model/decoder.py

import torch 
import torch.nn as nn  
import torch.functional as F
from bn import batch_norm
from residual import residual
class decoder(nn.Module):
    def __init__(self, inp, out):
        super().__init__()
        self.upsample = nn.Upsample(scale_factor=2, mode = 'bilinear', align_corners = True)
        self.block = residual(inp+out, out)
    def forward(self, x, skip):
        x = self.upsample(x)
        x = torch.cat([x, skip], axis = 1)
        x = self.block(x)
        return x