File size: 1,428 Bytes
256a159 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 |
from typing import List
from datasets import Dataset
from opencompass.openicl import BaseEvaluator
from opencompass.registry import ICL_EVALUATORS, LOAD_DATASET
from ..base import BaseDataset
from .utils import iter_jsonl
@LOAD_DATASET.register_module()
class InfiniteBenchendiaDataset(BaseDataset):
@staticmethod
def load(path: str):
dataset = list(iter_jsonl(path))
raw_data = []
for item in dataset:
context = item['context']
question = item['input']
answer = item['answer']
raw_data.append({
'context': context,
'question': question,
'answer': answer
})
dataset = Dataset.from_list(raw_data)
return dataset
@ICL_EVALUATORS.register_module()
class InfiniteBenchendiaEvaluator(BaseEvaluator):
def score(self, predictions: List, references: List) -> dict:
score = 0.
for i in range(len(predictions)):
prediction = predictions[i]
reference = references[i][0]
for c in ['\n', ':', '"', "'", '.', ',', '?', '!', '{', '}']:
prediction = prediction.replace(c, ' ')
words = prediction.split()
words = [x.upper() for x in words]
if reference in words:
score += 1
score = score / len(predictions) * 100
return {'score': score}
|