Open jingsongs opened 5 months ago
Hello, is MoverScore only support for Bert model? When I use the mdeberta-v3-base model, I got this error:
The below is my code: os.environ['MOVERSCORE_MODEL'] = "mdeberta-v3-base"
from moverscore_v2 import get_idf_dict, word_mover_score from collections import defaultdict
refs = ['English is a language', 'English'] sys = ['Language English', 'Chinese']
idf_dict_hyp = get_idf_dict(sys) # idf_dict_hyp = defaultdict(lambda: 1.) idf_dict_ref = get_idf_dict(refs) # idf_dict_ref = defaultdict(lambda: 1.) scores = word_mover_score(refs, sys, idf_dict_ref, idf_dict_hyp, stop_words=[], n_gram=2, remove_subwords=True) print(scores)
Thanks.
Hello, is MoverScore only support for Bert model? When I use the mdeberta-v3-base model, I got this error:![捕获](https://github.com/AIPHES/emnlp19-moverscore/assets/25786748/4f5b18fd-7357-43c9-9764-2b5626e331fe)
The below is my code: os.environ['MOVERSCORE_MODEL'] = "mdeberta-v3-base"
from moverscore_v2 import get_idf_dict, word_mover_score from collections import defaultdict
refs = ['English is a language', 'English'] sys = ['Language English', 'Chinese']
idf_dict_hyp = get_idf_dict(sys) # idf_dict_hyp = defaultdict(lambda: 1.) idf_dict_ref = get_idf_dict(refs) # idf_dict_ref = defaultdict(lambda: 1.) scores = word_mover_score(refs, sys, idf_dict_ref, idf_dict_hyp, stop_words=[], n_gram=2, remove_subwords=True) print(scores)
Thanks.