fix: AdaIN color correction and AMD ROCm compatibility (v2.5.4)

- Fix AdaIN non-contiguous tensor error by using reshape() instead of view()
- Add cuDNN availability checks to prevent ROCm 'ATen not compiled with cuDNN' error

README.md

@@ -36,6 +36,11 @@ We're actively working on improvements and new features. To stay informed:
 
 ## 🚀 Updates
 
+**2025.11.08 - Version 2.5.4**
+
+- 🎨 **Fix: AdaIN color correction** - Replace `.view()` with `.reshape()` to handle non-contiguous tensors after spatial padding, resolving "view size is not compatible with input tensor's size and stride" error
+- 🔴 **Fix: AMD ROCm compatibility** - Add cuDNN availability check in Conv3d workaround to prevent "ATen not compiled with cuDNN support" error on ROCm systems (AMD GPUs on Windows/Linux)
+
 **2025.11.08 - Version 2.5.3**
 
 - 🍎 **Fix: Apple Silicon MPS device handling** - Corrected MPS device enumeration to use `"mps"` instead of `"mps:0"`, resolving invalid device errors on M-series Macs

pyproject.toml

@@ -1,7 +1,7 @@
 [project]
 name = "seedvr2_videoupscaler"
 description = "SeedVR2 official ComfyUI integration: ByteDance-Seed's one-step diffusion-based video/image upscaling with memory-efficient inference"
-version = "2.5.3"
+version = "2.5.4"
 authors = [
     {name = "numz"},
     {name = "adrientoupet"}

src/models/video_vae_v3/modules/causal_inflation_lib.py

@@ -91,8 +91,12 @@ class InflatedCausalConv3d(Conv3d):
         Workaround: Call torch.cudnn_convolution directly to bypass buggy layer.
         Status is logged at startup in compatibility.py.
         """
-        if NVIDIA_CONV3D_MEMORY_BUG_WORKAROUND and weight.dtype in (torch.float16, torch.bfloat16):
-            # Direct cuDNN call bypasses buggy PyTorch dispatch layer
+        if (NVIDIA_CONV3D_MEMORY_BUG_WORKAROUND and 
+            weight.dtype in (torch.float16, torch.bfloat16) and 
+            hasattr(torch.backends.cudnn, 'is_available') and
+            torch.backends.cudnn.is_available() and
+            getattr(torch.backends.cudnn, 'enabled', True)):
+            # Direct cuDNN call bypasses buggy PyTorch dispatch layer (NVIDIA only)
             out = torch.cudnn_convolution(
                 input, weight, self.padding, self.stride, self.dilation, self.groups,
                 benchmark=False, deterministic=False, allow_tf32=True

src/utils/color_fix.py

@@ -84,9 +84,9 @@ def calc_mean_std(feat: Tensor, eps: float = 1e-5) -> tuple[Tensor, Tensor]:
     assert len(size) == 4, 'The input feature should be 4D tensor.'
     b, c = size[:2]
     
-    feat_var = feat.view(b, c, -1).var(dim=2) + eps
-    feat_std = feat_var.sqrt().view(b, c, 1, 1)
-    feat_mean = feat.view(b, c, -1).mean(dim=2).view(b, c, 1, 1)
+    feat_var = feat.reshape(b, c, -1).var(dim=2) + eps
+    feat_std = feat_var.sqrt().reshape(b, c, 1, 1)
+    feat_mean = feat.reshape(b, c, -1).mean(dim=2).reshape(b, c, 1, 1)
     
     return feat_mean, feat_std