Fix linting problems

lm_eval/models/dummy.py

@@ -20,4 +20,4 @@ class DummyLM(LM):
     
     def greedy_until(self, requests):
         # TODO: implement
-        pass
+        pass
\ No newline at end of file

lm_eval/models/gpt2.py

@@ -43,4 +43,4 @@ class GPT2LM(LM):
     
     def greedy_until(self, requests):
         # TODO: implement
-        pass
+        pass
\ No newline at end of file

lm_eval/tasks/arc.py

@@ -70,4 +70,4 @@ class ARCEasy(HFTask):
 
 class ARCChallenge(ARCEasy):
     DATASET_PATH = "ai2_arc"
-    DATASET_NAME = "ARC-Challenge"
+    DATASET_NAME = "ARC-Challenge"
\ No newline at end of file

lm_eval/tasks/drop.py

@@ -104,4 +104,4 @@ class DROP(Dataset):
             whether a higher value of the submetric is better
         """
         # TODO: implement evaluation.
-        raise NotImplementedError('Evaluation not implemented')
+        raise NotImplementedError('Evaluation not implemented')
\ No newline at end of file

lm_eval/tasks/lambada.py

@@ -67,4 +67,4 @@ class LAMBADA(Dataset):
         return {
             'perplexity': False,
             'accuracy': True
-        }
+        }
\ No newline at end of file

lm_eval/tasks/naturalqs.py

@@ -87,4 +87,4 @@ class NaturalQs(HFTask):
             whether a higher value of the submetric is better
         """
         # TODO: implement evaluation.
-        raise NotImplementedError('Evaluation not implemented')
+        raise NotImplementedError('Evaluation not implemented')
\ No newline at end of file

lm_eval/tasks/openbookqa.py

@@ -95,4 +95,4 @@ class OpenBookQA(HFTask):
             whether a higher value of the submetric is better
         """
         # TODO: implement evaluation.
-        raise NotImplementedError('Evaluation not implemented')
+        raise NotImplementedError('Evaluation not implemented')
\ No newline at end of file

lm_eval/tasks/piqa.py

@@ -74,4 +74,4 @@ class PiQA(Dataset):
     def higher_is_better(self):
         return {
             'acc': True
-        }
+        }
\ No newline at end of file

lm_eval/tasks/quac.py

@@ -103,4 +103,4 @@ class QuAC(Dataset):
             whether a higher value of the submetric is better
         """
         # TODO: implement evaluation.
-        raise NotImplementedError('Evaluation not implemented')
+        raise NotImplementedError('Evaluation not implemented')
\ No newline at end of file

lm_eval/tasks/race.py

@@ -23,7 +23,8 @@ class RACE(HFTask):
         return True
 
     def _collate_data(self, set):
-        if set in self.cache: return self.cache[set]
+        if set in self.cache:
+            return self.cache[set]
         # One big issue with HF's implementation of this dataset: it makes a
         # separate document for each question; meanwhile, in the GPT3 paper it
         # is shown that one document is made per passage.

lm_eval/tasks/squad.py

@@ -83,4 +83,4 @@ class SQuAD(HFTask):
             whether a higher value of the submetric is better
         """
         # TODO: implement evaluation.
-        raise NotImplementedError('Evaluation not implemented')
+        raise NotImplementedError('Evaluation not implemented')
\ No newline at end of file

lm_eval/tasks/storycloze.py

@@ -89,4 +89,4 @@ class StoryCloze(Dataset):
             whether a higher value of the submetric is better
         """
         # TODO: implement evaluation.
-        raise NotImplementedError('Evaluation not implemented')
+        raise NotImplementedError('Evaluation not implemented')
\ No newline at end of file

lm_eval/tasks/triviaqa.py

@@ -74,4 +74,4 @@ class TriviaQA(Dataset):
     def higher_is_better(self):
         return {
             "acc": True
-        }
+        }
\ No newline at end of file

lm_eval/tasks/webqs.py

@@ -70,4 +70,4 @@ class WebQs(HFTask):
             whether a higher value of the submetric is better
         """
         # TODO: implement evaluation.
-        raise NotImplementedError('Evaluation not implemented')
+        raise NotImplementedError('Evaluation not implemented')
\ No newline at end of file

lm_eval/tasks/wikitext.py

@@ -121,4 +121,4 @@ class WikiText2(NLP_TASK):
             whether a higher value of the submetric is better
         """
         # TODO: implement evaluation.
-        raise NotImplementedError('Evaluation not implemented')
+        raise NotImplementedError('Evaluation not implemented')
\ No newline at end of file

lm_eval/tasks/winogrande.py

@@ -90,4 +90,4 @@ class Winogrande(HFTask):
             whether a higher value of the submetric is better
         """
         # TODO: implement evaluation.
-        raise NotImplementedError('Evaluation not implemented')
+        raise NotImplementedError('Evaluation not implemented')
\ No newline at end of file

lm_eval/utils.py

@@ -28,4 +28,4 @@ def simple_parse_args_string(args_string):
 
 def join_iters(iters):
     for iter in iters:
-        yield from iter
+        yield from iter
\ No newline at end of file

lm_eval/utils_stream.py

@@ -5,11 +5,13 @@ from tqdm import tqdm
 import json
 
 
-class ExitCodeError(Exception): pass
+class ExitCodeError(Exception):
+    pass
 
 
 def sh(x):
-    if os.system(x): raise ExitCodeError()
+    if os.system(x):
+        raise ExitCodeError()
 
 def ls(x):
     return [x + '/' + fn for fn in os.listdir(x)]
@@ -64,7 +66,8 @@ class join:
         self.sep = sep
 
     def __rrshift__(self, other):
-        if other is None: return
+        if other is None:
+            return
         try:
             return self.sep.join(other)
         except:
@@ -156,4 +159,4 @@ def comp(*fs):
     return _f
 
 
-X = Reflective()
+X = Reflective()
\ No newline at end of file