DeOldify/fasterai/unet.py

from fastai.layers import *
from fasterai.layers import *
from fastai.torch_core import *
from fastai.callbacks.hooks import *

#The code below is meant to be merged into fastaiv1 ideally

__all__ = ['DynamicUnet2', 'UnetBlock2']

def _get_sfs_idxs(sizes:Sizes) -> List[int]:
    "Get the indexes of the layers where the size of the activation changes."
    feature_szs = [size[-1] for size in sizes]
    sfs_idxs = list(np.where(np.array(feature_szs[:-1]) != np.array(feature_szs[1:]))[0])
    if feature_szs[0] != feature_szs[1]: sfs_idxs = [0] + sfs_idxs
    return sfs_idxs

class PixelShuffle_ICNR2(nn.Module):
    "Upsample by `scale` from `ni` filters to `nf` (default `ni`), using `nn.PixelShuffle`, `icnr` init, and `weight_norm`."
    def __init__(self, ni:int, nf:int=None, scale:int=2, blur:bool=False, leaky:float=None, **kwargs):
        super().__init__()
        nf = ifnone(nf, ni)
        self.conv = conv_layer2(ni, nf*(scale**2), ks=1, use_activ=False, **kwargs)
        icnr(self.conv[0].weight)
        self.shuf = nn.PixelShuffle(scale)
        # Blurring over (h*w) kernel
        # "Super-Resolution using Convolutional Neural Networks without Any Checkerboard Artifacts"
        # - https://arxiv.org/abs/1806.02658
        self.pad = nn.ReplicationPad2d((1,0,1,0))
        self.blur = nn.AvgPool2d(2, stride=1)
        self.relu = relu(True, leaky=leaky)

    def forward(self,x):
        x = self.shuf(self.relu(self.conv(x)))
        return self.blur(self.pad(x)) if self.blur else x

class UnetBlock2(nn.Module):
    "A quasi-UNet block, using `PixelShuffle_ICNR upsampling`."
    def __init__(self, up_in_c:int, x_in_c:int, hook:Hook, final_div:bool=True, blur:bool=False, leaky:float=None,
                 self_attention:bool=False, nf_factor:float=1.0,  **kwargs):
        super().__init__()
        self.hook = hook
        self.shuf = PixelShuffle_ICNR2(up_in_c, up_in_c//2, blur=blur, leaky=leaky, **kwargs)
        self.bn = batchnorm_2d(x_in_c)
        ni = up_in_c//2 + x_in_c
        nf = int((ni if final_div else ni//2)*nf_factor)
        self.conv1 = conv_layer2(ni, nf, leaky=leaky, **kwargs)
        self.conv2 = conv_layer2(nf, nf, leaky=leaky, self_attention=self_attention, **kwargs)
        self.relu = relu(leaky=leaky)

    def forward(self, up_in:Tensor) -> Tensor:
        s = self.hook.stored
        up_out = self.shuf(up_in)
        ssh = s.shape[-2:]
        if ssh != up_out.shape[-2:]:
            up_out = F.interpolate(up_out, s.shape[-2:], mode='nearest')
        cat_x = self.relu(torch.cat([up_out, self.bn(s)], dim=1))
        return self.conv2(self.conv1(cat_x))


class DynamicUnet2(SequentialEx):
    "Create a U-Net from a given architecture."
    def __init__(self, encoder:nn.Module, n_classes:int, blur:bool=False, blur_final=True, self_attention:bool=False,
                 y_range:Optional[Tuple[float,float]]=None, last_cross:bool=True, bottle:bool=False,
                 norm_type:Optional[NormType]=NormType.Batch, nf_factor:float=1.0, **kwargs):
        #extra_bn =  norm_type in (NormType.Spectral, NormType.Weight)
        extra_bn =  norm_type == NormType.Spectral
        imsize = (256,256)
        sfs_szs = model_sizes(encoder, size=imsize)
        sfs_idxs = list(reversed(_get_sfs_idxs(sfs_szs)))
        self.sfs = hook_outputs([encoder[i] for i in sfs_idxs])
        x = dummy_eval(encoder, imsize).detach()

        ni = sfs_szs[-1][1]
        middle_conv = nn.Sequential(conv_layer2(ni, ni*2, norm_type=norm_type, extra_bn=extra_bn, **kwargs),
                                    conv_layer2(ni*2, ni, norm_type=norm_type, extra_bn=extra_bn, **kwargs)).eval()
        x = middle_conv(x)
        layers = [encoder, batchnorm_2d(ni), nn.ReLU(), middle_conv]

        for i,idx in enumerate(sfs_idxs):
            not_final = i!=len(sfs_idxs)-1
            up_in_c, x_in_c = int(x.shape[1]), int(sfs_szs[idx][1])
            do_blur = blur and (not_final or blur_final)
            sa = self_attention and (i==len(sfs_idxs)-3)
            unet_block = UnetBlock2(up_in_c, x_in_c, self.sfs[i], final_div=not_final, blur=blur, self_attention=sa,
                                   norm_type=norm_type, extra_bn=extra_bn, nf_factor=nf_factor, **kwargs).eval()
            layers.append(unet_block)
            x = unet_block(x)

