bert_score
bert_score copied to clipboard
Tiiiger
BERT score for text generation
During the IDF dict calculation, the weight associated with special tokens is zeroed: https://github.com/Tiiiger/bert_score/blob/dbcf6db37e8bd6ff68446f06b0ba5d0763b62d20/bert_score/score.py#L243-L246 But, to my understanding of the code, this weight never actually prevents a non-special token embedding...
Hello, Do you happen to still have the COCO human ratings, just to see the setup. Was M1 and M2 binary, or was it a more fine-grained scale? What is...
When I BERTScorer as follows: ``` bert_scorer = BERTScorer(lang='en-sci') _, _, f1 = bert_scorer.score([pred_text], [r_text]) ``` I get the following error: `RuntimeError: The expanded size of the tensor (556) must...
I am using a very basic call around a simple response of "true" and "false" variations, and am getting a very odd result that I don't quite understand. Here's the...
Does bert-score support DeBERTaV3? https://huggingface.co/microsoft/deberta-v3-large If not, what types of work would this require?
I'm trying to use bertscore to retrive top N similar docs, is there a a way to cache the embeddings for N docs ?
Small change to improve warning messages printed instead of Warnings and output to stderr using `sys`.
Warning
Hi~ How can i hide the warning as follows: Warning: Empty candidate sentence detected; setting raw BERTscores to 0.
I run this line: P, R, F1 = bert_score.score(cands, refs, lang='en', verbose=True, model_type='roberta-large', num_layers=17) and I got nan of P score, but not for R and F1 score. what does...
I think it would be good practice to separate main functionality dependencies and other stuff into separate requirements files. For example `matplotlib` should not be needed for calculating the scores....
