Hi,
I wish to investigate various methods which can assist with transfer learning. Surprisingly, I cannot see any de facto standard datasets for assessing transfer learning performance, like we have with ImageNet for training image classifiers from scratch.
My hope was to use networks pre-trained on ImageNet.
I’ve found a few interesting datasets - Stanford dogs seems an excellent candidate because obviously dogs are well represented in ImageNet, but I was thinking of something more general?
Many thanks