        ni = x.shape[1]
        if imsize != sfs_szs[0][-2:]: layers.append(PixelShuffle_ICNR(ni, **kwargs))
        if last_cross:
            layers.append(MergeLayer(dense=True))
            ni += in_channels(encoder)
            #TODO:  Missing norm_type argument here.  DOH!
            layers.append(res_block(ni, bottle=bottle, **kwargs))
        layers += [conv_layer2(ni, n_classes, ks=1, use_activ=False, norm_type=norm_type)]
        if y_range is not None: layers.append(SigmoidRange(*y_range))
        super().__init__(*layers)

    def __del__(self):
        if hasattr(self, "sfs"): self.sfs.remove()


class DynamicUnet3(SequentialEx):
    "Create a U-Net from a given architecture."
    def __init__(self, encoder:nn.Module, n_classes:int, blur:bool=False, blur_final=True, self_attention:bool=False,
                 y_range:Optional[Tuple[float,float]]=None, last_cross:bool=True, bottle:bool=False,
                 norm_type:Optional[NormType]=NormType.Batch, nf_factor:float=1.0, **kwargs):
        extra_bn =  norm_type == NormType.Spectral
        imsize = (256,256)
        sfs_szs = model_sizes(encoder, size=imsize)
        sfs_idxs = list(reversed(_get_sfs_idxs(sfs_szs)))
        self.sfs = hook_outputs([encoder[i] for i in sfs_idxs])
        x = dummy_eval(encoder, imsize).detach()

        ni = sfs_szs[-1][1]
        middle_conv = nn.Sequential(conv_layer2(ni, ni*2, norm_type=norm_type, extra_bn=extra_bn, **kwargs),
                                    conv_layer2(ni*2, ni, norm_type=norm_type, extra_bn=extra_bn, **kwargs)).eval()
        x = middle_conv(x)
        layers = [encoder, batchnorm_2d(ni), nn.ReLU(), middle_conv]

        for i,idx in enumerate(sfs_idxs):
            not_final = i!=len(sfs_idxs)-1
            up_in_c, x_in_c = int(x.shape[1]), int(sfs_szs[idx][1])
            do_blur = blur and (not_final or blur_final)
            sa = self_attention and (i==len(sfs_idxs)-3)
            unet_block = UnetBlock2(up_in_c, x_in_c, self.sfs[i], final_div=not_final, blur=blur, self_attention=sa,
                                   norm_type=norm_type, extra_bn=extra_bn, nf_factor=nf_factor, **kwargs).eval()
            layers.append(unet_block)
            x = unet_block(x)

        ni = x.shape[1]
        if imsize != sfs_szs[0][-2:]: layers.append(PixelShuffle_ICNR(ni, **kwargs))
        if last_cross:
            layers.append(MergeLayer(dense=True))
            ni += in_channels(encoder)
            layers.append(res_block(ni, bottle=bottle, norm_type=norm_type, **kwargs))
        layers += [conv_layer2(ni, n_classes, ks=1, use_activ=False, norm_type=norm_type)]
        if y_range is not None: layers.append(SigmoidRange(*y_range))
        super().__init__(*layers)

    def __del__(self):
        if hasattr(self, "sfs"): self.sfs.remove()

#No batch norm
class DynamicUnet4(SequentialEx):
    "Create a U-Net from a given architecture."
    def __init__(self, encoder:nn.Module, n_classes:int, blur:bool=False, blur_final=True, self_attention:bool=False,
                 y_range:Optional[Tuple[float,float]]=None, last_cross:bool=True, bottle:bool=False,
                 norm_type:Optional[NormType]=NormType.Batch, nf_factor:float=1.0, **kwargs):
        #extra_bn =  norm_type == NormType.Spectral
        extra_bn = False
        imsize = (256,256)
        sfs_szs = model_sizes(encoder, size=imsize)
        sfs_idxs = list(reversed(_get_sfs_idxs(sfs_szs)))
        self.sfs = hook_outputs([encoder[i] for i in sfs_idxs])
        x = dummy_eval(encoder, imsize).detach()

        ni = sfs_szs[-1][1]
        middle_conv = nn.Sequential(conv_layer2(ni, ni*2, norm_type=norm_type, extra_bn=extra_bn, **kwargs),
                                    conv_layer2(ni*2, ni, norm_type=norm_type, extra_bn=extra_bn, **kwargs)).eval()
        x = middle_conv(x)
        #layers = [encoder, batchnorm_2d(ni), nn.ReLU(), middle_conv]
        layers = [encoder, nn.ReLU(), middle_conv]

