fix(offload): skip accelerator pre-moves of on-demand components

Tab runners pre-move the vae module to the gpu before generation,
which parks an on-demand component on the accelerator for the whole
denoise. Moves of a stamped module toward the accelerator now return
early; the entry bridge onloads it when its encode or decode runs.

modules/sd_models.py

@@ -231,6 +231,8 @@ def move_model(model, device=None, force=False):
 
     if model is None or device is None:
         return
+    if getattr(model, 'sdnext_ondemand', False) and device == devices.device: # on-demand components onload at their entry points instead of pre-moves
+        return
 
     if hasattr(model, 'pipe'):
         move_model(model.pipe, device, force)