import logging from dataclasses import dataclass from typing import Literal, Optional, Tuple, Union import torch import torch.nn as nn import torch.nn.functional as F from transformers import PreTrainedModel from transformers.utils import ModelOutput from .configuration_isnet import ISNetConfig logger = logging.getLogger(__name__) @dataclass class ISNetStageOutput(ModelOutput): d1: torch.Tensor d2: Optional[torch.Tensor] = None d3: Optional[torch.Tensor] = None d4: Optional[torch.Tensor] = None d5: Optional[torch.Tensor] = None d6: Optional[torch.Tensor] = None @dataclass class ISNetModelOutput(ModelOutput): activated: ISNetStageOutput hidden_states: Optional[ISNetStageOutput] = None bce_loss = nn.BCELoss(size_average=True) def muti_loss_fusion(preds, target): loss0 = 0.0 loss = 0.0 for i in range(0, len(preds)): # print("i: ", i, preds[i].shape) if preds[i].shape[2] != target.shape[2] or preds[i].shape[3] != target.shape[3]: # tmp_target = _upsample_like(target,preds[i]) tmp_target = F.interpolate( target, size=preds[i].size()[2:], mode="bilinear", align_corners=True ) loss = loss + bce_loss(preds[i], tmp_target) else: loss = loss + bce_loss(preds[i], target) if i == 0: loss0 = loss return loss0, loss fea_loss = nn.MSELoss(size_average=True) kl_loss = nn.KLDivLoss(size_average=True) l1_loss = nn.L1Loss(size_average=True) smooth_l1_loss = nn.SmoothL1Loss(size_average=True) LossMode = Literal["MSE", "KL", "MAE", "SmoothL1"] def muti_loss_fusion_kl( preds, target, dfs, fs, mode: LossMode = "MSE" ) -> Tuple[torch.Tensor, torch.Tensor]: loss0 = 0.0 loss = 0.0 for i in range(0, len(preds)): # print("i: ", i, preds[i].shape) if preds[i].shape[2] != target.shape[2] or preds[i].shape[3] != target.shape[3]: # tmp_target = _upsample_like(target,preds[i]) tmp_target = F.interpolate( target, size=preds[i].size()[2:], mode="bilinear", align_corners=True ) loss = loss + bce_loss(preds[i], tmp_target) else: loss = loss + bce_loss(preds[i], target) if i == 0: loss0 = loss for i in range(0, len(dfs)): if mode == "MSE": loss = loss + fea_loss( dfs[i], fs[i] ) ### add the mse loss of features as additional constraints # print("fea_loss: ", fea_loss(dfs[i],fs[i]).item()) elif mode == "KL": loss = loss + kl_loss(F.log_softmax(dfs[i], dim=1), F.softmax(fs[i], dim=1)) # print("kl_loss: ", kl_loss(F.log_softmax(dfs[i],dim=1),F.softmax(fs[i],dim=1)).item()) elif mode == "MAE": loss = loss + l1_loss(dfs[i], fs[i]) # print("ls_loss: ", l1_loss(dfs[i],fs[i])) elif mode == "SmoothL1": loss = loss + smooth_l1_loss(dfs[i], fs[i]) # print("SmoothL1: ", smooth_l1_loss(dfs[i],fs[i]).item()) return loss0, loss class REBNCONV(nn.Module): def __init__(self, in_ch=3, out_ch=3, dirate=1, stride=1): super(REBNCONV, self).__init__() self.conv_s1 = nn.Conv2d( in_ch, out_ch, 3, padding=1 * dirate, dilation=1 * dirate, stride=stride ) self.bn_s1 = nn.BatchNorm2d(out_ch) self.relu_s1 = nn.ReLU(inplace=True) def forward(self, x): hx = x xout = self.relu_s1(self.bn_s1(self.conv_s1(hx))) return xout def _upsample_like(src, tar: torch.Tensor) -> torch.Tensor: """upsample tensor 'src' to have the same spatial size with tensor 'tar'""" return F.upsample(src, size=tar.shape[2:], mode="bilinear") ### RSU-7 ### class RSU7(nn.Module): def __init__(self, in_ch=3, mid_ch=12, out_ch=3, img_size=512) -> None: super().