[1.0.8] load_unlabeled_targets -> load_train_targets

README.rst

@@ -38,7 +38,7 @@ You can find some research projects that are built on top of Torchreid `here <ht
 
 What's new
 ---------------
-- [Nov 19] ``ImageDataManager`` can load training data from target datasets by setting ``load_unlabeled_targets=True``, and the train-loader can be accessed with ``train_loader_u = datamanager.train_loader_u``. This feature is useful for domain adaptation research.
+- [Nov 19] ``ImageDataManager`` can load training data from target datasets by setting ``load_train_targets=True``, and the train-loader can be accessed with ``train_loader_t = datamanager.train_loader_t``. This feature is useful for domain adaptation research.
 
 
 Installation

torchreid/__init__.py

@@ -1,7 +1,7 @@
 from __future__ import absolute_import
 from __future__ import print_function
 
-__version__ = '1.0.7'
+__version__ = '1.0.8'
 __author__ = 'Kaiyang Zhou'
 __homepage__ = 'https://kaiyangzhou.github.io/'
 __description__ = 'Deep learning person re-identification in PyTorch'

torchreid/data/datamanager.py

@@ -93,8 +93,8 @@ class ImageDataManager(DataManager):
         split_id (int, optional): split id (*0-based*). Default is 0.
         combineall (bool, optional): combine train, query and gallery in a dataset for
             training. Default is False.
-        load_unlabeled_targets (bool, optional): construct train loader for unlabeled target
-            datasets. Default is False.
+        load_train_targets (bool, optional): construct train-loader for target datasets.
+            Default is False. This is useful for domain adaptation research.
         batch_size_train (int, optional): number of images in a training batch. Default is 32.
         batch_size_test (int, optional): number of images in a test batch. Default is 32.
         workers (int, optional): number of workers. Default is 4.
@@ -126,7 +126,7 @@ class ImageDataManager(DataManager):
         test_loader = datamanager.test_loader
 
         # return train loader of target data
-        train_loader_u = datamanager.train_loader_u
+        train_loader_t = datamanager.train_loader_t
     """
     data_type = 'image'
 
@@ -143,7 +143,7 @@ class ImageDataManager(DataManager):
         use_gpu=True,
         split_id=0,
         combineall=False,
-        load_unlabeled_targets=False,
+        load_train_targets=False,
         batch_size_train=32,
         batch_size_test=32,
         workers=4,
@@ -194,33 +194,33 @@ class ImageDataManager(DataManager):
             drop_last=True
         )
 
-        self.train_loader_u = None
-        if load_unlabeled_targets:
+        self.train_loader_t = None
+        if load_train_targets:
             # check if sources and targets are identical
             assert len(set(self.sources) & set(self.targets)) == 0, \
                 'sources={} and targets={} must not have overlap'.format(self.sources, self.targets)
 
             print('=> Loading train (target) dataset')
-            trainset_u = []
+            trainset_t = []
             for name in self.targets:
-                trainset_u_ = init_image_dataset(
+                trainset_t_ = init_image_dataset(
                     name,
                     transform=self.transform_tr,
                     mode='train',
-                    combineall=combineall,
+                    combineall=False, # only use the training data
                     root=root,
                     split_id=split_id,
                     cuhk03_labeled=cuhk03_labeled,
                     cuhk03_classic_split=cuhk03_classic_split,
                     market1501_500k=market1501_500k
                 )
-                trainset_u.append(trainset_u_)
-            trainset_u = sum(trainset_u)
+                trainset_t.append(trainset_t_)
+            trainset_t = sum(trainset_t)
 
-            self.train_loader_u = torch.utils.data.DataLoader(
-                trainset_u,
+            self.train_loader_t = torch.utils.data.DataLoader(
+                trainset_t,
                 sampler=build_train_sampler(
-                    trainset_u.train, train_sampler,
+                    trainset_t.train, train_sampler,
                     batch_size=batch_size_train,
                     num_instances=num_instances
                 ),
@@ -289,9 +289,9 @@ class ImageDataManager(DataManager):
         print('  # source ids      : {}'.format(self.num_train_pids))
         print('  # source images   : {}'.format(len(trainset)))
         print('  # source cameras  : {}'.format(self.num_train_cams))
+        if load_train_targets:
+            print('  # target images   : {} (unlabeled)'.format(len(trainset_t)))
         print('  target            : {}'.format(self.targets))
-        if load_unlabeled_targets:
-            print('  # target images   : {} (unlabeled)'.format(len(trainset_u)))
         print('  *****************************************')
         print('\n')