search

huawei-noah / Efficient-AI-Backbones

Efficient AI Backbones including GhostNet, TNT and MLP, developed by Huawei Noah's Ark Lab.
4.07k stars 708 forks source link

VIG - Custom dataset #117

Closed awur978 closed 2 years ago

awur978 commented 2 years ago

Hello, thank you for awesome work, really enjoyed your paper. I am currently working on how I can apply ViG to my custom dataset. Can you please provide a small instruction on how your code can be adapted to custom dataset.

For example: in pyramid_vig.py, am I right to assume that the variable num_classes is not generic and only works for imagenet?, if my custom dataset has 100 classes, which other aspect of the code do I need to change and so on?

Finally, the evaluation example you provided,

python train.py /path/to/imagenet/ --model pvig_s_224_gelu -b 256 --pretrain_path /path/to/pretrained/model/ --evaluate

seems to only work for directory with train and val, what if I have a test set separately and I want to check performance of my trained model against the test set, how do I go about this?

Thank you in advance.

iamhankai commented 2 years ago

Thanks for the attention. It's a general question. Please ask Google.