FadedCosine
FadedCosine
@xutaima so have you updated the code till now?
This may seem like a faiss related error, which is often a headache issue🤕️. You can search this error info on Google and there may be corresponding solutions.
I wish you good luck🍀
These baselines all come with source code links in their original paper, and you should be able to easily replicate the results from their open source code.
> However, the results i get are quite different from the reported results in the paper. > > Medical Law Koran > reproduced results 19.06 11.09 12.41 > reported results...
Sorry, as I have already graduated, I cannot obtain my experimental materials at the time. Can you directly use the checkpoint we provided to continue your experiment?
I am sorry for the late reply (because I will graduate soon, and I have been busy looking for a job). We provid the processed data and the checkpoints in...
> Also, should this be revised to `transformer_iwslt_de_en_with_datastore` ? > > https://github.com/FadedCosine/kNN-KD/blob/72fb20cdcb31f4018ae9e3f00002fe2a542fbc80/commands/vanilla_knn_mt_inference.sh#L18 Yes, as we have stated in our paper, for iwslt datasets, we use transformer_iwslt_de_en configuration.
> how do you decide the value of knntarget_size? @Hannibal046 @FadedCosine ``knntarget_size'' is the precessed token numbers. You can get this number by running ``fairseq-preprocess xxx''
I am sorry, but I have no idea about this error 😢