daitao / SAN

Second-order Attention Network for Single Image Super-resolution (CVPR-2019)
554 stars 106 forks source link

Paper error!! (RCAN's parameter is 15.44M, which is less than SAN 15.7M)!!! #48

Open Senwang98 opened 3 years ago

Senwang98 commented 3 years ago

Hi, @daitao Sorry to disturb, I think you have made an error in your SAN paper. RCAN have less parameters than SAN that RCAN contains 15.44M and SAN contains 15.7M. (This value is reported by RCAN's author) Because of your wrong conclusion, now all the work has made the same mistakes as you, which is completely deviated from the development track. image

EchoXu98 commented 1 year ago

求助!请问兄弟们是用几块GPUs训练的啊?我用一块1080ti,12G显存的卡之前一直在MPNCOV.py大约第24行那里报CUDA out of memory,我参考别人给的代码对24行那里做了修改,同时只有将训练时的batch_size修改为1时,在训练时才不会报错。但是在测试时又在MPNCOV.py文件下报了CUDA out of memory错误。

Senwang98 commented 1 year ago

@EchoXu98 我不做这个方向已经很久了,有点记不清楚了,应该是两卡训练的我