About the multi-scale traing

[suzhenghang]

Hi @soeaver ,
I try to add the multi-scale traing, but the convergence seems to be difficult; Without multi-scale traing, converge quickly. Do you meet this situation? Thanks in advance

[suzhenghang]

During training, some loss will be larger, such as 0.99, 0.73 ..., I try to imshow the preprocessed image and mask, I do not find something wrong

[soeaver]

Hi, you mean multi-scale training for semantic seg?
Usually, ms training will lead to a slightly larger and unstable loss. I think you can look at the final result of the training.

[suzhenghang]

@soeaver , Thanks, ms training does lead to unstable loss.By the way, does ms training increase the IOU in your experiments?

[soeaver]

Actually, I didn't do a single-scale training experiment.
But as many papers say, ms training and random flipping will improve 1-3% mIoU as PASCAL VOC dataset.

[shiyuangogogo]

Hi. @suzhenghang @soeaver . How do you implement multi-scale traing in Caffe?