SKBL5694

I get the same bug, and I change pytorch 2.0.0 to torch1.13.1+cu117(I guess 1.x.x all works well), it works well.

> Hi, for the VQA task, the text encoder and decoder do not share parameters (their parameters are shared during pre-training). Thanks for your reply. I think you mean the...

> Hi, the primary reason for using LM is because we want to enable image-to-text generation capability. Both losses perform similarly in terms of VL representation learning (MLM can be...

I think the key lies in how to define **open-vocabulary**. The following is my personal understanding of ram after reading code and papers: In ram, we define a large enough...

It seems that I am indeed doing gradcam on the recognition task, because your code does not open the generation task for RAM, but I have added the generation task...

I think I have some problem with image-tag interaction encoder performing backward calculation grad. The calculation method I use is to pre-define a hook, and then register it to the...