SKKU-ESLAB / Auto-Compression

Automatic DNN compression tool with various model compression and neural architecture search techniques
MIT License
20 stars 20 forks source link

Hyperparameter Settings #66

Open MurakamiTestudo opened 9 months ago

MurakamiTestudo commented 9 months ago

For the special hyperparameters like alpha, beta, w_share_in_train, thetas_lr, and train_thetas_from_the_epoch, could you provide some examples or best practices on how to set these values? How do they affect the training and the final model?