search

vturrisi / solo-learn

solo-learn: a library of self-supervised methods for visual representation learning powered by Pytorch Lightning
MIT License
1.41k stars 182 forks source link

How to evaluate the ResNet18 backbone on object detection? #266

Closed yutaro-s closed 2 years ago

yutaro-s commented 2 years ago

Hi. Thank you for your work! This is a really cool project!

I'm trying to evaluate a trained model whose backbone is ResNet18 on the downstream task. The program (./downstream/object_detection/train_object_detection.py) assumes that ResNet50 will be used for the backbone. Could you teach me how to evaluate the ResNet18 backbone on object detection?

vturrisi commented 2 years ago

Hey. The code that we have assumes that you want to evaluate a resnet50 and we don't support a resnet18. To chance this, you will need to re-write the model conversion so that it works with a resnet18 and also the detectron code.

yutaro-s commented 2 years ago

Thanks for your quick reply. I got it.