idstcv / ZenNAS

219 stars 35 forks source link

A training question about GENet #29

Closed FisherWY closed 1 year ago

FisherWY commented 1 year ago

Hi, everyone. I'm currently trying to reproduce your previous work "Neural Architecture Design for GPU-Efficient Networks", here is the repo link. After determining the model structure, such as "GENet_large", "GENet_small", etc. Can we refer to the training script(train_image_classification.py) in this repo to train the model and get the effect consistent with the paper description?

MingLin-home commented 1 year ago

Hi FisherWY,

Sorry for the late reply. We no longer maintain GENet repo. Yes you can use Zen-NAS training script. Actually you can use the more popular TIMM framework to train our model. Teacher-Student Distillation is strongly recommended for better accuracy. The TS training script is here:

https://github.com/idstcv/ZenNAS/blob/main/ts_train_image_classification.py

FisherWY commented 1 year ago

Thank you for your guidance, I will try it according to your suggestion. :-)