__init__() self.in_ch = in_ch self.mid_ch = mid_ch self.out_ch = out_ch self.rebnconvin = REBNCONV(in_ch, out_ch, dirate=1) ## 1 -> 1/2 self.rebnconv1 = REBNCONV(out_ch, mid_ch, dirate=1) self.pool1 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.rebnconv2 = REBNCONV(mid_ch, mid_ch, dirate=1) self.pool2 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.rebnconv3 = REBNCONV(mid_ch, mid_ch, dirate=1) self.pool3 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.rebnconv4 = REBNCONV(mid_ch, mid_ch, dirate=1) self.pool4 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.rebnconv5 = REBNCONV(mid_ch, mid_ch, dirate=1) self.pool5 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.rebnconv6 = REBNCONV(mid_ch, mid_ch, dirate=1) self.rebnconv7 = REBNCONV(mid_ch, mid_ch, dirate=2) self.rebnconv6d = REBNCONV(mid_ch * 2, mid_ch, dirate=1) self.rebnconv5d = REBNCONV(mid_ch * 2, mid_ch, dirate=1) self.rebnconv4d = REBNCONV(mid_ch * 2, mid_ch, dirate=1) self.rebnconv3d = REBNCONV(mid_ch * 2, mid_ch, dirate=1) self.rebnconv2d = REBNCONV(mid_ch * 2, mid_ch, dirate=1) self.rebnconv1d = REBNCONV(mid_ch * 2, out_ch, dirate=1) def forward(self, x: torch.Tensor) -> torch.Tensor: # b, c, h, w = x.shape hx = x hxin = self.rebnconvin(hx) hx1 = self.rebnconv1(hxin) hx = self.pool1(hx1) hx2 = self.rebnconv2(hx) hx = self.pool2(hx2) hx3 = self.rebnconv3(hx) hx = self.pool3(hx3) hx4 = self.rebnconv4(hx) hx = self.pool4(hx4) hx5 = self.rebnconv5(hx) hx = self.pool5(hx5) hx6 = self.rebnconv6(hx) hx7 = self.rebnconv7(hx6) hx6d = self.rebnconv6d(torch.cat((hx7, hx6), 1)) hx6dup = _upsample_like(hx6d, hx5) hx5d = self.rebnconv5d(torch.cat((hx6dup, hx5), 1)) hx5dup = _upsample_like(hx5d, hx4) hx4d = self.rebnconv4d(torch.cat((hx5dup, hx4), 1)) hx4dup = _upsample_like(hx4d, hx3) hx3d = self.rebnconv3d(torch.cat((hx4dup, hx3), 1)) hx3dup = _upsample_like(hx3d, hx2) hx2d = self.rebnconv2d(torch.cat((hx3dup, hx2), 1)) hx2dup = _upsample_like(hx2d, hx1) hx1d = self.rebnconv1d(torch.cat((hx2dup, hx1), 1)) return hx1d + hxin ### RSU-6 ### class RSU6(nn.Module): def __init__(self, in_ch=3, mid_ch=12, out_ch=3) -> None: super().__init__() self.rebnconvin = REBNCONV(in_ch, out_ch, dirate=1) self.rebnconv1 = REBNCONV(out_ch, mid_ch, dirate=1) self.pool1 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.rebnconv2 = REBNCONV(mid_ch, mid_ch, dirate=1) self.pool2 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.rebnconv3 = REBNCONV(mid_ch, mid_ch, dirate=1) self.pool3 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.rebnconv4 = REBNCONV(mid_ch, mid_ch, dirate=1) self.pool4 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.rebnconv5 = REBNCONV(mid_ch, mid_ch, dirate=1) self.rebnconv6 = REBNCONV(mid_ch, mid_ch, dirate=2) self.rebnconv5d = REBNCONV(mid_ch * 2, mid_ch, dirate=1) self.rebnconv4d = REBNCONV(mid_ch * 2, mid_ch, dirate=1) self.rebnconv3d = REBNCONV(mid_ch * 2, mid_ch, dirate=1) self.rebnconv2d = REBNCONV(mid_ch * 2, mid_ch, dirate=1) self.rebnconv1d = REBNCONV(mid_ch * 2, out_ch, dirate=1) def forward(self, x: torch.Tensor) -> torch.Tensor: hx = x hxin = self.rebnconvin(hx) hx1 = self.rebnconv1(hxin) hx = self.pool1(hx1) hx2 = self.rebnconv2(hx) hx = self.pool2(hx2) hx3 = self.rebnconv3(hx) hx = self.pool3(hx3) hx4 = self.rebnconv4(hx) hx = self.pool4(hx4) hx5 = self.rebnconv5(hx) hx6 = self.rebnconv6(hx5) hx5d = self.rebnconv5d(torch.cat((hx6, hx5), 1)) hx5dup = _upsample_like(hx5d, hx4) hx4d = self.rebnconv4d(torch.cat((hx5dup, hx4), 1)) hx4dup = _upsample_like(hx4d, hx3) hx3d = self.rebnconv3d(torch.cat((hx4dup, hx3), 1)) hx3dup = _upsample_like(hx3d, hx2) hx2d = self.rebnconv2d(torch.cat((hx3dup, hx2), 1)) hx2dup = _upsample_like(hx2d, hx1) hx1d = self.rebnconv1d(torch.cat((hx2dup, hx1), 1)) return hx1d + hxin ### RSU-5 ### class RSU5(nn.Module): def __init__(self, in_ch=3, mid_ch=12, out_ch=3) -> None: super().__init__() self.rebnconvin = REBNCONV(in_ch, out_ch, dirate=1) self.rebnconv1 = REBNCONV(out_ch, mid_ch, dirate=1) self.pool1 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.rebnconv2 = REBNCONV(mid_ch, mid_ch, dirate=1) self.pool2 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.rebnconv3 = REBNCONV(mid_ch, mid_ch, dirate=1) self.pool3 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.rebnconv4 = REBNCONV(mid_ch, mid_ch, dirate=1) self.rebnconv5 = REBNCONV(mid_ch, mid_ch, dirate=2) self.rebnconv4d = REBNCONV(mid_ch * 2, mid_ch, dirate=1) self.rebnconv3d = REBNCONV(mid_ch * 2, mid_ch, dirate=1) self.rebnconv2d = REBNCONV(mid_ch * 2, mid_ch, dirate=1) self.rebnconv1d = REBNCONV(mid_ch * 2, out_ch, dirate=1) def forward(self, x: torch.Tensor) -> torch.Tensor: hx = x hxin = self.rebnconvin(hx) hx1 = self.rebnconv1(hxin) hx = self.pool1(hx1) hx2 = self.rebnconv2(hx) hx = self.pool2(hx2) hx3 = self.rebnconv3(hx) hx = self.pool3(hx3) hx4 = self.rebnconv4(hx) hx5 = self.rebnconv5(hx4) hx4d = self.rebnconv4d(torch.cat((hx5, hx4), 1)) hx4dup = _upsample_like(hx4d, hx3) hx3d = self.rebnconv3d(torch.cat((hx4dup, hx3), 1)) hx3dup = _upsample_like(hx3d, hx2) hx2d = self.rebnconv2d(torch.cat((hx3dup, hx2), 1)) hx2dup = _upsample_like(hx2d, hx1) hx1d = self.rebnconv1d(torch.cat((hx2dup, hx1), 1)) return hx1d + hxin ### RSU-4 ### class RSU4(nn.Module): def __init__(self, in_ch=3, mid_ch=12, out_ch=3) -> None: super(RSU4, self).__init__() self.rebnconvin = REBNCONV(in_ch, out_ch, dirate=1) self.rebnconv1 = REBNCONV(out_ch, mid_ch, dirate=1) self.pool1 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.rebnconv2 = REBNCONV(mid_ch, mid_ch, dirate=1) self.pool2 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.rebnconv3 = REBNCONV(mid_ch, mid_ch, dirate=1) self.rebnconv4 = REBNCONV(mid_ch, mid_ch, dirate=2) self.rebnconv3d = REBNCONV(mid_ch * 2, mid_ch, dirate=1) self.rebnconv2d = REBNCONV(mid_ch * 2, mid_ch, dirate=1) self.rebnconv1d = REBNCONV(mid_ch * 2, out_ch, dirate=1) def forward(self, x: torch.Tensor) -> torch.Tensor: hx = x hxin = self.rebnconvin(hx) hx1 = self.rebnconv1(hxin) hx = self.pool1(hx1) hx2 = self.rebnconv2(hx) hx = self.pool2(hx2) hx3 = self.rebnconv3(hx) hx4 = self.rebnconv4(hx3) hx3d = self.rebnconv3d(torch.cat((hx4, hx3), 1)) hx3dup = _upsample_like(hx3d, hx2) hx2d = self.rebnconv2d(torch.cat((hx3dup, hx2), 1)) hx2dup = _upsample_like(hx2d, hx1) hx1d = self.rebnconv1d(torch.cat((hx2dup, hx1), 1)) return hx1d + hxin ### RSU-4F ### class RSU4F(nn.Module): def __init__(self, in_ch=3, mid_ch=12, out_ch=3) -> None: super().