|
|
|
|
|
|
|
|
|
|
|
|
|
from copy import deepcopy |
|
import torch |
|
import os |
|
from packaging import version |
|
import torch.nn as nn |
|
import dust3r.utils.path_to_croco |
|
inf = float('inf') |
|
import torch.nn as nn |
|
import torch.nn.functional as F |
|
|
|
class ConvBnReLU(nn.Module): |
|
def __init__(self, in_channels, out_channels, |
|
kernel_size=3, stride=1, pad=1): |
|
super(ConvBnReLU, self).__init__() |
|
self.conv = nn.Conv2d(in_channels, out_channels, |
|
kernel_size, stride=stride, padding=pad, bias=False) |
|
self.relu = nn.ReLU(inplace=True) |
|
def forward(self, x): |
|
return self.relu(self.conv(x)) |
|
|
|
class FeatureNet(nn.Module): |
|
def __init__(self, norm_act=nn.BatchNorm2d): |
|
super(FeatureNet, self).__init__() |
|
self.conv0 = nn.Sequential( |
|
ConvBnReLU(3, 8, 3, 1, 1), |
|
ConvBnReLU(8, 8, 3, 1, 1)) |
|
self.conv1 = nn.Sequential( |
|
ConvBnReLU(8, 16, 5, 2, 2), |
|
ConvBnReLU(16, 16, 3, 1, 1)) |
|
self.conv2 = nn.Sequential( |
|
ConvBnReLU(16, 32, 5, 2, 2), |
|
ConvBnReLU(32, 32, 3, 1, 1)) |
|
|
|
self.toplayer = nn.Conv2d(32, 32, 1) |
|
self.lat1 = nn.Conv2d(16, 32, 1) |
|
self.lat0 = nn.Conv2d(8, 32, 1) |
|
|
|
self.smooth1 = nn.Conv2d(32, 32, 3, padding=1) |
|
self.smooth0 = nn.Conv2d(32, 32, 3, padding=1) |
|
|
|
def _upsample_add(self, x, y): |
|
return F.interpolate(x, scale_factor=2, mode='bilinear', align_corners=True) + y |
|
|
|
def forward(self, x): |
|
conv0 = self.conv0(x) |
|
conv1 = self.conv1(conv0) |
|
conv2 = self.conv2(conv1) |
|
feat2 = self.toplayer(conv2) |
|
feat1 = self._upsample_add(feat2, self.lat1(conv1)) |
|
feat0 = self._upsample_add(feat1, self.lat0(conv0)) |
|
feat1 = self.smooth1(feat1) |
|
feat0 = self.smooth0(feat0) |
|
return feat2, feat1, feat0 |
|
|
