diff --git a/invokeai/backend/stable_diffusion/diffusion/regional_ip_data.py b/invokeai/backend/stable_diffusion/diffusion/regional_ip_data.py
index d3b4505f587..792c97114da 100644
--- a/invokeai/backend/stable_diffusion/diffusion/regional_ip_data.py
+++ b/invokeai/backend/stable_diffusion/diffusion/regional_ip_data.py
@@ -59,8 +59,11 @@ def _prepare_masks(
             if downscale_factor <= max_downscale_factor:
                 # We use max pooling because we downscale to a pretty low resolution, so we don't want small mask
                 # regions to be lost entirely.
+                #
+                # ceil_mode=True is set to mirror the downsampling behavior of SD and SDXL.
+                #
                 # TODO(ryand): In the future, we may want to experiment with other downsampling methods.
-                mask_tensor = torch.nn.functional.max_pool2d(mask_tensor, kernel_size=2, stride=2)
+                mask_tensor = torch.nn.functional.max_pool2d(mask_tensor, kernel_size=2, stride=2, ceil_mode=True)
 
         return masks_by_seq_len
 
diff --git a/invokeai/backend/stable_diffusion/diffusion/regional_prompt_data.py b/invokeai/backend/stable_diffusion/diffusion/regional_prompt_data.py
index 85331013d5f..f09cc0a0d21 100644
--- a/invokeai/backend/stable_diffusion/diffusion/regional_prompt_data.py
+++ b/invokeai/backend/stable_diffusion/diffusion/regional_prompt_data.py
@@ -61,9 +61,12 @@ def _prepare_spatial_masks(
                 if downscale_factor <= max_downscale_factor:
                     # We use max pooling because we downscale to a pretty low resolution, so we don't want small prompt
                     # regions to be lost entirely.
+                    #
+                    # ceil_mode=True is set to mirror the downsampling behavior of SD and SDXL.
+                    #
                     # TODO(ryand): In the future, we may want to experiment with other downsampling methods (e.g.
                     # nearest interpolation), and could potentially use a weighted mask rather than a binary mask.
-                    batch_sample_masks = F.max_pool2d(batch_sample_masks, kernel_size=2, stride=2)
+                    batch_sample_masks = F.max_pool2d(batch_sample_masks, kernel_size=2, stride=2, ceil_mode=True)
 
         return batch_sample_masks_by_seq_len