KimSoybean
@foralliance Hi！I think we should speak English because your questions may help the people in other countries.
@foralliance Hi！I think we should speak English because your questions may help the people in other countries.
-
No. When you train the model with large input image size (e.g., 800x1300), the batch-size will reduce to 1-2 due to the limited GPU memory. Then the effect of BN will be constrained. If so, please replace BN with GN. GN does not care batch-size.
-
This question is very complex. I can just answer your question that they are not relevant.
Originally posted by @KimSoybean in https://github.com/KimSoybean/ScratchDet/issues/3#issuecomment-475104596
@KimSoybean Hi I read the answer above,you mean one GPU can not reach the large batch_size, I think 128 means accum_batch_size, we can use one GPU read 4 batch_size by 32 iter_size. So the batch_size of one GPU(such as 4) will Influence the effect of BN? Or why you recommend GN.
Or B of BN means batch_size that one gpu can reach.Looking forward to your answer。
BN is evaluated by the statistics in a batch, so batchsize will impact the perfoemance. I have tried GN on mmdetection on ScratchDet, and get 1mAP lower result. I haven't found the reason.
