涂云斌
> I think it is because you are running it on python 2.7 Thank you, you are right, and I sovled this problem when changing version of python to 3+
> 请问您的问题解决了吗，我也遇到了相同的问题。就是用bert自带的optimizer就会OOM，如果用Adam就不会。 请问vcr这个数据集你怎么下的，为什么我点了‘I agree'以后，下面那两个蓝底的Annotations和Images点了没有任何反应
Hi, I want to know how many GPU memories do you use for successfully running this code? @jaeyun95
Now, I use 4 RTX 2080Ti, but it still report CUDA out of memory. I'm look forward to your early reply.
Yes, I have same confusion and I have a little trouble to align the code with paper equation.
@runzeer Have you understood this reasoning module? I try my best to read it, but I cannot align these codes with the Eq.(3) ~ Eq.(9).
@gorou-gnm Hi, can you run this code successfully?
Thank you for your prompt reply! And I have another question that you said the resulting files are about 200GB, but the feats_att.h5 I extracted via your code is 396.0...
My reproduced results are also much better than their reported results in the paper. My conjecture is that the authors finetuned the model after the paper's been accpeted.
Hi，did you train your own model? I have a question that when I evaluate all of the iteration models on the validation set, I will get a summary result as...