Train/Test split when training linear classifier (ImageNet)

Question

Train/Test split when training linear classifier (ImageNet)

jnyjxn opened this issue 4 years ago · comments

Hello,

Thank you for this great work and well-written paper.

When training the linear classifier (e.g. Alexnet with ImageNet), how do you perform the train/test split? 1. Do you use all of the classes in your test set?
2. How do you choose views for each sample?

Many thanks,
Jonny

Yonglong Tian · Answer 1 · Tue May 26 2020 03:05:16 GMT+0800 (China Standard Time)

Hi Jonny,

ImageNet dataset itself has a "train" set and a "val" set, which will be used as train and test sets, respectively.
The same as the CMC pre-training stage. If you train CMC with Y/DbDr split, then you use Y/DbDr split in linear evaluation stage as well.

Jonny Jackson · Answer 2 · Mon Jun 01 2020 18:21:55 GMT+0800 (China Standard Time)

Thanks @HobbitLong - unfortunately from what I can tell ImageNet is defunct (I've had no response for > 3 months) so I can't access the dataset.

Could you briefly describe the approach they've used to do the split? Specifically, do they split according to e.g. 80 different chairs in train, 20 different chairs in val, or is it the same 100 chairs in train and val but with different perspectives?

Thank you!

Yonglong Tian · Answer 3 · Mon Jun 01 2020 23:02:14 GMT+0800 (China Standard Time)

I believe it's the former way.