hustvl / Vim

[ICML 2024] Vision Mamba: Efficient Visual Representation Learning with Bidirectional State Space Model
Apache License 2.0
2.55k stars 160 forks source link

How to test pretrained model on CIFAR dataset and coco2017 #22

Open DongHan9722 opened 4 months ago

DongHan9722 commented 4 months ago

hi,

Since the pretrained model's head is 1000 classes, how can I test it on the CIFAR dataset which only has 100 classes?

For the coco2017 dataset, could you provide the data preprocessing code about how to construct it properly for evaluation? here is my data folder structure for coco dataset: -dataset -coco -annotations -images -test -train