pre-commit update

configs/mmdet/instance-seg/instance-seg_tensorrt_dynamic-320x320-1344x1344.py

@@ -4,7 +4,7 @@
 ]
 
 backend_config = dict(
-    common_config=dict(max_workspace_size=1 << 30),
+    common_config=dict(max_workspace_size=1 << 32),
     model_inputs=[
         dict(
             input_shapes=dict(

mmdeploy/codebase/mmdet/models/dense_heads/__init__.py

@@ -1,6 +1,7 @@
 # Copyright (c) OpenMMLab. All rights reserved.
 from . import base_dense_head  # noqa: F401,F403
 from . import centernet_head  # noqa: F401,F403
+from . import condinst_head  # noqa: F401,F403
 from . import detr_head  # noqa: F401,F403
 from . import fovea_head  # noqa: F401,F403
 from . import gfl_head  # noqa: F401,F403
@@ -12,4 +13,3 @@
 from . import solov2_head  # noqa: F401,F403
 from . import yolo_head  # noqa: F401,F403
 from . import yolox_head  # noqa: F401,F403
-from . import condinst_head # noqa: F401,F403

mmdeploy/codebase/mmdet/models/dense_heads/condinst_head.py

@@ -137,8 +137,8 @@ def condinst_mask_head__forward(self, x: tuple,
     weights, biases = _parse_dynamic_params(self, param_preds)
     mask_preds = _dynamic_conv_forward(mask_head_inputs, weights, biases)
     mask_preds = mask_preds.reshape(batch_size, num_insts, hw[0], hw[1])
-    mask_preds = aligned_bilinear(mask_preds, 
-                                  int(self.mask_feat_stride / self.mask_out_stride))
+    mask_preds = aligned_bilinear(
+        mask_preds, int(self.mask_feat_stride / self.mask_out_stride))
     return (mask_preds, )
 
 

mmdeploy/pytorch/functions/repeat.py

@@ -20,7 +20,8 @@ def tensor__repeat__tensorrt(input: torch.Tensor, *size: Union[torch.Size,
     origin_func = ctx.origin_func
     if input.dim() == 1 and len(size) == 1:
         if isinstance(*size, tuple):
-            return origin_func(input.unsqueeze(0), *([1] + list(*size))).squeeze(0)
+            return origin_func(input.unsqueeze(0),
+                               *([1] + list(*size))).squeeze(0)
         return origin_func(input.unsqueeze(0), *([1] + list(size))).squeeze(0)
     else:
         return origin_func(input, *size)

tests/test_codebase/test_mmdet/test_mmdet_models.py

@@ -2487,7 +2487,7 @@ def test_condinst_bbox_head_predict_by_feat(backend_type):
         wrapped_model=wrapped_model,
         model_inputs=rewrite_inputs,
         deploy_cfg=deploy_cfg)
-    
+
     if is_backend_output:
         dets = rewrite_outputs[0]
         labels = rewrite_outputs[1]
@@ -2571,18 +2571,18 @@ def test_condinst_mask_head_predict_by_feat(backend_type):
                                 max_shape=[1, 3, 1344, 1344])))
                 ]),
             onnx_config=dict(output_names=output_names, input_shape=None),
-            codebase_config=dict(
-                type='mmdet',
-                task='ObjectDetection')))
-
+            codebase_config=dict(type='mmdet', task='ObjectDetection')))
+
     class TestCondInstMaskHeadModel(torch.nn.Module):
+
         def __init__(self, condinst_mask_head):
             super(TestCondInstMaskHeadModel, self).__init__()
             self.mask_head = condinst_mask_head
 
         def predict_by_feat(self, mask_preds, det, label, batch_img_metas):
             results = dict(dets=det, labels=label)
-            return self.mask_head.predict_by_feat(mask_preds, results, batch_img_metas)
+            return self.mask_head.predict_by_feat(mask_preds, results,
+                                                  batch_img_metas)
 
     head = get_condinst_mask_head()
     condinst_mask_head = TestCondInstMaskHeadModel(head)
@@ -2597,11 +2597,7 @@ def predict_by_feat(self, mask_preds, det, label, batch_img_metas):
     # to get outputs of onnx/tensorrt model after rewrite
     wrapped_model = WrapModel(
         condinst_mask_head, 'predict_by_feat', batch_img_metas=batch_img_metas)
-    rewrite_inputs = {
-        'mask_preds': mask_preds,
-        'det': dets,
-        'label': labels
-    }
+    rewrite_inputs = {'mask_preds': mask_preds, 'det': dets, 'label': labels}
     rewrite_outputs, is_backend_output = get_rewrite_outputs(
         wrapped_model=wrapped_model,
         model_inputs=rewrite_inputs,