__init__() self.rebnconvin = REBNCONV(in_ch, out_ch, dirate=1) self.rebnconv1 = REBNCONV(out_ch, mid_ch, dirate=1) self.rebnconv2 = REBNCONV(mid_ch, mid_ch, dirate=2) self.rebnconv3 = REBNCONV(mid_ch, mid_ch, dirate=4) self.rebnconv4 = REBNCONV(mid_ch, mid_ch, dirate=8) self.rebnconv3d = REBNCONV(mid_ch * 2, mid_ch, dirate=4) self.rebnconv2d = REBNCONV(mid_ch * 2, mid_ch, dirate=2) self.rebnconv1d = REBNCONV(mid_ch * 2, out_ch, dirate=1) def forward(self, x: torch.Tensor) -> torch.Tensor: hx = x hxin = self.rebnconvin(hx) hx1 = self.rebnconv1(hxin) hx2 = self.rebnconv2(hx1) hx3 = self.rebnconv3(hx2) hx4 = self.rebnconv4(hx3) hx3d = self.rebnconv3d(torch.cat((hx4, hx3), 1)) hx2d = self.rebnconv2d(torch.cat((hx3d, hx2), 1)) hx1d = self.rebnconv1d(torch.cat((hx2d, hx1), 1)) return hx1d + hxin class myrebnconv(nn.Module): def __init__( self, in_ch=3, out_ch=1, kernel_size=3, stride=1, padding=1, dilation=1, groups=1, ) -> None: super().__init__() self.conv = nn.Conv2d( in_ch, out_ch, kernel_size=kernel_size, stride=stride, padding=padding, dilation=dilation, groups=groups, ) self.bn = nn.BatchNorm2d(out_ch) self.rl = nn.ReLU(inplace=True) def forward(self, x: torch.Tensor) -> torch.Tensor: return self.rl(self.bn(self.conv(x))) class ISNetGTEncoder(nn.Module): def __init__(self, in_ch=1, out_ch=1) -> None: super(ISNetGTEncoder, self).__init__() self.conv_in = myrebnconv( in_ch, 16, 3, stride=2, padding=1 ) # nn.Conv2d(in_ch,64,3,stride=2,padding=1) self.stage1 = RSU7(16, 16, 64) self.pool12 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.stage2 = RSU6(64, 16, 64) self.pool23 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.stage3 = RSU5(64, 32, 128) self.pool34 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.stage4 = RSU4(128, 32, 256) self.pool45 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.stage5 = RSU4F(256, 64, 512) self.pool56 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.stage6 = RSU4F(512, 64, 512) self.side1 = nn.Conv2d(64, out_ch, 3, padding=1) self.side2 = nn.Conv2d(64, out_ch, 3, padding=1) self.side3 = nn.Conv2d(128, out_ch, 3, padding=1) self.side4 = nn.Conv2d(256, out_ch, 3, padding=1) self.side5 = nn.Conv2d(512, out_ch, 3, padding=1) self.side6 = nn.Conv2d(512, out_ch, 3, padding=1) def compute_loss(self, preds, targets): return muti_loss_fusion(preds, targets) def forward( self, x: torch.Tensor ) -> Tuple[ Tuple[ torch.Tensor, torch.Tensor, torch.Tensor, torch.Tensor, torch.Tensor, torch.Tensor, ], Tuple[ torch.Tensor, torch.Tensor, torch.Tensor, torch.Tensor, torch.Tensor, torch.Tensor, ], ]: hx = x hxin = self.conv_in(hx) # hx = self.pool_in(hxin) # stage 1 hx1 = self.stage1(hxin) hx = self.pool12(hx1) # stage 2 hx2 = self.stage2(hx) hx = self.pool23(hx2) # stage 3 hx3 = self.stage3(hx) hx = self.pool34(hx3) # stage 4 hx4 = self.stage4(hx) hx = self.pool45(hx4) # stage 5 hx5 = self.stage5(hx) hx = self.pool56(hx5) # stage 6 hx6 = self.stage6(hx) # side output d1 = self.side1(hx1) d1 = _upsample_like(d1, x) d2 = self.side2(hx2) d2 = _upsample_like(d2, x) d3 = self.side3(hx3) d3 = _upsample_like(d3, x) d4 = self.side4(hx4) d4 = _upsample_like(d4, x) d5 = self.side5(hx5) d5 = _upsample_like(d5, x) d6 = self.side6(hx6) d6 = _upsample_like(d6, x) # d0 = self.outconv(torch.cat((d1,d2,d3,d4,d5,d6),1)) activated = ( F.sigmoid(d1), F.sigmoid(d2), F.sigmoid(d3), F.sigmoid(d4), F.sigmoid(d5), F.sigmoid(d6), ) hidden_states = ( hx1, hx2, hx3, hx4, hx5, hx6, ) return activated, hidden_states class ISNetModel(PreTrainedModel): config_class = ISNetConfig def __init__(self, config: ISNetConfig) -> None: super().