fix legacy typos in code

src/lib/opts.py

@@ -273,6 +273,7 @@ class opts(object):
     opt.exp_dir = os.path.join(opt.root_dir, 'exp', opt.task)
     opt.save_dir = os.path.join(opt.exp_dir, opt.exp_id)
     opt.debug_dir = os.path.join(opt.save_dir, 'debug')
+    print('The output will be saved to ', opt.save_dir)
     
     if opt.resume and opt.load_model == '':
       model_path = opt.save_dir[:-4] if opt.save_dir.endswith('TEST') \
@@ -354,7 +355,7 @@ class opts(object):
       def __init__(self, entries):
         for k, v in entries.items():
           self.__setattr__(k, v)
-    opt = self.parse()
+    opt = self.parse(args)
     dataset = Struct(default_dataset_info[opt.task])
     opt.dataset = dataset.dataset
     opt = self.update_dataset_info_and_set_heads(opt, dataset)

src/lib/trains/ddd.py

@@ -77,7 +77,7 @@ class DddTrainer(BaseTrainer):
       opt = self.opt
       wh = output['wh'] if opt.reg_bbox else None
       reg = output['reg'] if opt.reg_offset else None
-      dets = ctadd_decode(output['hm'], output['rot'], output['dep'],
+      dets = ddd_decode(output['hm'], output['rot'], output['dep'],
                           output['dim'], wh=wh, reg=reg, K=opt.K)
 
       # x, y, score, r1-r8, depth, dim1-dim3, cls
@@ -86,10 +86,10 @@ class DddTrainer(BaseTrainer):
       # x, y, score, rot, depth, dim1, dim2, dim3
       # if opt.dataset == 'gta':
       #   dets[:, 12:15] /= 3
-      dets_pred = ctadd_post_process(
+      dets_pred = ddd_post_process(
         dets.copy(), batch['meta']['c'].detach().numpy(), 
         batch['meta']['s'].detach().numpy(), calib, opt)
-      dets_gt = ctadd_post_process(
+      dets_gt = ddd_post_process(
         batch['meta']['gt_det'].detach().numpy().copy(),
         batch['meta']['c'].detach().numpy(), 
         batch['meta']['s'].detach().numpy(), calib, opt)
@@ -138,14 +138,14 @@ class DddTrainer(BaseTrainer):
     opt = self.opt
     wh = output['wh'] if opt.reg_bbox else None
     reg = output['reg'] if opt.reg_offset else None
-    dets = ctadd_decode(output['hm'], output['rot'], output['dep'],
+    dets = ddd_decode(output['hm'], output['rot'], output['dep'],
                         output['dim'], wh=wh, reg=reg, K=opt.K)
 
     # x, y, score, r1-r8, depth, dim1-dim3, cls
     dets = dets.detach().cpu().numpy().reshape(1, -1, dets.shape[2])
     calib = batch['meta']['calib'].detach().numpy()
     # x, y, score, rot, depth, dim1, dim2, dim3
-    dets_pred = ctadd_post_process(
+    dets_pred = ddd_post_process(
       dets.copy(), batch['meta']['c'].detach().numpy(), 
       batch['meta']['s'].detach().numpy(), calib, opt)
     img_id = batch['meta']['img_id'].detach().numpy()[0]

src/lib/utils/debugger.py

@@ -222,16 +222,16 @@ class Debugger(object):
     else:
       self.ax = None
       nImgs = len(self.imgs)
-      fig=plt.figure(figsize=(nImgs * 10,10))
+      fig=self.plt.figure(figsize=(nImgs * 10,10))
       nCols = nImgs
       nRows = nImgs // nCols
       for i, (k, v) in enumerate(self.imgs.items()):
         fig.add_subplot(1, nImgs, i + 1)
         if len(v.shape) == 3:
-          plt.imshow(cv2.cvtColor(v, cv2.COLOR_BGR2RGB))
+          self.plt.imshow(cv2.cvtColor(v, cv2.COLOR_BGR2RGB))
         else:
-          plt.imshow(v)
-      plt.show()
+          self.plt.imshow(v)
+      self.plt.show()
 
   def save_img(self, imgId='default', path='./cache/debug/'):
     cv2.imwrite(path + '{}.png'.format(imgId), self.imgs[imgId])