Di Jin

I see, I need to find the code for NLI and will post it these two days. Thank you for the patience!

Hi, I am so sorry for delay since I have been busy with my conference ddl. I just now spent half an hour digging into my hard drive to find...

Hi, although effective batch size is more important, but if the gradient steps is too large, the performance may be degraded. I used the gradient steps between 4 and 6...

Here it is: https://github.com/jind11/MMM-MCQA/blob/a4edfb035b11858f26114a7b9130e4d7eccfdce9/pytorch_pretrained_bert/modeling.py#L1347 .

Yes!

I am sorry for the late reply. Thanks for reaching out to me! This code base provides the elastic search based IR baseline and you can follow the readme file...

The c here should be the top-K retrieved sentences/paragraphs in the textbooks so that we do not need to concatenate all textbooks.

I just now updated the data in the github repo by adding the 4 options version of data I previously used. Let me whether you can replicate the numbers in...

hmmm, this code was written a year ago and I tried my best to release the old code here without running and verifying. Thanks for helping me find out the...

Hi, I plan to submit one paper to EMNLP and want to incorporate your method in this paper to my modules. So if possible, it would be great that the...