__init__(config) self.conv_in = nn.Conv2d(config.in_channels, 64, 3, stride=2, padding=1) self.pool_in = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.stage1 = RSU7(64, 32, 64) self.pool12 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.stage2 = RSU6(64, 32, 128) self.pool23 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.stage3 = RSU5(128, 64, 256) self.pool34 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.stage4 = RSU4(256, 128, 512) self.pool45 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.stage5 = RSU4F(512, 256, 512) self.pool56 = nn.MaxPool2d(2, stride=2, ceil_mode=True) self.stage6 = RSU4F(512, 256, 512) # decoder self.stage5d = RSU4F(1024, 256, 512) self.stage4d = RSU4(1024, 128, 256) self.stage3d = RSU5(512, 64, 128) self.stage2d = RSU6(256, 32, 64) self.stage1d = RSU7(128, 16, 64) self.side1 = nn.Conv2d(64, config.out_channels, 3, padding=1) self.side2 = nn.Conv2d(64, config.out_channels, 3, padding=1) self.side3 = nn.Conv2d(128, config.out_channels, 3, padding=1) self.side4 = nn.Conv2d(256, config.out_channels, 3, padding=1) self.side5 = nn.Conv2d(512, config.out_channels, 3, padding=1) self.side6 = nn.Conv2d(512, config.out_channels, 3, padding=1) # self.outconv = nn.Conv2d(6*out_ch,out_ch,1) def compute_loss_kl(self, preds, targets, dfs, fs, mode: LossMode = "MSE"): # return muti_loss_fusion(preds,targets) return muti_loss_fusion_kl(preds, targets, dfs, fs, mode=mode) def compute_loss(self, preds, targets): # return muti_loss_fusion(preds,targets) return muti_loss_fusion(preds, targets) def forward( self, pixel_values: torch.Tensor, return_dict: Optional[bool] = None ) -> Union[Tuple, ISNetModelOutput]: x = pixel_values hx = x hxin = self.conv_in(hx) # hx = self.pool_in(hxin) # stage 1 hx1 = self.stage1(hxin) hx = self.pool12(hx1) # stage 2 hx2 = self.stage2(hx) hx = self.pool23(hx2) # stage 3 hx3 = self.stage3(hx) hx = self.pool34(hx3) # stage 4 hx4 = self.stage4(hx) hx = self.pool45(hx4) # stage 5 hx5 = self.stage5(hx) hx = self.pool56(hx5) # stage 6 hx6 = self.stage6(hx) hx6up = _upsample_like(hx6, hx5) # -------------------- decoder -------------------- hx5d = self.stage5d(torch.cat((hx6up, hx5), 1)) hx5dup = _upsample_like(hx5d, hx4) hx4d = self.stage4d(torch.cat((hx5dup, hx4), 1)) hx4dup = _upsample_like(hx4d, hx3) hx3d = self.stage3d(torch.cat((hx4dup, hx3), 1)) hx3dup = _upsample_like(hx3d, hx2) hx2d = self.stage2d(torch.cat((hx3dup, hx2), 1)) hx2dup = _upsample_like(hx2d, hx1) hx1d = self.stage1d(torch.cat((hx2dup, hx1), 1)) # side output d1 = self.side1(hx1d) d1 = _upsample_like(d1, x) d2 = self.side2(hx2d) d2 = _upsample_like(d2, x) d3 = self.side3(hx3d) d3 = _upsample_like(d3, x) d4 = self.side4(hx4d) d4 = _upsample_like(d4, x) d5 = self.side5(hx5d) d5 = _upsample_like(d5, x) d6 = self.side6(hx6) d6 = _upsample_like(d6, x) # d0 = self.outconv(torch.cat((d1,d2,d3,d4,d5,d6),1)) activated = ( F.sigmoid(d1), F.sigmoid(d2), F.sigmoid(d3), F.sigmoid(d4), F.sigmoid(d5), F.sigmoid(d6), ) hidden_states = ( hx1d, hx2d, hx3d, hx4d, hx5d, hx6, ) if not return_dict: return activated, hidden_states return ISNetModelOutput( activated=ISNetStageOutput(*activated), hidden_states=ISNetStageOutput(*hidden_states), ) def convert_from_checkpoint( repo_id: str, filename: str, config: Optional[ISNetConfig] = None ) -> ISNetModel: from huggingface_hub import hf_hub_download checkpoint_path = hf_hub_download(repo_id=repo_id, filename=filename) config = config or ISNetConfig() model = ISNetModel(config) logger.info(f"Loading checkpoint from {checkpoint_path}") state_dict = torch.load(checkpoint_path) model.load_state_dict(state_dict, strict=True) model.eval() return model