search

yaoing / DAN

Official implementation of DAN
MIT License
163 stars 35 forks source link

About pre-trained model of MSCeleb #13

Closed RockPine520 closed 2 years ago

RockPine520 commented 2 years ago

Thanks for your work which inspire me a lot! There are a few questions I would like to ask.

Could you tell me whether the MSCeleb-resnet18 pre-training model weights are trained by your own? When I used the MSCeleb pre-training weights directly on resnet18 and trained the raf-db dataset according to your training strategy, I was able to achieve 88.75% accuracy, which I think is quite high. But in your paper, the accuracy of baseline resnet18 is just 86.25%, which confusing me a lot.

I‘m looking forward to your reply. Thank you!

yaoing commented 2 years ago

Thanks for your interest! This MSCeleb-resnet18 pretrained model is not trained by myself. In fact, it's also used in many other FER works like SCN and DACL. However, the best performances of all these model are below your stated 88.75%.

If you're still unsure, try using their pre-trained MSCeleb-resnet18 model for training and use it as a reference.

RockPine520 commented 2 years ago

Thanks for your advice. I'll try it.