search

ohhhyeahhh / SiamCAR

SiamCAR: Siamese Fully Convolutional Classification and Regression for Visual Tracking (CVPR 2020, Oral)
https://openaccess.thecvf.com/content_CVPR_2020/html/Guo_SiamCAR_Siamese_Fully_Convolutional_Classification_and_Regression_for_Visual_Tracking_CVPR_2020_paper.html
322 stars 63 forks source link

About Distributed Training Difference #20

Open chienyiwang opened 4 years ago

chienyiwang commented 4 years ago

Thanks for sharing the codebase. I found that you modified the DistModule part in train.py compared with the original pysot repo, and did not use torch.distributed.launch for multi-gpu training. Do you have specific reason for the modification? Thanks a lot!

twotwo2 commented 4 years ago

There is no difference between the two methods.