callbacks

_callbacks

This file is a base class for implementation of Callbacks.

class scalr.nn.callbacks._callbacks.CallbackBase(dirpath='.')

Bases: object

Base class to build callbacks.

classmethod get_default_params()

Class method to get default params for callbacks config.

class scalr.nn.callbacks._callbacks.CallbackExecutor(dirpath: str, callbacks: list[dict])

Bases: object

Wrapper class to execute all enabled callbacks.

Enabled callbacks are executed with the early stopping callback executed last to return a flag for continuation or stopping of model training

execute(**kwargs) → bool

Execute all the enabled callbacks. Returns early stopping condition.

early_stopping

This file is an implementation of early stopping callback.

class scalr.nn.callbacks.early_stopping.EarlyStopping(dirpath: str = None, patience: int = 3, min_delta: float = 0.0001)

Bases: CallbackBase

Implements early stopping based upon validation loss.

patience

Number of epochs with no improvement after which training will be stopped.

min_delta

Minimum change in the monitored quantity to qualify as an improvement,

i.e. an absolute change of less than min_delta, will count as no improvement.

classmethod get_default_params()

Class method to get default params for model_config.

model_checkpoint

This file is an implementation of model checkpoint callback.

class scalr.nn.callbacks.model_checkpoint.ModelCheckpoint(dirpath: str, interval: int = 5)

Bases: CallbackBase

Model checkpointing to save model weights at regular intervals.

epoch

An interger count of epochs trained.

max_validation_acc

Keeps track of the maximum validation accuracy across all epochs.

interval

Regular interval of model checkpointing.

classmethod get_default_params()

Class method to get default params for model_config.

save_checkpoint(model_state_dict: dict, opt_state_dict: dict, path: str)

A function to save model & optimizer state dict to the given path.

Parameters:

• model_state_dict – Model’s state dict.

• opt_state_dict – Optimizer’s state dict.

• path – Path to store checkpoint to.

tensorboard_logger

This file is an implementation of Tensorboard logging callback.

class scalr.nn.callbacks.tensorboard_logger.TensorboardLogger(dirpath: str = '.')

Bases: CallbackBase

Tensorboard logging of the training process.

epoch

An interger count of epochs trained.

writer

Object that writes to tensorboard.

classmethod get_default_params()

Class method to get default params for model_config.

test_early_stopping

This is a test file for early_stopping.py

scalr.nn.callbacks.test_early_stopping.test_early_stopping()

This function tests early stopping of the model.