File size: 732 Bytes
256a159 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 |
from ..utils.function_utils import compute_rouge
#法条记忆问答
def compute_ftcs(data_dict):
"""
Compute the ROUGE-L score between the prediction and the reference
"""
references, predictions = [], []
for example in data_dict:
question, prediction, answer = example["origin_prompt"], example["prediction"], example["refr"]
answer = answer.replace("答案:", "")
predictions.append(prediction)
references.append(answer)
# compute the accuracy of score_list
rouge_scores = compute_rouge(predictions, references)
rouge_ls = [score["rouge-l"]["f"] for score in rouge_scores]
average_rouge_l = sum(rouge_ls) / len(rouge_ls)
return {"score": average_rouge_l}
|