| 37 | return sampled_df |
| 38 | |
| 39 | def load_fever(self, cache_dir="data/fever", sample_size=100, type="test"): |
| 40 | assert type in ["train", "validation", "test"] |
| 41 | data = datasets.load_dataset('copenlu/fever_gold_evidence', cache_dir=cache_dir) |
| 42 | df = data[type].to_pandas() |
| 43 | sampled_df = df.sample(sample_size, random_state=self.seed)[["claim", "label"]].reset_index(drop=True) |
| 44 | return sampled_df |
| 45 | |
| 46 | def load_trivia_qa(self, cache_dir="data/trivia_qa", sample_size=100, type="test"): |
| 47 | assert type in ["train", "validation", "test"] |