ICLR_VTNet

ICLR_VTNet copied to clipboard

Is the activation map visualization done through attention rollout? How do you implement it? Thanks for your time!

izilu

可以提供一下文章中已经训练好的模型吗？

11

你好，非常感谢您能分享代码，但是我们在训练结果性能较低，请问可以提供一下您训练好的模型吗？想做进一步的测试，非常感谢。

colinzhaoxp

keeper1314

由于实验中需要修改这部分，在论文中提到了codes and pretrained model will be released for reproducibility

Ubastic

Hello! Why not use multi-GPUs to pretrain your model? unnecessary or have some problems?

Rh-Dang