Closed XM-WANG closed 2 years ago
Hi, Thanks for pointing this out, It depends on the way that the dataset is preprocessed, the template we put in the paper which is {prem}[P][T]. . . [T]{hyp} is right. Keeping the sample template for label search and the prompt search will help the performance.
Thanks for catching this issue @XM-WANG as well as your kind words!
I've updated the example in the README to be consistent with the template and model used for trigger search.
Hi,
I noticed that the template is '[CLS] {sentence} [T] [T] [T] [P]. [SEP]' in the given command for NLI label search (See readme.md).
Is that correct? I suppose it should be '[CLS] {sentence_A} [P] [T] [T] [T] [T] {sentence_B}. [SEP]'.
And the keys of label_map should be in the capital, i.e entailment -> ENTAILMENT.
By the way, the repo is the clearest one I have ever seen! Many thanks for your share.