luyug / mores_plus Goto Github PK

Thank you for publishing the code. However, it seems to be incomplete. For example, there is no code/guidelines regarding encoding queries & documents, no hyperparameters (such as chunk size) are provided. I don't think it's possible to use this code without substantial rework and reimplementation.

Do you plan to publish a more complete version? Thanks!

@luyug Thanks for providing this code. I have an implementation based on this repo and the paper. However, FT the model on msmarco datasets, the model performance was very low (MRR of 0.24). Do you want to provide more information on the hyperparameters you used during the training?

Also could you provide additional information on the chunking implementation?