update readme and MLP model.

README.md

@@ -1,4 +1,4 @@
-# torch cross modal (torchcmh)
+# torch cross modal hashing (torchcmh)
 
 torchcmh is a library built on PyTorch for deep learning cross modal hashing.\
 if you want use the dataset i make, please download mat file and image file by readme file in dataset package.
@@ -11,14 +11,20 @@ you need to install these package to run
 - create a configuration file as ./script/default_config.yml
 ```yaml
 training:
+  # the name of python file in training
   method: DCMH
+  # the data set name, you can choose mirflickr25k, nus wide, ms coco, iapr tc-12
   dataName: Mirflickr25K
   batchSize: 64
+  # the bit of hash codes
   bit: 64
+  # if true, the program will be run on gpu
   cuda: True
+  # the device id you want to use, if you want to multi gpu, you can use [id1, id2]
   device: 0
 datasetPath:
   Mirflickr25k:
+    # the path you download the image of data set.
     img_dir: I:\dataset\mirflickr25k\mirflickr
 
 ```

torchcmh/dataset/__init__.py

@@ -10,11 +10,11 @@
 def triplet_data(dataset_name: str, img_dir: str, **kwargs):
     if dataset_name.lower() == 'mirflickr25k':
         from torchcmh.dataset.mirflckr25k import get_triplet_datasets
-    elif dataset_name.lower() == 'nus wide':
+    elif dataset_name.lower() in ['nus wide', 'nuswide']:
         from torchcmh.dataset.nus_wide import get_triplet_datasets
-    elif dataset_name.lower() == 'coco2014':
+    elif dataset_name.lower() in ['coco2014', 'coco', 'mscoco', 'ms coco']:
         from torchcmh.dataset.coco2014 import get_triplet_datasets
-    elif dataset_name.lower() == 'iapr tc-12':
+    elif dataset_name.lower() in ['iapr tc-12', 'iapr', 'tc-12', 'tc12']:
         from torchcmh.dataset.tc12 import get_triplet_datasets
     else:
         raise ValueError("there is no dataset name is %s" % dataset_name)
@@ -25,11 +25,11 @@ def triplet_data(dataset_name: str, img_dir: str, **kwargs):
 def single_data(dataset_name: str, img_dir: str, **kwargs):
     if dataset_name.lower() == 'mirflickr25k':
         from torchcmh.dataset.mirflckr25k import get_single_datasets
-    elif dataset_name.lower() == 'nus wide':
+    elif dataset_name.lower() in ['nus wide', 'nuswide']:
         from torchcmh.dataset.nus_wide import get_single_datasets
-    elif dataset_name.lower() == 'coco2014':
+    elif dataset_name.lower() in ['coco2014', 'coco', 'mscoco', 'ms coco']:
         from torchcmh.dataset.coco2014 import get_single_datasets
-    elif dataset_name.lower() == 'iapr tc-12':
+    elif dataset_name.lower() in ['iapr tc-12', 'iapr', 'tc-12', 'tc12']:
         from torchcmh.dataset.tc12 import get_single_datasets
     else:
         raise ValueError("there is no dataset name is %s" % dataset_name)
@@ -40,11 +40,11 @@ def single_data(dataset_name: str, img_dir: str, **kwargs):
 def pairwise_data(dataset_name: str, img_dir: str, **kwargs):
     if dataset_name.lower() == 'mirflickr25k':
         from torchcmh.dataset.mirflckr25k import get_pairwise_datasets
-    elif dataset_name.lower() == 'nus wide':
+    elif dataset_name.lower() in ['nus wide', 'nuswide']:
         from torchcmh.dataset.nus_wide import get_pairwise_datasets
-    elif dataset_name.lower() == 'coco2014':
+    elif dataset_name.lower() in ['coco2014', 'coco', 'mscoco', 'ms coco']:
         from torchcmh.dataset.coco2014 import get_pairwise_datasets
-    elif dataset_name.lower() == 'iapr tc-12':
+    elif dataset_name.lower() in ['iapr tc-12', 'iapr', 'tc-12', 'tc12']:
         from torchcmh.dataset.tc12 import get_pairwise_datasets
     else:
         raise ValueError("there is no dataset name is %s" % dataset_name)

torchcmh/models/.gitignore

@@ -10,5 +10,8 @@ MSText_capsnet.py
 MSText_norm.py
 pcb.py
 resnet_gcn.py
-resnet_norm.py
 pretrain_model
+ASCHN
+ACMH
+GCH
+LCMH

torchcmh/models/MLP.py

@@ -20,7 +20,7 @@ def weights_init(m):
 
 
 class MLP(BasicModule):
-    def __init__(self, input_dim, output_dim, hidden_nodes=[8192], dropout=None, leakRelu=False):
+    def __init__(self, input_dim, output_dim, hidden_nodes=[8192], dropout=None, leakRelu=False, norm=False):
         """
         :param input_dim: dimension of input
         :param output_dim: bit number of the final binary code
@@ -35,6 +35,8 @@ def __init__(self, input_dim, output_dim, hidden_nodes=[8192], dropout=None, lea
             kernel_size = input_dim if in_channel == 1 else 1
             full_conv_layers.append(nn.Conv1d(in_channel, hidden_node, kernel_size=kernel_size, stride=1))
             in_channel = hidden_node
+            if norm:
+                full_conv_layers.append(nn.BatchNorm1d(hidden_node))
             if dropout:
                 full_conv_layers.append(nn.Dropout(dropout))
             if leakRelu:

torchcmh/models/resnet.py

@@ -114,7 +114,6 @@ def __init__(self, num_classes, loss, block, layers,
                  dropout_p=None,
                  **kwargs):
         self.inplanes = 64
-        self.out_feature = kwargs['out_feature'] if 'out_feature' in kwargs.keys() else False
         super(ResNet, self).__init__()
         self.loss = loss
         self.feature_dim = 512 * block.expansion
@@ -207,7 +206,7 @@ def featuremaps(self, x):
         x = self.layer4(x)
         return x
 
-    def forward(self, x):
+    def forward(self, x, out_feature=False):
         f = self.featuremaps(x)
         v = self.global_avgpool(f)
         v = v.view(v.size(0), -1)
@@ -216,7 +215,7 @@ def forward(self, x):
             v = self.fc(v)
 
         y = self.classifier(v)
-        if self.out_feature and self.training:
+        if out_feature:
             return y, v
         return y
 

torchcmh/utils/plotter.py

@@ -8,11 +8,12 @@
 
 class VisdomLinePlotter(object):
     """Plots to Visdom"""
-    def __init__(self, env_name='challenge'):
+    def __init__(self, env_name='plotter'):
         self.viz = Visdom()
         self.env = env_name
         self.plots = {}
         self.epoch = 0
+
     def plot(self, var_name, split_name, y, x=None):
         if x is None:
             x = self.epoch
@@ -25,9 +26,12 @@ def plot(self, var_name, split_name, y, x=None):
             ))
         else:
             self.viz.line(X=np.array([x]), Y=np.array([y]), env=self.env, win=self.plots[var_name], name=split_name, update='append')
+
     def next_epoch(self):
         self.epoch += 1
 
+    def reset_epoch(self):
+        self.epoch = 0
 
 def get_plotter(env_name: str):
     return VisdomLinePlotter(env_name)