diff --git a/pyproject.toml b/pyproject.toml
index 318e4f1..ea8a901 100644
--- a/pyproject.toml
+++ b/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 = "2.5.1"
+version = "2.5.2"
 authors = [
     {name = "numz"},
     {name = "adrientoupet"}
diff --git a/src/common/distributed/basic.py b/src/common/distributed/basic.py
index 195a714..8b0360f 100644
--- a/src/common/distributed/basic.py
+++ b/src/common/distributed/basic.py
@@ -47,7 +47,7 @@ def get_device() -> torch.device:
     """
     Get current rank device.
     """
-    if torch.mps.is_available():
+    if hasattr(torch, 'mps') and callable(getattr(torch.mps, 'is_available', None)) and torch.mps.is_available():
         return torch.device("mps")
     return torch.device("cuda", get_local_rank())
 
diff --git a/src/data/image/transforms/area_resize.py b/src/data/image/transforms/area_resize.py
index a360c1f..2b6dc22 100644
--- a/src/data/image/transforms/area_resize.py
+++ b/src/data/image/transforms/area_resize.py
@@ -31,7 +31,7 @@ class AreaResize:
         self.max_area = max_area
         self.downsample_only = downsample_only
         self.interpolation = interpolation
-        if torch.mps.is_available():
+        if hasattr(torch, 'mps') and callable(getattr(torch.mps, 'is_available', None)) and torch.mps.is_available():
             self.interpolation = InterpolationMode.BILINEAR
 
     def __call__(self, image: Union[torch.Tensor, Image.Image]):
diff --git a/src/data/image/transforms/na_resize.py b/src/data/image/transforms/na_resize.py
index 14ddb5b..ecb49d5 100644
--- a/src/data/image/transforms/na_resize.py
+++ b/src/data/image/transforms/na_resize.py
@@ -26,7 +26,7 @@ def NaResize(
     max_resolution: int = 0,
     interpolation: InterpolationMode = InterpolationMode.BICUBIC,
 ):
-    Interpolation = InterpolationMode.BILINEAR if torch.mps.is_available() else interpolation
+    Interpolation = InterpolationMode.BILINEAR if (hasattr(torch, 'mps') and callable(getattr(torch.mps, 'is_available', None)) and torch.mps.is_available()) else interpolation
     if mode == "area":
         return AreaResize(
             max_area=resolution**2,
diff --git a/src/data/image/transforms/side_resize.py b/src/data/image/transforms/side_resize.py
index 4b6c664..5334e98 100644
--- a/src/data/image/transforms/side_resize.py
+++ b/src/data/image/transforms/side_resize.py
@@ -30,7 +30,7 @@ class SideResize:
         self.max_size = max_size
         self.downsample_only = downsample_only
         self.interpolation = interpolation
-        if torch.mps.is_available():
+        if hasattr(torch, 'mps') and callable(getattr(torch.mps, 'is_available', None)) and torch.mps.is_available():
             self.interpolation = InterpolationMode.BILINEAR
 
     def __call__(self, image: Union[torch.Tensor, Image.Image]):
diff --git a/src/optimization/compatibility.py b/src/optimization/compatibility.py
index f270c17..f640f02 100644
--- a/src/optimization/compatibility.py
+++ b/src/optimization/compatibility.py
@@ -207,7 +207,7 @@ class FP8CompatibleDiT(torch.nn.Module):
             self._convert_rope_freqs(target_dtype=self.compute_dtype)
             self.debug.end_timer("_convert_rope_freqs", "RoPE freqs conversion")
             
-            if torch.mps.is_available():
+            if hasattr(torch, 'mps') and callable(getattr(torch.mps, 'is_available', None)) and torch.mps.is_available():
                 self.debug.log(f"Also converting NaDiT parameters/buffers for MPS backend", category="setup", force=True)
                 self.debug.start_timer("_force_nadit_precision")
                 self._force_nadit_precision(target_dtype=self.compute_dtype)
@@ -510,7 +510,7 @@ class FP8CompatibleDiT(torch.nn.Module):
             k = k.view(batch_size, seq_len, num_heads, head_dim).transpose(1, 2)
             v = v.view(batch_size, seq_len, num_heads, head_dim).transpose(1, 2)
             
