Since this is the first time I tried on detectron, I am familiar with this framework. I would like to have 2 questions. Please help me.
After downloading the pre-trained model of RetinaNet baselines R-50-FPN (https://github.com/facebookresearch/Detectron/blob/master/MODEL_ZOO.md), I extracted the weight layer by layer and I got the data like below figures. However, I do not know where these layer in Retina Net. How can I know the architecture of RetinaNet?
In these data, I realized that there are bn_s and bn_b, I guess that they are batch norm. Are they variance and mean?
Dear all,
I hope that everyone stays safe!
Since this is the first time I tried on detectron, I am familiar with this framework. I would like to have 2 questions. Please help me.
After downloading the pre-trained model of RetinaNet baselines R-50-FPN (https://github.com/facebookresearch/Detectron/blob/master/MODEL_ZOO.md), I extracted the weight layer by layer and I got the data like below figures. However, I do not know where these layer in Retina Net. How can I know the architecture of RetinaNet?
In these data, I realized that there are bn_s and bn_b, I guess that they are batch norm. Are they variance and mean?
Thank you for your time and consideration.