scheduler

Module containing custom learning rate schedulers for use in minerva.

class CosineLR(optimizer, min_lr: int, max_lr: int, max_epochs: int, n_periods: int = 1, last_epoch: int = -1, verbose='deprecated')

Cosine learning rate scheduler.

Parameters:

• optimizer (Optimizer) – Wrapped optimizer.

• min_lr (int) – Minimum learning rate.

• max_lr (int) – Maximum learning rate.

• max_epochs (int) – Epoch number to run learning rate cosine oscilation up to.

• n_periods (int) – Optional; Number of periods of the cosine oscilation over the scheduler. Default 1.

• last_epoch (int) – The index of last epoch. Default: -1.

• verbose (bool) –

  If True, prints a message to stdout for each update. Default: False.

  Deprecated since version 2.2: verbose is deprecated. Please use get_last_lr() to access the learning rate.

get_lr() → list[float]

Compute learning rate using chainable form of the scheduler.