(channels[8], channels[9], kernel_size=kernel_size, stride=1)
self.pool4 = ME.MinkowskiMaxPooling(kernel_size=3, stride=2, dimension=D)
self.conv5_1 = self.get_conv_block(channels[9], channels[10], kernel_size=kernel_size, stride=1)
self.conv5_2 = self.get_conv_block(channels[10], channels[11], kernel_size=kernel_size, stride=1)
self.conv5_3 = self.get_conv_block(channels[11], channels[12], kernel_size=kernel_size, stride=1)
self.pool5 = ME.MinkowskiMaxPooling(kernel_size=3, stride=2, dimension=D)
self.global_pool = ME.MinkowskiGlobalPooling()
self.final = nn.Sequential(
self.get_mlp_block(512, 512),
ME.MinkowskiDropout(),
self.get_mlp_block(512, 512),
self.get_mlp_block(512, 2),
#ME.MinkowskiFunctional.softmax(),
)
def __init__(self,
en_ch=[3, 64, 128, 256],
de_ch=[256, 128, 32, 1],
D=3,
lang_feat_dim=256):
super().__init__()
self.en1 = nn.Sequential(
ME.MinkowskiConvolution(en_ch[0],
en_ch[1],
kernel_size=3,
dimension=D),
ME.MinkowskiReLU(inplace=True),
ME.MinkowskiMaxPooling(kernel_size=2, stride=2, dimension=D)) # 15
self.en2 = nn.Sequential(
ME.MinkowskiConvolution(en_ch[1],
en_ch[2],
kernel_size=3,
dimension=D),
ME.MinkowskiReLU(inplace=True),
ME.MinkowskiMaxPooling(kernel_size=2, stride=2, dimension=D))
self.en3 = nn.Sequential(
ME.MinkowskiConvolution(en_ch[2],
en_ch[3],
kernel_size=3,
dimension=D),
ME.MinkowskiReLU(inplace=True),
ME.MinkowskiMaxPooling(kernel_size=2, stride=2, dimension=D))
self.fuse_layer = nn.Sequential(
ME.MinkowskiConvolution(en_ch[3] + lang_feat_dim,
en_ch[3] + lang_feat_dim,
kernel_size=1,
dimension=D),
ME.MinkowskiReLU(inplace=True),
ME.MinkowskiConvolution(en_ch[3] + lang_feat_dim,
en_ch[3] + lang_feat_dim // 2,
kernel_size=1,
dimension=D),
ME.MinkowskiReLU(inplace=True),
ME.MinkowskiConvolution(en_ch[3] + lang_feat_dim // 2,
en_ch[3],
kernel_size=1,
dimension=D),
ME.MinkowskiReLU(inplace=True))
self.de3 = upblock(down_channels=en_ch[3],
skip_channels=en_ch[2],
out_channels=de_ch[0])
self.de2 = upblock(down_channels=de_ch[0],
skip_channels=en_ch[1],
out_channels=de_ch[1])
self.de1 = upblock(down_channels=de_ch[1],
skip_channels=en_ch[0],
out_channels=de_ch[2])
self.final_layer = nn.Sequential(
ME.MinkowskiConvolution(de_ch[2],
de_ch[3],
kernel_size=1,
dimension=D), )
def network_initialization(self, in_channels, out_channels, D):
self.conv1 = ME.MinkowskiConvolution(in_channels,
8,
kernel_size=3,
dimension=D)
self.conv2 = ME.MinkowskiConvolution(8, 8, kernel_size=3, dimension=D)
self.pool = ME.MinkowskiMaxPooling(kernel_size=3,
stride=2,
dimension=D)
self.conv3 = ME.MinkowskiConvolution(8, 16, kernel_size=3, dimension=D)
self.conv4 = ME.MinkowskiConvolution(16,
16,
kernel_size=3,
dimension=D)
def network_initialization(self, in_channels, out_channels, D):
self.inplanes = self.INIT_DIM
self.conv1 = nn.Sequential(
ME.MinkowskiConvolution(in_channels,
self.inplanes,
kernel_size=3,
stride=2,
dimension=D),
ME.MinkowskiBatchNorm(self.inplanes),
ME.MinkowskiReLU(inplace=True),
ME.MinkowskiMaxPooling(kernel_size=2, stride=2, dimension=D),
)
self.layer1 = self._make_layer(self.BLOCK,
self.PLANES[0],
self.LAYERS[0],
stride=2)
self.layer2 = self._make_layer(self.BLOCK,
self.PLANES[1],
self.LAYERS[1],
stride=2)
self.layer3 = self._make_layer(self.BLOCK,
self.PLANES[2],
self.LAYERS[2],
stride=2)
self.layer4 = self._make_layer(self.BLOCK,
self.PLANES[3],
self.LAYERS[3],
stride=2)
self.conv5 = nn.Sequential(
ME.MinkowskiDropout(),
ME.MinkowskiConvolution(self.inplanes,
self.inplanes,
kernel_size=3,
stride=3,
dimension=D),
ME.MinkowskiBatchNorm(self.inplanes),
ME.MinkowskiGELU(),
)
self.glob_pool = ME.MinkowskiGlobalMaxPooling()
self.final = ME.MinkowskiLinear(self.inplanes, out_channels, bias=True)
def network_initialization(
self, in_channel, out_channel, channels, embedding_channel, kernel_size, D=3,
):
self.mlp1 = self.get_mlp_block(in_channel, channels[0])
self.conv1 = self.get_conv_block(
