diff --git a/mmdeploy/codebase/mmdet/models/dense_heads/sparseinst_head.py b/mmdeploy/codebase/mmdet/models/dense_heads/sparseinst_head.py
index 6ddbb7744c..8e4c3e487d 100644
--- a/mmdeploy/codebase/mmdet/models/dense_heads/sparseinst_head.py
+++ b/mmdeploy/codebase/mmdet/models/dense_heads/sparseinst_head.py
@@ -1,11 +1,8 @@
 # Copyright (c) OpenMMLab. All rights reserved.
-from typing import Dict, List, Tuple
+from typing import List
 
 import torch
 import torch.nn.functional as F
-from mmdet.models.utils import aligned_bilinear
-from mmdet.structures import OptSampleList, SampleList
-from mmengine.config import ConfigDict
 from torch import Tensor
 
 from mmdeploy.core import FUNCTION_REWRITER
@@ -30,7 +27,7 @@ def sparseinst__predict(
     max_shape = batch_inputs.shape[-2:]
     x = self.extract_feat(batch_inputs)
     output = self.decoder(x)
-    
+
     pred_scores = output['pred_logits'].sigmoid()
     pred_masks = output['pred_masks'].sigmoid()
     pred_objectness = output['pred_scores'].sigmoid()
@@ -47,15 +44,13 @@ def sparseinst__predict(
 
     img_meta = batch_data_samples[0].metainfo
     # rescoring mask using maskness
-    scores = rescoring_mask(scores,
-                            pred_masks > self.mask_threshold,
+    scores = rescoring_mask(scores, pred_masks > self.mask_threshold,
                             pred_masks)
     h, w = img_meta['img_shape'][:2]
-    pred_masks = F.interpolate(pred_masks,
-                               size=max_shape,
-                               mode='bilinear',
-                               align_corners=False)[:, :, :h, :w]
-    
+    pred_masks = F.interpolate(
+        pred_masks, size=max_shape, mode='bilinear',
+        align_corners=False)[:, :, :h, :w]
+
     bboxes = torch.zeros(scores.shape[0], scores.shape[1], 4)
     dets = torch.cat([bboxes, scores.unsqueeze(-1)], dim=-1)
     masks = (pred_masks > self.mask_threshold).float()