Flake8

fairseq/models/fconv.py

@@ -122,9 +122,9 @@ class FConvEncoder(FairseqEncoder):
             self.projections.append(Linear(in_channels, out_channels)
                                     if in_channels != out_channels else None)
             if kernel_size % 2 == 1:
-               padding = kernel_size //2
+                padding = kernel_size // 2
             else:
-               padding = 0
+                padding = 0
             self.convolutions.append(
                 ConvTBC(in_channels, out_channels * 2, kernel_size,
                         dropout=dropout, padding=padding)

fairseq/models/lstm.py

@@ -105,7 +105,7 @@ class LSTMModel(FairseqModel):
 class LSTMEncoder(FairseqEncoder):
     """LSTM encoder."""
     def __init__(self, dictionary, embed_dim=512, embed_dict=None,
-        num_layers=1, dropout_in=0.1, dropout_out=0.1):
+                 num_layers=1, dropout_in=0.1, dropout_out=0.1):
         super().__init__(dictionary)
         self.num_layers = num_layers
         self.dropout_in = dropout_in
@@ -117,8 +117,7 @@ class LSTMEncoder(FairseqEncoder):
         self.padding_idx = dictionary.pad()
         self.embed_tokens = Embedding(num_embeddings, embed_dim, self.padding_idx)
         if embed_dict:
-            self.embed_tokens = utils.load_embedding(
-                embed_dict, self.dictionary, self.embed_tokens)
+            self.embed_tokens = utils.load_embedding(embed_dict, self.dictionary, self.embed_tokens)
 
         self.lstm = LSTM(
             input_size=embed_dim,
@@ -233,7 +232,7 @@ class AttentionLayer(nn.Module):
 
 class LSTMDecoder(FairseqIncrementalDecoder):
     """LSTM decoder."""
-    def __init__(self, dictionary, encoder_embed_dim=512, 
+    def __init__(self, dictionary, encoder_embed_dim=512,
                  embed_dim=512, embed_dict=None,
                  out_embed_dim=512, num_layers=1, dropout_in=0.1,
                  dropout_out=0.1, attention=True):
@@ -246,9 +245,7 @@ class LSTMDecoder(FairseqIncrementalDecoder):
         padding_idx = dictionary.pad()
         self.embed_tokens = Embedding(num_embeddings, embed_dim, padding_idx)
         if embed_dict:
-            self.embed_tokens = utils.load_embedding(
-                embed_dict, self.dictionary, self.embed_tokens)
-
+            self.embed_tokens = utils.load_embedding(embed_dict, self.dictionary, self.embed_tokens)
 
         self.layers = nn.ModuleList([
             LSTMCell(

fairseq/utils.py

@@ -258,10 +258,11 @@ def load_align_dict(replace_unk):
 
 
 def print_embed_overlap(embed_dict, vocab_dict):
-     embed_keys = set(embed_dict.keys())
-     vocab_keys = set(vocab_dict.symbols)
-     overlap = len(embed_keys & vocab_keys)
-     print("| Found {}/{} types in embedding file.".format(overlap, len(vocab_dict)))
+    embed_keys = set(embed_dict.keys())
+    vocab_keys = set(vocab_dict.symbols)
+    overlap = len(embed_keys & vocab_keys)
+    print("| Found {}/{} types in embedding file.".format(overlap, len(vocab_dict)))
+
 
 def parse_embedding(embed_path):
     """Parse embedding text file into a dictionary of word and embedding tensors.
@@ -276,12 +277,13 @@ def parse_embedding(embed_path):
     """
     embed_dict = dict()
     with open(embed_path) as f_embed:
-        _ = next(f_embed) #skip header
+        _ = next(f_embed)  # skip header
         for line in f_embed:
             pieces = line.strip().split()
             embed_dict[pieces[0]] = torch.Tensor([float(weight) for weight in pieces[1:]])
     return embed_dict
 
+
 def load_embedding(embed_dict, vocab, embedding):
     for idx in range(len(vocab)):
         token = vocab[idx]
@@ -289,6 +291,7 @@ def load_embedding(embed_dict, vocab, embedding):
             embedding.weight.data[idx] = embed_dict[token]
     return embedding
 
+
 def replace_unk(hypo_str, src_str, alignment, align_dict, unk):
     from fairseq import tokenizer
     # Tokens are strings here