code
stringlengths
3
6.57k
cv2.imshow(str(c)
cpu()
numpy()
copy()
int(realy)
int(realx)
int(realy + height)
int(realx)
int(top_x - height * 0.41/2)
int(top_y)
int(down_x + height * 0.41/2)
cv2.rectangle(img_now, top_left, down_right, (255, 255, 125)
img_nows.append(img_now)
cv2.imshow(str(i)
cv2.waitKey(0)
show_mot_input_debug(self, img, classification_maps, scale_maps, offset_maps)
range(img.shape[0])
img.cpu()
numpy()
copy()
np.transpose(img_numpy, [1, 2, 0])
np.transpose(img_numpy, [1, 2, 0])
img_numpy.astype(np.uint8)
enumerate(strides)
img_numpy.copy()
cpu()
numpy()
copy()
cpu()
numpy()
copy()
cpu()
numpy()
copy()
cpu()
numpy()
copy()
cpu()
numpy()
copy()
cls_numpy.nonzero()
zip(xs, ys)
cv2.imshow(str(c)
cpu()
numpy()
copy()
int(realy - height/2)
int(realx)
int(realy + height/2)
int(realx)
int(top_x - height * 0.1)
int(top_y)
int(down_x + height * 0.1)
cv2.rectangle(img_now, top_left, down_right, (255, 255, 5*int(c)
cv2.putText(img_now, str(instance)
img_nows.append(img_now)
cv2.imshow(str(i)
cv2.waitKey(0)
refine(self)
hasattr(self, 'refine_head')
type(img)
dict()
self.extract_feat(img)
self.show_input_debug(img, classification_maps, scale_maps, offset_maps)
self.show_input_debug_caltech(img, classification_maps, scale_maps, offset_maps)
self.show_mot_input_debug(img, classification_maps, scale_maps, offset_maps)
self.show_input_debug_head(img, classification_maps, scale_maps, offset_maps)
self.bbox_head(x)
losses.update(losses_bbox)
tuple([i.detach()
self.bbox_head.get_bboxes(*bbox_inputs, no_strides=False)
bbox2result(det_bboxes, det_labels, self.bbox_head.num_classes)
build_assigner(self.train_cfg.rcnn.assigner)
img.size(0)
range(num_imgs)
range(num_imgs)
torch.tensor(bbox_list[i])
float()
cuda()
sampling_results.append(sampling_result)
len(samp_list)
losses.update(dict(loss_refine_cls=torch.tensor(0)
float()
cuda()
torch.tensor(0)
float()
cuda()
bbox2roi(samp_list)
float()
torch.cat([torch.tensor(bbox[:, 4])
float()
cuda()
bbox2roi([torch.tensor(bbox)
float()
cuda()
self.refine_head(pred_feats)
self.refine_head.compute_opinion_loss(pred_scores, pred_scores_refine)
losses.update(loss_opinion)
self.refine_head(bbox_feats)
losses.update(dict(loss_refine_cls=loss_refine["loss_cls"], distL1=loss_refine["dist"])