        for i,idx in enumerate(sfs_idxs):
            not_final = i!=len(sfs_idxs)-1
            up_in_c, x_in_c = int(x.shape[1]), int(sfs_szs[idx][1])
            do_blur = blur and (not_final or blur_final)
            sa = self_attention and (i==len(sfs_idxs)-3)
            unet_block = UnetBlock2(up_in_c, x_in_c, self.sfs[i], final_div=not_final, blur=blur, self_attention=sa,
                                   norm_type=norm_type, extra_bn=extra_bn, nf_factor=nf_factor, **kwargs).eval()
            layers.append(unet_block)
            x = unet_block(x)

        ni = x.shape[1]
        if imsize != sfs_szs[0][-2:]: layers.append(PixelShuffle_ICNR(ni, **kwargs))
        if last_cross:
            layers.append(MergeLayer(dense=True))
            ni += in_channels(encoder)
            layers.append(res_block(ni, bottle=bottle, norm_type=norm_type, **kwargs))
        layers += [conv_layer2(ni, n_classes, ks=1, use_activ=False, norm_type=norm_type)]
        if y_range is not None: layers.append(SigmoidRange(*y_range))
        super().__init__(*layers)

    def __del__(self):
        if hasattr(self, "sfs"): self.sfs.remove()

class UnetBlock5(nn.Module):
    "A quasi-UNet block, using `PixelShuffle_ICNR upsampling`."
    def __init__(self, up_in_c:int, x_in_c:int, out_c:int, hook:Hook, final_div:bool=True, blur:bool=False, leaky:float=None,
                 self_attention:bool=False,  **kwargs):
        super().__init__()
        self.hook = hook
        self.shuf = PixelShuffle_ICNR2(up_in_c, up_in_c//2, blur=blur, leaky=leaky, **kwargs)
        self.bn = batchnorm_2d(x_in_c)
        ni = up_in_c//2 + x_in_c
        nf = out_c
        self.conv = conv_layer2(ni, nf, leaky=leaky, self_attention=self_attention, **kwargs)
        self.relu = relu(leaky=leaky)

    def forward(self, up_in:Tensor) -> Tensor:
        s = self.hook.stored
        up_out = self.shuf(up_in)
        ssh = s.shape[-2:]
        if ssh != up_out.shape[-2:]:
            up_out = F.interpolate(up_out, s.shape[-2:], mode='nearest')
        cat_x = self.relu(torch.cat([up_out, self.bn(s)], dim=1))
        return self.conv(cat_x)

#custom filter widths
class DynamicUnet5(SequentialEx):
    "Create a U-Net from a given architecture."
    def __init__(self, encoder:nn.Module, n_classes:int, blur:bool=False, blur_final=True, self_attention:bool=False,
                 y_range:Optional[Tuple[float,float]]=None, last_cross:bool=True, bottle:bool=True,
                 norm_type:Optional[NormType]=NormType.Batch, nf:int=256, **kwargs):
        extra_bn =  norm_type == NormType.Spectral
        imsize = (256,256)
        sfs_szs = model_sizes(encoder, size=imsize)
        sfs_idxs = list(reversed(_get_sfs_idxs(sfs_szs)))
        self.sfs = hook_outputs([encoder[i] for i in sfs_idxs])
        x = dummy_eval(encoder, imsize).detach()

        ni = sfs_szs[-1][1]
        middle_conv = nn.Sequential(conv_layer2(ni, ni*2, norm_type=norm_type, extra_bn=extra_bn, **kwargs),
                                    conv_layer2(ni*2, ni, norm_type=norm_type, extra_bn=extra_bn, **kwargs)).eval()
        x = middle_conv(x)
        layers = [encoder, batchnorm_2d(ni), nn.ReLU(), middle_conv]

        for i,idx in enumerate(sfs_idxs):
            not_final = i!=len(sfs_idxs)-1
            up_in_c = int(x.shape[1]) if i == 0 else nf
            x_in_c = int(sfs_szs[idx][1])
            do_blur = blur and (not_final or blur_final)
            sa = self_attention and (i==len(sfs_idxs)-3)
            unet_block = UnetBlock5(up_in_c, x_in_c, nf, self.sfs[i], final_div=not_final, blur=blur, self_attention=sa,
                                   norm_type=norm_type, extra_bn=extra_bn, **kwargs).eval()
            layers.append(unet_block)
            x = unet_block(x)

        ni = x.shape[1]
        if imsize != sfs_szs[0][-2:]: layers.append(PixelShuffle_ICNR(ni, **kwargs))
        if last_cross:
            layers.append(MergeLayer(dense=True))
            ni += in_channels(encoder)
            layers.append(res_block(ni, bottle=bottle, norm_type=norm_type, **kwargs))
        layers += [conv_layer2(ni, n_classes, ks=1, use_activ=False, norm_type=norm_type)]
        if y_range is not None: layers.append(SigmoidRange(*y_range))
        super().__init__(*layers)

    def __del__(self):
        if hasattr(self, "sfs"): self.sfs.remove()