image resize

[ShiAngWang]

Hi， while training， have you resized the images into a specific size or used the original image without transformation? If you resized the images, can you show where the code is?

[yun-liu]

We just used the original image size without resizing. To do this, we used a batch size of 1 (code) and set the iter_size in Caffe to 5 (code). In this way, the gradients during training are averaged among 5 images, similar to the effect of the batch size of 5 (we don't use batchnorm, so iter_size=5 euqals to batch size=5 here).