search

yxgeee / MMT

[ICLR-2020] Mutual Mean-Teaching: Pseudo Label Refinery for Unsupervised Domain Adaptation on Person Re-identification.
https://yxgeee.github.io/projects/mmt
MIT License
469 stars 73 forks source link

Some questions #8

Closed WangWenhao0716 closed 4 years ago

WangWenhao0716 commented 4 years ago

If I only have 2 GPUs, for example, 2 V100 32G. Can I reproduce your experiments by enlarging batch size?

yxgeee commented 4 years ago

I am not sure a larger batch size would contribute to better performance, so I encourage you to try a batch size of 64 or 32 first.

WangWenhao0716 commented 4 years ago

OK! Thanks!