Token Labeling

[monney]

Hi, thank you for the paper.

Do you have any numbers for how the networks perform without token labeling, only using MixToken and other augmentations?

[yuanli2333]

Hi, token labeling can improve volo_d1 by 0.4% (from 83.8% to 84.2%). For larger volo models, token labeling can achieve improvement from 0.0 to 0.4%.

[monney]

Thank you, I think these numbers are probably worth adding to the paper as well for context.

[KleinXin]

Hi, token labeling can improve volo_d1 by 0.4% (from 83.8% to 84.2%). For larger volo models, token labeling can achieve improvement from 0.0 to 0.4%.

Could you please also put a copy of token labeling data and pretrained models on Baidu Netdisk？

[Ree1s]

Hi,thank you for sharing this work. If I want to use VOLO to finetune on my own dataset, how can I create my own token labeling?

[zihangJiang]

Hi @Ree1s , You can refer to https://github.com/zihangJiang/TokenLabeling/issues/7#issuecomment-863130757 for generating token label data for your own dataset.

[yuanli2333]

Hi,thank you for sharing this work. If I want to use VOLO to finetune on my own dataset, how can I create my own token labeling?

If you only want to 'finetune' VOLO on your own dataset, one option is to directly use the 'class token' or mean the feature tokens without using token labeling, thus you don't need to generate your own token labels.

[Ree1s]

Hi @Ree1s , You can refer to zihangJiang/TokenLabeling#7 (comment) for generating token label data for your own dataset.

Thanks, and I will try on my dataset through this way mentioned.

[Ree1s]

Hi,thank you for sharing this work. If I want to use VOLO to finetune on my own dataset, how can I create my own token labeling?

If you only want to 'finetune' VOLO on your own dataset, one option is to directly use the 'class token' or mean the feature tokens without using token labeling, thus you don't need to generate your own token labels.

I'll have a try. Thanks a lot!

[chengengliu]

Hi,thank you for sharing this work. If I want to use VOLO to finetune on my own dataset, how can I create my own token labeling?

If you only want to 'finetune' VOLO on your own dataset, one option is to directly use the 'class token' or mean the feature tokens without using token labeling, thus you don't need to generate your own token labels.

Hi, when you say class token, do you mean configure return_mean to True the feeding parameters of VOLO?