no mask for default case

utils/drag_utils.py

@@ -104,6 +104,7 @@ def drag_diffusion_update(model,
     # prepare for point tracking and background regularization
     handle_points_init = copy.deepcopy(handle_points)
     interp_mask = F.interpolate(mask, (init_code.shape[2],init_code.shape[3]), mode='nearest')
+    using_mask = interp_mask.sum() != 0.0
 
     # prepare amp scaler for mixed-precision training
     scaler = torch.cuda.amp.GradScaler()
@@ -145,7 +146,8 @@ def drag_diffusion_update(model,
                 loss += ((2*args.r_m+1)**2)*F.l1_loss(f0_patch, f1_patch)
 
             # masked region must stay unchanged
-            loss += args.lam * ((x_prev_updated-x_prev_0)*(1.0-interp_mask)).abs().sum()
+            if using_mask:
+                loss += args.lam * ((x_prev_updated-x_prev_0)*(1.0-interp_mask)).abs().sum()
             # loss += args.lam * ((init_code_orig-init_code)*(1.0-interp_mask)).abs().sum()
             print('loss total=%f'%(loss.item()))
 
@@ -210,6 +212,7 @@ def drag_diffusion_update_gen(model,
     # prepare for point tracking and background regularization
     handle_points_init = copy.deepcopy(handle_points)
     interp_mask = F.interpolate(mask, (init_code.shape[2],init_code.shape[3]), mode='nearest')
+    using_mask = interp_mask.sum() != 0.0
 
     # prepare amp scaler for mixed-precision training
     scaler = torch.cuda.amp.GradScaler()
@@ -267,7 +270,8 @@ def drag_diffusion_update_gen(model,
                 loss += ((2*args.r_m+1)**2)*F.l1_loss(f0_patch, f1_patch)
 
             # masked region must stay unchanged
-            loss += args.lam * ((x_prev_updated-x_prev_0)*(1.0-interp_mask)).abs().sum()
+            if using_mask:
+                loss += args.lam * ((x_prev_updated-x_prev_0)*(1.0-interp_mask)).abs().sum()
             # loss += args.lam * ((init_code_orig - init_code)*(1.0-interp_mask)).abs().sum()
             print('loss total=%f'%(loss.item()))