add self.with_neck (open-mmlab#796)

mmaction/models/recognizers/base.py

@@ -91,6 +91,11 @@ def __init__(self,
 
  self.fp16_enabled = False
 
+ @property
+ def with_neck(self):
+ """bool: whether the detector has a neck"""
+ return hasattr(self, 'neck') and self.neck is not None
+
  def init_weights(self):
  """Initialize the model network weights."""
  if self.backbone_from in ['mmcls', 'mmaction2']:
@@ -105,7 +110,7 @@ def init_weights(self):
  f'{self.backbone_from}!')
 
  self.cls_head.init_weights()
- if hasattr(self, 'neck'):
+ if self.with_neck:
  self.neck.init_weights()
 
  @auto_fp16()

mmaction/models/recognizers/recognizer2d.py

@@ -26,7 +26,7 @@ def forward_train(self, imgs, labels, **kwargs):
  x = x.reshape((x.shape[0], -1))
  x = x.reshape(x.shape + (1, 1))
 
- if hasattr(self, 'neck'):
+ if self.with_neck:
  x = [
  each.reshape((-1, num_segs) +
  each.shape[1:]).transpose(1, 2).contiguous()
@@ -60,7 +60,7 @@ def _do_test(self, imgs):
  x = x.reshape((x.shape[0], -1))
  x = x.reshape(x.shape + (1, 1))
 
- if hasattr(self, 'neck'):
+ if self.with_neck:
  x = [
  each.reshape((-1, num_segs) +
  each.shape[1:]).transpose(1, 2).contiguous()
@@ -97,7 +97,7 @@ def _do_fcn_test(self, imgs):
  imgs = torch.flip(imgs, [-1])
  x = self.extract_feat(imgs)
 
- if hasattr(self, 'neck'):
+ if self.with_neck:
  x = [
  each.reshape((-1, num_segs) +
  each.shape[1:]).transpose(1, 2).contiguous()
@@ -147,7 +147,7 @@ def forward_dummy(self, imgs, softmax=False):
  num_segs = imgs.shape[0] // batches
 
  x = self.extract_feat(imgs)
- if hasattr(self, 'neck'):
+ if self.with_neck:
  x = [
  each.reshape((-1, num_segs) +
  each.shape[1:]).transpose(1, 2).contiguous()

mmaction/models/recognizers/recognizer3d.py

@@ -15,7 +15,7 @@ def forward_train(self, imgs, labels, **kwargs):
  losses = dict()
 
  x = self.extract_feat(imgs)
- if hasattr(self, 'neck'):
+ if self.with_neck:
  x, loss_aux = self.neck(x, labels.squeeze())
  losses.update(loss_aux)
 
@@ -42,15 +42,15 @@ def _do_test(self, imgs):
  while view_ptr < total_views:
  batch_imgs = imgs[view_ptr:view_ptr + self.max_testing_views]
  x = self.extract_feat(batch_imgs)
- if hasattr(self, 'neck'):
+ if self.with_neck:
  x, _ = self.neck(x)
  cls_score = self.cls_head(x)
  cls_scores.append(cls_score)
  view_ptr += self.max_testing_views
  cls_score = torch.cat(cls_scores)
  else:
  x = self.extract_feat(imgs)
- if hasattr(self, 'neck'):
+ if self.with_neck:
  x, _ = self.neck(x)
  cls_score = self.cls_head(x)
 
@@ -76,7 +76,7 @@ def forward_dummy(self, imgs, softmax=False):
  imgs = imgs.reshape((-1, ) + imgs.shape[2:])
  x = self.extract_feat(imgs)
 
- if hasattr(self, 'neck'):
+ if self.with_neck:
  x, _ = self.neck(x)
 
  outs = self.cls_head(x)