Modify the _finetunecli.py file. Now two functions are available: finetune and finetune_advanced.
For the finetune function, users only need to give pretrained_checkpoint_path and finetune_data_dir as inputs, all other parameters are set to be default.
For the finetune_advanced function, users can specify all parameters as they want.
Put two callbacks, i.e. PrintLearningRate and ToleranceCallback, into _optimutils.py and import them in train.py and finetune.py.
Modify the _finetunecli.py file. Now two functions are available: finetune and finetune_advanced. For the finetune function, users only need to give pretrained_checkpoint_path and finetune_data_dir as inputs, all other parameters are set to be default. For the finetune_advanced function, users can specify all parameters as they want.
Put two callbacks, i.e. PrintLearningRate and ToleranceCallback, into _optimutils.py and import them in train.py and finetune.py.