def __init__(self,rep_in,rep_out,group,ch=384,num_layers=3):
super().__init__()
self.rep_in =rep_in(group)
self.rep_out = rep_out(group)
self.G = group
chs = [self.rep_in.size()] + num_layers*[ch]
cout = self.rep_out.size()
logging.info("Initing MLP")
self.net = Sequential(
*[Sequential(nn.Linear(cin,cout),swish) for cin,cout in zip(chs,chs[1:])],
nn.Linear(chs[-1],cout)
)
def __init__(self,
num_blocks: list,
num_classes: int = 1000,
width_multiplier: Optional[list] = None,
override_groups_map=None,
deploy=False):
super(RepVGG, self).__init__()
assert len(width_multiplier) == 4
self.deploy = deploy
self.override_groups_map = override_groups_map or dict()
assert 0 not in self.override_groups_map
self.in_planes = min(64, int(64 * width_multiplier[0]))
self.stage0 = RepVGGBlock(in_channels=3,
out_channels=self.in_planes,
kernel_size=3,
stride=2,
padding=1,
deploy=self.deploy)
self.cur_layer_idx = 1
self.stage1 = self._make_stage(int(64 * width_multiplier[0]),
num_blocks[0],
stride=2)
self.stage2 = self._make_stage(int(128 * width_multiplier[1]),
num_blocks[1],
stride=2)
self.stage3 = self._make_stage(int(256 * width_multiplier[2]),
num_blocks[2],
stride=2)
self.stage4 = self._make_stage(int(512 * width_multiplier[3]),
num_blocks[3],
stride=2)
self.gap = F.average_pool_2d
self.linear = nn.Linear(int(512 * width_multiplier[3]), num_classes)
def MLPBlock(cin,cout):
return Sequential(nn.Linear(cin,cout),swish)#,nn.BatchNorm0D(cout,momentum=.9),swish)#,