-            if torch.mps.is_available():
+            if hasattr(torch, 'mps') and callable(getattr(torch.mps, 'is_available', None)) and torch.mps.is_available():
                 attn_output = torch.nn.functional.scaled_dot_product_attention(
                     q, k, v,
                     dropout_p=0.0,
diff --git a/src/optimization/memory_manager.py b/src/optimization/memory_manager.py
index 71fa501..19715b0 100644
--- a/src/optimization/memory_manager.py
+++ b/src/optimization/memory_manager.py
@@ -80,7 +80,7 @@ def get_basic_vram_info(device: Optional[torch.device] = None) -> Dict[str, Any]
             elif not isinstance(device, torch.device):
                 device = torch.device(device)
             free_memory, total_memory = torch.cuda.mem_get_info(device)
-        elif torch.mps.is_available():
+        elif hasattr(torch, 'mps') and callable(getattr(torch.mps, 'is_available', None)) and torch.mps.is_available():
             # MPS doesn't support per-device queries or mem_get_info
             # Use system memory as proxy
             mem = psutil.virtual_memory()
@@ -100,7 +100,7 @@ def get_basic_vram_info(device: Optional[torch.device] = None) -> Dict[str, Any]
 # Initial VRAM check at module load
 vram_info = get_basic_vram_info(device=None)
 if "error" not in vram_info:
-    backend = "MPS" if torch.mps.is_available() else "CUDA"
+    backend = "MPS" if (hasattr(torch, 'mps') and callable(getattr(torch.mps, 'is_available', None)) and torch.mps.is_available()) else "CUDA"
     print(f"📊 Initial {backend} memory: {vram_info['free_gb']:.2f}GB free / {vram_info['total_gb']:.2f}GB total")
 else:
     print(f"⚠️ Memory check failed: {vram_info['error']} - No available backend!")
@@ -129,7 +129,7 @@ def get_vram_usage(device: Optional[torch.device] = None, debug: Optional['Debug
             reserved = torch.cuda.memory_reserved(device) / (1024**3)
             max_allocated = torch.cuda.max_memory_allocated(device) / (1024**3)
             return allocated, reserved, max_allocated
-        elif torch.mps.is_available():
+        elif hasattr(torch, 'mps') and callable(getattr(torch.mps, 'is_available', None)) and torch.mps.is_available():
             # MPS doesn't support per-device queries - uses global memory tracking
             allocated = torch.mps.current_allocated_memory() / (1024**3)
             reserved = torch.mps.driver_allocated_memory() / (1024**3)
@@ -235,11 +235,11 @@ def clear_memory(debug: Optional['Debug'] = None, deep: bool = False, force: boo
             if free_ratio < 0.05:
                 should_clear = True
                 if debug:
-                    backend = "MPS" if torch.mps.is_available() else "VRAM"
+                    backend = "MPS" if (hasattr(torch, 'mps') and callable(getattr(torch.mps, 'is_available', None)) and torch.mps.is_available()) else "VRAM"
                     debug.log(f"{backend} pressure: {mem_info['free_gb']:.2f}GB free of {mem_info['total_gb']:.2f}GB", category="memory")
         
         # For non-MPS systems, also check system RAM separately
-        if not should_clear and not torch.mps.is_available():
+        if not should_clear and not (hasattr(torch, 'mps') and callable(getattr(torch.mps, 'is_available', None)) and torch.mps.is_available()):
             mem = psutil.virtual_memory()
             if mem.available < mem.total * 0.05:
                 should_clear = True
@@ -265,7 +265,7 @@ def clear_memory(debug: Optional['Debug'] = None, deep: bool = False, force: boo
     if torch.cuda.is_available():
         torch.cuda.empty_cache()
         torch.cuda.ipc_collect()
-    elif torch.mps.is_available():
+    elif hasattr(torch, 'mps') and callable(getattr(torch.mps, 'is_available', None)) and torch.mps.is_available():
         torch.mps.empty_cache()
     
     if debug:
@@ -302,7 +302,7 @@ def clear_memory(debug: Optional['Debug'] = None, deep: bool = False, force: boo
                 handle = _os_memory_lib.GetCurrentProcess()
                 _os_memory_lib.SetProcessWorkingSetSize(handle, -1, -1)
                 
-            elif torch.mps.is_available():
+            elif hasattr(torch, 'mps') and callable(getattr(torch.mps, 'is_available', None)) and torch.mps.is_available():
                 # macOS with MPS
                 import ctypes  # Import only when needed
                 import ctypes.util
diff --git a/src/utils/debug.py b/src/utils/debug.py
index e1040f6..24d0739 100644
--- a/src/utils/debug.py
+++ b/src/utils/debug.py
@@ -333,7 +333,7 @@ class Debug:
         }
         
         # VRAM metrics
-        if torch.cuda.is_available() or torch.mps.is_available():
+        if torch.cuda.is_available() or (hasattr(torch, 'mps') and callable(getattr(torch.mps, 'is_available', None)) and torch.mps.is_available()):
             metrics['vram_allocated'], metrics['vram_reserved'], current_global_peak = get_vram_usage(device=None, debug=self)
 
             # Calculate peak since last log_memory_state
@@ -346,7 +346,7 @@ class Debug:
                 metrics['vram_free'] = vram_info["free_gb"]
                 metrics['vram_total'] = vram_info["total_gb"]
                 
-                backend = "MPS" if torch.mps.is_available() else "VRAM"
+                backend = "MPS" if (hasattr(torch, 'mps') and callable(getattr(torch.mps, 'is_available', None)) and torch.mps.is_available()) else "VRAM"
                 metrics['summary_vram'] = (f"  [{backend}] {metrics['vram_allocated']:.2f}GB allocated / "
                         f"{metrics['vram_reserved']:.2f}GB reserved / "
                         f"Peak: {metrics['vram_peak_since_last']:.2f}GB / "
@@ -369,7 +369,7 @@ class Debug:
             metrics['summary_ram'] = ""
         
         # Update VRAM history for tracking
-        if torch.cuda.is_available() or torch.mps.is_available():
+        if torch.cuda.is_available() or (hasattr(torch, 'mps') and callable(getattr(torch.mps, 'is_available', None)) and torch.mps.is_available()):
             self.vram_history.append(metrics['vram_allocated'])
         
         return metrics