not know

README.md

@@ -28,7 +28,7 @@ $ conda install python=3.7.4
 ```
 $ git clone https://github.com/yifding/hetseq.git
 $ cd /path/to/hetseq
-$ pip install -r requirement.txt 
+$ pip install -r requirements.txt 
 $ pip install --editable . 
 ```
 
@@ -72,7 +72,8 @@ this repository is MIT-licensed. It is created based on [fairseq](https://github
 
 Please send us e-mail or leave comments on github if have any questions.
 
-Copyright (c) 2020 Yifan Ding, Weninger Lab
+
+Copyright (c) 2020 Yifan Ding and [Weninger Lab](https://www3.nd.edu/~tweninge/)
 
 
 

docs/source/installing.rst

@@ -19,7 +19,7 @@ Git Clone and Install Packages
 
 	$ git clone https://github.com/yifding/hetseq.git
 	$ cd /path/to/hetseq
-	$ pip install -r requirement.txt 
+	$ pip install -r requirements.txt 
 	$ pip install --editable . 
 
 Download BERT Processed File

requirements.txt

@@ -1,11 +1,12 @@
 cython == 0.29.13
 numpy == 1.17.0
 h5py == 2.10.0
-torch == 1.2.0
+torch == 1.6.0
 tqdm == 4.36.1
-boto3 == 1.9.244
 chardet == 3.0.4
 idna == 2.8
 python-dateutil == 2.8.0
 sphinx-rtd-theme == 0.5.0
 sphinx == 3.2.1
+boto3 == 1.9.244
+torchvision == 0.7.0

tasks.py

@@ -48,7 +48,7 @@ def dataset(self, split):
         Args:
             split (str): name of the split (e.g., train, valid, test)
         Returns:
-            a :class:`~fairseq.data.FairseqDataset` corresponding to *split*
+            a :class:`~torch.utils.data.Dataset` corresponding to *split*
         """
         if split not in self.datasets:
             raise KeyError('Dataset not loaded: ' + split)
@@ -87,7 +87,7 @@ def get_batch_iterator(
             epoch (int, optional): the epoch to start the iterator from
                 (default: 0).
         Returns:
-            ~fairseq.iterators.EpochBatchIterator: a batched iterator over the
+            ~iterators.EpochBatchIterator: a batched iterator over the
                 given dataset split
         """
         # For default fairseq task, return same iterator across epochs
@@ -127,12 +127,12 @@ def get_batch_iterator(
 
     def build_model(self, args):
         """
-        Build the :class:`~fairseq.models.BaseFairseqModel` instance for this
+        Build the :class:`~torch.nn.Module` instance for this
         task.
         Args:
             args (argparse.Namespace): parsed command-line arguments
         Returns:
-            a :class:`~fairseq.models.BaseFairseqModel` instance
+            a :class:`~torch.nn.Module` instance
         """
         raise NotImplementedError
 
@@ -142,10 +142,9 @@ def train_step(self, sample, model, optimizer, ignore_grad=False):
         for the given *model* and *sample*.
         Args:
             sample (dict): the mini-batch. The format is defined by the
-                :class:`~fairseq.data.FairseqDataset`.
-            model (~fairseq.models.BaseFairseqModel): the model
-            criterion (~fairseq.criterions.FairseqCriterion): the criterion
-            optimizer (~fairseq.optim.FairseqOptimizer): the optimizer
+                :class:`~torch.utils.data.Dataset`.
+            model (~torch.nn.Module): the model
+            optimizer (~optim._Optimizer): the optimizer
             ignore_grad (bool): multiply loss by 0 if this is set to True
         Returns:
             tuple:
@@ -187,12 +186,9 @@ def update_step(self, num_updates):
 class LanguageModelingTask(Task):
     """
     Train a language model, currently support BERT.
-
-
-    Args: 
-	args: parsed from command line
-
-    	dictionary: the BPE dictionary for the input of the language model
+    Args:
+        args: parsed from command line
+        dictionary: the BPE dictionary for the input of the language model
     """
 
     def __init__(self, args, dictionary):
@@ -215,7 +211,6 @@ def setup_task(cls, args, **kwargs):
         """
         dictionary = cls.load_dictionary(cls, args.dict)
 
-        #return cls(args, dictionary, output_dictionary, targets=targets)
         return cls(args, dictionary)
 
     def build_model(self, args):
@@ -323,8 +318,6 @@ def __init__(self):
         self.fc2 = nn.Linear(128, 10)
 
     def forward(self, x, target, eval=False):
-        # print('shape', x.shape, target.shape)
-        # print(target)
 
         x = self.conv1(x)
         x = F.relu(x)
@@ -339,5 +332,4 @@ def forward(self, x, target, eval=False):
         x = self.fc2(x)
         output = F.log_softmax(x, dim=1)
         loss = F.nll_loss(output, target)
-        # return loss if not eval else output, loss
         return loss