channels[0], channels[1], kernel_size=kernel_size, stride=1,
)
self.conv2 = self.get_conv_block(
channels[1], channels[2], kernel_size=kernel_size, stride=2,
)
self.conv3 = self.get_conv_block(
channels[2], channels[3], kernel_size=kernel_size, stride=2,
)
self.conv4 = self.get_conv_block(
channels[3], channels[4], kernel_size=kernel_size, stride=2,
)
self.conv5 = nn.Sequential(
self.get_conv_block(
channels[1] + channels[2] + channels[3] + channels[4],
embedding_channel // 4,
kernel_size=3,
stride=2,
),
self.get_conv_block(
embedding_channel // 4, embedding_channel // 2, kernel_size=3, stride=2,
),
self.get_conv_block(
embedding_channel // 2, embedding_channel, kernel_size=3, stride=2,
),
)
self.pool = ME.MinkowskiMaxPooling(kernel_size=3, stride=2, dimension=D)
self.global_max_pool = ME.MinkowskiGlobalMaxPooling()
self.global_avg_pool = ME.MinkowskiGlobalAvgPooling()
self.final = nn.Sequential(
self.get_mlp_block(embedding_channel * 2, 512),
ME.MinkowskiDropout(),
self.get_mlp_block(512, 512),
ME.MinkowskiLinear(512, out_channel, bias=True),
)
def max_pool(kernel_size,
stride=1,
dilation=1,
conv_type=ConvType.HYPERCUBE,
in_coords_key=None,
D=-1):
assert D > 0, 'Dimension must be a positive integer'
region_type, axis_types, kernel_size = convert_conv_type(
conv_type, kernel_size, D)
kernel_generator = ME.KernelGenerator(kernel_size,
stride,
dilation,
region_type=region_type,
axis_types=axis_types,
dimension=D)
return ME.MinkowskiMaxPooling(kernel_size=kernel_size,
stride=stride,
dilation=dilation,
kernel_generator=kernel_generator,
dimension=D)
def __init__(self,
in_planes,
out_planes,
kernel_size,
stride,
with_uncertainty=False,
add_rgb=False):
super().__init__()
self.layer3 = nn.Sequential(
ME.MinkowskiConvolution(in_planes,
out_planes * 2,
kernel_size=kernel_size,
stride=1,
dimension=2),
ME.MinkowskiBatchNorm(out_planes * 2),
ME.MinkowskiReLU(inplace=True),
ME.MinkowskiConvolution(out_planes * 2,
out_planes * 2,
kernel_size=kernel_size,
stride=1,
dimension=2),
ME.MinkowskiBatchNorm(out_planes * 2),
ME.MinkowskiReLU(inplace=True),
ME.MinkowskiConvolution(out_planes * 2,
out_planes,
kernel_size=kernel_size,
stride=1,
dimension=2),
)
self.layer2 = nn.Sequential(
ME.MinkowskiConvolution(in_planes,
out_planes * 2,
kernel_size=kernel_size,
stride=1,
dimension=2),
ME.MinkowskiBatchNorm(out_planes * 2),
ME.MinkowskiReLU(inplace=True),
ME.MinkowskiConvolution(out_planes * 2,
out_planes,
kernel_size=kernel_size,
stride=1,
dimension=2),
)
self.layer1 = nn.Sequential(
ME.MinkowskiConvolution(in_planes,
out_planes,
kernel_size=kernel_size,
stride=1,
dimension=2), )
self.layer_final = nn.Sequential(ME.MinkowskiBatchNorm(out_planes),
ME.MinkowskiReLU(inplace=True))
self.pool = None if stride == 1 else ME.MinkowskiMaxPooling(
3, stride, dimension=2)
self.add_rgb = add_rgb
self.with_uncertainty = with_uncertainty
if with_uncertainty:
self.unc_layer = nn.Sequential(
ME.MinkowskiConvolution(out_planes,
1,
kernel_size=3,
stride=1,
dimension=2), ME.MinkowskiSigmoid())
# coords = torch.rand([N_voxel])*10
# coords = torch.cat([torch.zeros(N_voxel,1), coords], dim=1)
coords = torch.tensor([[0, 2., 4., 2.],
[0, 1., 3., 1.],
# [0, 0., 3., 1.]
[0, 4., 6., 3.]
# [0, 1., 5., 1.]
])
conv1 = ME.MinkowskiConvolution(1,1,kernel_size=1,dimension=3)
convt1 = ME.MinkowskiConvolutionTranspose(1,1,kernel_size=1,dimension=3)
conv2 = nn.Conv2d(1,1,kernel_size=1,bias=False)
conv_s2 = ME.MinkowskiConvolution(1,1,kernel_size=2,stride=2,dimension=3)
pool_s2 = ME.MinkowskiMaxPooling(2,stride=2,dimension=3)
for n, m in conv1.named_parameters():
print(n,m)
conv2.weight = nn.Parameter(m.reshape(1,1,1,1))
x1 = ME.SparseTensor(coordinates=coords, features=feats.reshape([-1,1]))\
y1 = conv1(x1)
y1t = convt1(x1) # different weight, but they are equal
x2 = ME.SparseTensor(coordinates=coords, features=feats2.reshape([-1,1]))
x1_s2 = conv_s2(x1)
x1_p2 = pool_s2(x1)