search

cccorn / AP-loss

The implementation of "Towards accurate one-stage object detection with AP-loss".
MIT License
175 stars 28 forks source link

Converting other backbones #7

Open cancam opened 4 years ago

cancam commented 4 years ago

Hi,

I am planning to use different backbones (resnext-101, resnet-152) different than that you've already provided. As you said in readme, the backbone models are converted from caffe version. Could you please provide the converter tool that you used?

Thank you for your time. Baris.

cccorn commented 4 years ago

Our converter code only works for ResNet-50 and ResNet-101. For ResNet-152 and ResNeXt-101, I suggest using the model provided by Pytorch: https://github.com/pytorch/vision/blob/master/torchvision/models/resnet.py and then change the backbone description accordingly in

lib/model/model.py

and image normalization method in

lib/dataloader/augmentation.py, SubtractMeans()

for training, and in

lib/dataloader/dataloader.py, Normalizer()

for testing.

suresh-s commented 3 years ago

I also like to change to the backbone to resnet152. why changes required in image normalization method (lib/dataloader/augmentation.py, SubtractMeans()) and in lib/dataloader/dataloader.py, Normalizer(). Could you tell me?

and i want to check detection in test images using the trained model, is there any demo code in this repository?