update

tests/test_mmdet_inference.py

@@ -1,61 +1,60 @@
-from mmdet import build_detector
+from mmdet.models import build_detector
+
+from mmcls.models import (MobileNetV2, MobileNetV3, RegNet, ResNeSt, ResNet,
+                          ResNeXt, SEResNet, SEResNeXt, SwinTransformer)
 
 backbone_configs = dict(
     mobilenetv2=dict(
         backbone=dict(
-            _delete_=True,
-            type='mmcv.MobileNetV2',
+            type='mmcls.MobileNetV2',
             widen_factor=1.0,
             norm_cfg=dict(type='GN', num_groups=2, requires_grad=True),
-            out_indices=range(3, 7))),
+            out_indices=(4, 7))),
     mobilenetv3=dict(
         backbone=dict(
-            _delete_=True,
-            type='mmcv.MobileNetV3',
-            widen_factor=1.0,
+            type='mmcls.MobileNetV3',
             norm_cfg=dict(type='GN', num_groups=2, requires_grad=True),
             out_indices=range(7, 12))),
-    regnet=dict(
-        backbone=dict(
-            _delete_=True, type='mmcv.RegNet', arch_name='regnetx_400mf')),
+    regnet=dict(backbone=dict(type='mmcls.RegNet', arch='regnetx_400mf')),
     resnext=dict(
         backbone=dict(
-            _delete_=True,
-            type='mmcv.ResNeXt',
-            depth=50,
-            groups=32,
-            base_width=4)),
-    resnet=dict(backbone=dict(_delete_=True, type='mmcv.ResNet', depth=50)),
-    seresnet=dict(
-        backbone=dict(_delete_=True, type='mmcv.SEResNet', depth=50)),
+            type='mmcls.ResNeXt', depth=50, groups=32, width_per_group=4)),
+    resnet=dict(backbone=dict(type='mmcls.ResNet', depth=50)),
+    seresnet=dict(backbone=dict(type='mmcls.SEResNet', depth=50)),
     seresnext=dict(
         backbone=dict(
-            _delete_=True,
-            type='mmcv.SEResNeXt',
-            depth=50,
-            groups=32,
-            base_width=4)),
+            type='mmcls.SEResNeXt', depth=50, groups=32, width_per_group=4)),
     resnest=dict(
         backbone=dict(
-            _delete_=True,
-            type='mmcv.ResNeSt',
+            type='mmcls.ResNeSt',
             depth=50,
             radix=2,
             reduction_factor=4,
             out_indices=(0, 1, 2, 3))),
     swin=dict(
         backbone=dict(
-            _delete_=True,
-            type='mmcv.SwinTransformer',
-            arch='small',
-            drop_path_rate=0.2)))
+            type='mmcls.SwinTransformer', arch='small', drop_path_rate=0.2)))
+
+module_mapping = {
+    'mobilenetv2': MobileNetV2,
+    'mobilenetv3': MobileNetV3,
+    'regnet': RegNet,
+    'resnext': ResNeXt,
+    'resnet': ResNet,
+    'seresnext': SEResNeXt,
+    'seresnet': SEResNet,
+    'resnest': ResNeSt,
+    'swin': SwinTransformer
+}
 
 
 def test_mmdet_inference():
     from mmcv import Config
-    config_path = './test/data/retinanet.py'
+    config_path = './tests/data/retinanet.py'
     config = Config.fromfile(config_path)
 
-    for backbone_config in backbone_configs.items():
-        config.merge_from_dict(backbone_config)
-        build_detector(config)
+    for module_name, backbone_config in backbone_configs.items():
+        config.model.backbone = backbone_config['backbone']
+        model = build_detector(config.model)
+        module = module_mapping[module_name]
+        assert isinstance(model.backbone, module)