lambada_cloze.py 561 Bytes
Newer Older
1
2
3
4
import json
from lm_eval.base import Task, rf
from lm_eval.metrics import mean, perplexity
from lm_eval.utils import sh
Leo Gao's avatar
Leo Gao committed
5
from lm_eval.tasks.lambada import LAMBADA
6
7
8
from best_download import download_file


Leo Gao's avatar
Leo Gao committed
9
class LAMBADA_cloze(LAMBADA):
Leo Gao's avatar
Leo Gao committed
10
    VERSION = 0
11
12
13
    def doc_to_text(self, doc):
        return doc['text'].rsplit(' ', 1)[0] + " ____. ->"

14
15
16
17
18
19
    def should_decontaminate(self):
        return True

    def doc_to_decontamination_query(self, doc):
        return doc['text']

20
21
    def doc_to_target(self, doc):
        return " " + doc['text'].rsplit(' ', 1)[1]