Questions about VTAB-1K CIFAR100

dongzelian / SSF

[NeurIPS'22] This is an official implementation for "Scaling & Shifting Your Features: A New Baseline for Efficient Model Tuning".

https://arxiv.org/pdf/2210.08823.pdf

MIT License

163 stars 12 forks source link

Questions about VTAB-1K CIFAR100 #20

Open WangYZ1608 opened 9 months ago

WangYZ1608 commented 9 months ago

Hello, congratulations! Thank you for released VTAB-1K, it help me a lot. I tried to experiment on CIFAR100, but found that the results of SSF and VPT were far worse than those in the paper. I suspect there is a problem with the labels of the released CIFAR-100 dataset. So I wanted to ask you to check it out. grateful.

2catycm commented 1 month ago

I got a 67.47 acc1

2catycm commented 1 month ago

what about you?