fix: 🐛 Fix the torch.median() on apple gpu (#94)

✅ Closes: #1

lib_layerdiffusion/models.py

@@ -299,7 +299,16 @@ class TransparentVAEDecoder:
             result += [eps]
 
         result = torch.stack(result, dim=0)
-        median = torch.median(result, dim=0).values
+        if self.load_device == torch.device("mps"):
+            '''
+            In case that apple silicon devices would crash when calling torch.median() on tensors
+            in gpu vram with dimensions higher than 4, we move it to cpu, call torch.median()
+            and then move the result back to gpu.
+            '''
+            median = torch.median(result.cpu(), dim=0).values
+            median = median.to(device=self.load_device, dtype=self.dtype)
+        else:
+            median = torch.median(result, dim=0).values
         return median
 
     @torch.no_grad()