Lint whole repo

tasks/race/data.py

@@ -39,16 +39,13 @@ class RaceDataset(Dataset):
         print_rank_0('  >> total number of samples: {}'.format(
             len(self.samples)))
 
-
     def __len__(self):
         return len(self.samples)
 
-
     def __getitem__(self, idx):
         return self.samples[idx]
 
 
-
 def process_single_datapath(datapath, tokenizer, max_qa_length, max_seq_length):
     """Read in RACE files, combine, clean-up, tokenize, and convert to
     samples."""

tasks/zeroshot_gpt2/datasets.py

@@ -64,12 +64,12 @@ class _LMDataset(torch.utils.data.Dataset):
     def __getitem__(self, idx):
         start_idx = idx * self.overalapping_eval
         end_idx = start_idx + self.seq_len
-        tokens = self.tokens[start_idx:end_idx+1]
+        tokens = self.tokens[start_idx:end_idx + 1]
         num_tokens = len(tokens)
-        pad_mask = [1]*num_tokens
-        if num_tokens < self.seq_len+1:
-            num_pad = (self.seq_len+1-num_tokens)
-            pad_mask += [0]*(num_pad)
+        pad_mask = [1] * num_tokens
+        if num_tokens < self.seq_len + 1:
+            num_pad = (self.seq_len + 1 - num_tokens)
+            pad_mask += [0] * (num_pad)
             tokens += [self.pad_idx] * num_pad
         pad_mask = np.array(pad_mask[1:])
         if self.overalapping_eval != self.seq_len and idx != 0:
@@ -103,7 +103,7 @@ class _LambadaDataset(torch.utils.data.Dataset):
         last_token = text.split()[-1]
         start_idx = text.rfind(last_token)
         beginning_tokens = self.tokenizer.tokenize(text[:start_idx].strip())
-        last_token = self.tokenizer.tokenize(' '+last_token)
+        last_token = self.tokenizer.tokenize(' ' + last_token)
         return beginning_tokens, last_token
 
     def __len__(self):
@@ -112,14 +112,14 @@ class _LambadaDataset(torch.utils.data.Dataset):
     def __getitem__(self, idx):
         tokens = self.tokens[idx]
         num_tokens = len(tokens)
-        pad_mask = [0]*num_tokens
+        pad_mask = [0] * num_tokens
         labels = self.labels[idx]
-        pad_mask += [1]*len(labels)
-        tokens = tokens+labels
+        pad_mask += [1] * len(labels)
+        tokens = tokens + labels
         num_tokens = len(tokens)
-        if num_tokens < self.seq_len+1:
-            num_pad = (self.seq_len+1-num_tokens)
-            pad_mask += [0]*(num_pad)
+        if num_tokens < self.seq_len + 1:
+            num_pad = (self.seq_len + 1 - num_tokens)
+            pad_mask += [0] * (num_pad)
             tokens += [self.pad_idx] * num_pad
         pad_mask = np.array(pad_mask[1:])
 

tasks/zeroshot_gpt2/detokenizer.py

@@ -23,21 +23,21 @@ def ptb_detokenizer(string):
     string = string.replace(" \n", "\n")
     string = string.replace("\n ", "\n")
     string = string.replace(" n't", "n't")
-	string = string.replace(" N ","1 ")
+    string = string.replace(" N ", "1 ")
     string = string.replace("$ 1", "$1")
     string = string.replace("# 1", "#1")
     return string
 
 
 def wikitext_detokenizer(string):
-	#contractions
+    # contractions
     string = string.replace("s '", "s'")
     string = re.sub(r"/' [0-9]/", r"/'[0-9]/", string)
     # number separators
     string = string.replace(" @-@ ", "-")
     string = string.replace(" @,@ ", ",")
     string = string.replace(" @.@ ", ".")
-	#punctuation
+    # punctuation
     string = string.replace(" : ", ": ")
     string = string.replace(" ; ", "; ")
     string = string.replace(" . ", ". ")
@@ -54,7 +54,7 @@ def wikitext_detokenizer(string):
     string = string.replace("= = = =", "====")
     string = string.replace("= = =", "===")
     string = string.replace("= =", "==")
-	string = string.replace(" "+chr(176)+" ", chr(176))
+    string = string.replace(" " + chr(176) + " ", chr(176))
     string = string.replace(" \n", "\n")
     string = string.replace("\n ", "\n")
     string = string.replace(" N ", " 1 ")
@@ -78,5 +78,3 @@ def get_detokenizer(path):
     for key in _DETOKENIZERS.keys():
         if key in path:
             return _DETOKENIZERS[key]
-
-