Remove unused imports and format imports

lm_eval/metrics.py

 import math
 from collections import Iterable
-from pprint import pprint
 
 import numpy as np
 import sacrebleu

lm_eval/tasks/anli.py

@@ -3,6 +3,7 @@ from lm_eval.base import rf
 from ..metrics import mean
 from . common import HFTask
 
+
 class ANLIBase(HFTask):
     DATASET_PATH = "anli"
     DATASET_NAME = None

lm_eval/tasks/arc.py

-import numpy as np
 from lm_eval.base import MultipleChoiceTask
-from ..metrics import mean
 from . common import HFTask
 
 

lm_eval/tasks/arithmetic.py

@@ -8,6 +8,7 @@ from best_download import download_file
 
 ArithmeticDoc = namedtuple('ArithmeticDoc', ['context', 'completion'])
 
+
 class Arithmetic(Task):
     directory = 'data/arithmetic/'
 

lm_eval/tasks/common.py

 import datasets
-import numpy as np
-
 import lm_eval.metrics
 from ..base import Task
 

lm_eval/tasks/coqa.py

 import os
 import json
+import transformers.data.metrics.squad_metrics as squad_metrics
 from lm_eval.base import Task, rf, mean
 from ..utils import sh
 from itertools import zip_longest
-import transformers.data.metrics.squad_metrics as squad_metrics
-import collections
-import datasets
-import numpy as np
-from lm_eval.base import rf, mean
-from . common import HFTask
-from tqdm import tqdm
-import string, re
+
 
 class CoQA(Task):
 

lm_eval/tasks/ethics.py

-from lm_eval.base import Task, rf
-from lm_eval.metrics import mean
-from lm_eval.utils import sh
-from .common import yesno
-
 import abc
 import csv
 import os
 import random
 import numpy as np
+from lm_eval.base import Task, rf
+from lm_eval.metrics import mean
+from lm_eval.utils import sh
+from .common import yesno
+
 
 class Ethics(Task):
     def download(self):

lm_eval/tasks/glue.py

 import numpy as np
 from lm_eval.base import rf
 from ..metrics import mean, matthews_corrcoef, f1_score
-from scipy.stats import pearsonr, spearmanr
-from tqdm import auto as tqdm_lib
 from . common import HFTask, yesno
 from ..utils import general_detokenize
 

lm_eval/tasks/lambada.py

+import json
 from lm_eval.base import Task, rf
 from lm_eval.metrics import mean, perplexity
 from lm_eval.utils import sh
-import json
-import math
 from best_download import download_file
 
 

lm_eval/tasks/math.py

+import abc
 import json
-import random
 from lm_eval.utils import sh
 from lm_eval.metrics import mean
 from lm_eval.base import Task, rf
 from pathlib import Path
-import abc
 
 
 class Math(Task):

lm_eval/tasks/mathqa.py

-from . common import HFTask
-from lm_eval.base import mean, rf, MultipleChoiceTask
 import re
+from lm_eval.base import MultipleChoiceTask
+from . common import HFTask
+
 
 class MathQA(HFTask, MultipleChoiceTask):
     DATASET_PATH = "math_qa"

lm_eval/tasks/naturalqs.py

+import random
 from . common import HFTask
 from itertools import islice
-import random
+
 
 class NaturalQs(HFTask):
     # TODO: naturalqs has a *really* large train set that huggingface just

lm_eval/tasks/pubmedqa.py

 import numpy as np
-import json
-import random
-from .common import HFTask 
+from .common import HFTask
 from lm_eval.base import rf
 from ..metrics import mean
 

lm_eval/tasks/qa4mre.py

 import os
-import numpy as np
-from best_download import download_file
-from lm_eval.base import MultipleChoiceTask, rf
-from lm_eval.metrics import mean
 import xml.etree.ElementTree as ET
-import random
+from best_download import download_file
+from lm_eval.base import MultipleChoiceTask
+
 
 class QA4MRE(MultipleChoiceTask):
     YEAR = None

lm_eval/tasks/quac.py

 import json
-import random
 import os
 from lm_eval.base import Task
 from ..utils import sh

lm_eval/tasks/race.py

@@ -5,11 +5,6 @@ from lm_eval.base import rf
 from ..metrics import mean
 from . common import HFTask
 
-import os
-from functools import reduce
-import operator
-from tqdm import tqdm
-import json
 
 class each:
     def __init__(self, f):

lm_eval/tasks/sat.py

-import json
-import random
 import os
-from lm_eval.base import MultipleChoiceTask, rf
-from ..metrics import mean
-from tqdm import auto as tqdm_lib
-from . common import simple_accuracy_metric
-import numpy as np
-from ..utils import sh
+from lm_eval.base import MultipleChoiceTask
 
 
 class SATAnalogies(MultipleChoiceTask):    

lm_eval/tasks/sciq.py

 import os
 import json
-from ..utils import sh
-from lm_eval.base import MultipleChoiceTask, rf
-from ..metrics import mean
 import zipfile
+from lm_eval.base import MultipleChoiceTask
 from best_download import download_file
 
 

lm_eval/tasks/squad.py

-import numpy as np
-from scipy.stats import pearsonr, spearmanr
-from sklearn.metrics import f1_score, matthews_corrcoef
-from tqdm import auto as tqdm_lib
-from . common import HFTask, simple_accuracy_metric, yesno
+from . common import HFTask
+
 
 class SQuAD(HFTask):
     DATASET_PATH = "squad_v2"

lm_eval/tasks/storycloze.py

-import json
-import random
-from lm_eval.base import Task
-from ..utils import sh
 import csv
+from lm_eval.base import Task
+
 
 class StoryCloze(Task):
     NEEDS_MANUAL_DL = True