-
Notifications
You must be signed in to change notification settings - Fork 8
Advanced Features
Ravin D edited this page Oct 9, 2024
·
1 revision
You can define your custom loss function and pass it to the model.
Utilize GPU with CuPy for faster training.
Customizable learning rate schedulers for dynamic adjustment of the learning rate during training.
# Define a step decay learning rate scheduler
scheduler = LearningRateScheduler
(lr_type='step_decay', initial_lr=0.01, drop_factor=0.5, step_size=10)