Support DiffBIR SwinIR models.

comfy_extras/chainner_models/architecture/SwinIR.py

@@ -846,6 +846,7 @@ def __init__(
         num_in_ch = in_chans
         num_out_ch = in_chans
         supports_fp16 = True
+        self.start_unshuffle = 1
 
         self.model_arch = "SwinIR"
         self.sub_type = "SR"
@@ -874,6 +875,11 @@ def __init__(
             else 64
         )
 
+        if "conv_first.1.weight" in self.state:
+            self.state["conv_first.weight"] = self.state.pop("conv_first.1.weight")
+            self.state["conv_first.bias"] = self.state.pop("conv_first.1.bias")
+            self.start_unshuffle = round(math.sqrt(self.state["conv_first.weight"].shape[1] // 3))
+
         num_in_ch = self.state["conv_first.weight"].shape[1]
         in_chans = num_in_ch
         if "conv_last.weight" in state_keys:
@@ -968,7 +974,7 @@ def __init__(
         self.depths = depths
         self.window_size = window_size
         self.mlp_ratio = mlp_ratio
-        self.scale = upscale
+        self.scale = upscale / self.start_unshuffle
         self.upsampler = upsampler
         self.img_size = img_size
         self.img_range = img_range
@@ -1101,6 +1107,9 @@ def __init__(
             self.conv_up1 = nn.Conv2d(num_feat, num_feat, 3, 1, 1)
             if self.upscale == 4:
                 self.conv_up2 = nn.Conv2d(num_feat, num_feat, 3, 1, 1)
+            elif self.upscale == 8:
+                self.conv_up2 = nn.Conv2d(num_feat, num_feat, 3, 1, 1)
+                self.conv_up3 = nn.Conv2d(num_feat, num_feat, 3, 1, 1)
             self.conv_hr = nn.Conv2d(num_feat, num_feat, 3, 1, 1)
             self.conv_last = nn.Conv2d(num_feat, num_out_ch, 3, 1, 1)
             self.lrelu = nn.LeakyReLU(negative_slope=0.2, inplace=True)
@@ -1157,6 +1166,9 @@ def forward(self, x):
         self.mean = self.mean.type_as(x)
         x = (x - self.mean) * self.img_range
 
+        if self.start_unshuffle > 1:
+            x = torch.nn.functional.pixel_unshuffle(x, self.start_unshuffle)
+
         if self.upsampler == "pixelshuffle":
             # for classical SR
             x = self.conv_first(x)
@@ -1186,6 +1198,9 @@ def forward(self, x):
                         )
                     )
                 )
+            elif self.upscale == 8:
+                x = self.lrelu(self.conv_up2(torch.nn.functional.interpolate(x, scale_factor=2, mode='nearest')))
+                x = self.lrelu(self.conv_up3(torch.nn.functional.interpolate(x, scale_factor=2, mode='nearest')))
             x = self.conv_last(self.lrelu(self.conv_hr(x)))
         else:
             # for image denoising and JPEG compression artifact reduction

comfy_extras/nodes_upscale_model.py

@@ -18,6 +18,8 @@ def INPUT_TYPES(s):
     def load_model(self, model_name):
         model_path = folder_paths.get_full_path("upscale_models", model_name)
         sd = comfy.utils.load_torch_file(model_path, safe_load=True)
+        if "module.layers.0.residual_group.blocks.0.norm1.weight" in sd:
+            sd = comfy.utils.state_dict_prefix_replace(sd, {"module.":""})
         out = model_loading.load_state_dict(sd).eval()
         return (out, )