Fix timestep_embedding for torch.compile compatibility

Revert to CPU-first tensor creation then .to(device) instead of
device= parameter. The device= parameter causes torch.compile to fail
with ConstantVariable assertion error for torch.device type.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>

comfy/ldm/modules/diffusionmodules/util.py

@@ -268,8 +268,10 @@ def timestep_embedding(timesteps, dim, max_period=10000, repeat_only=False):
     if not repeat_only:
         half = dim // 2
         freqs = torch.exp(
-            -math.log(max_period) * torch.arange(start=0, end=half, dtype=torch.float32, device=timesteps.device) / half
-        )
+            -math.log(max_period)
+            * torch.arange(start=0, end=half, dtype=torch.float32)
+            / half
+        ).to(timesteps.device)
         args = timesteps[:, None].float() * freqs[None]
         embedding = torch.cat([torch.cos(args), torch.sin(args)], dim=-1)
         if dim % 2: