Transitions

Base classes

class ignite_framework.transitions.base_transitions.BaseConditionalTransition(name)

condition() → bool

Any condition returning a bool :param engine:

Returns:True or False

false_tranform()

Transformation either setting self.output or any other value in state if self.condition() == False.

Parameters:engine – Returns:self.output is the output state variable therefore no return is required.

get(*name_or_parts)

Get method for parameterized attribute names, especially for state object containers, but not only.

Main reason is to simply avoid unhandy commands, e.g. getattr(state.transitions, ‘{}_{}’.format(name, suffix) commands and simply write state.transitions.get(name, suffix).

get_bso_names()

n_false_transforms

Iteration counter is a state variable, only allowing ‘’int’’ with increasing values by step size ‘+`1`.

n_true_transforms

Iteration counter is a state variable, only allowing ‘’int’’ with increasing values by step size ‘+`1`.

name = 'BaseConditionalTransition'

run()

set(*name_or_parts_and_value)

true_transform()

Transformation either setting self.output or any other value in state if self.condition() == True.

Parameters:engine – Returns:self.output is the output state variable therefore no return is required.

class ignite_framework.transitions.base_transitions.BaseMultiInputsTransition(name, input_refs, caller_refs='')

n_calls

Iteration counter is a state variable, only allowing ‘’int’’ with increasing values by step size ‘+`1`.

name = 'BaseMultiInputsTransition'

output

StateVariable is the basic state variable without any additional type checking or setting conditions.

Features:

• gets/sets/calls a _value and gets/sets/calls a list of callbacks which are callables triggered when _value is set

• overloaded suffixes: [‘_callbacks’, _ref’, ‘_every’, ‘_once’]

  • _ref: returns StateObjectsReference-instance of itself
  • _callbacks: returns list of callbacks (functions called when value is set)

• set-value type dependent __set__() overload:

  • CALLABLE_TYPES & tuple`([<CALLABLE_TYPES>, <dict>]): appended to its `_callbacks-list
  • any other type: sets its _value

• state shortcut auto-generation

class ignite_framework.transitions.base_transitions.BaseTransformer(name)

get(*name_or_parts)

Get method for parameterized attribute names, especially for state object containers, but not only.

Main reason is to simply avoid unhandy commands, e.g. getattr(state.transitions, ‘{}_{}’.format(name, suffix) commands and simply write state.transitions.get(name, suffix).

get_bso_names()

n_transform

Iteration counter is a state variable, only allowing ‘’int’’ with increasing values by step size ‘+`1`.

name = 'BaseTransformer'

run()

set(*name_or_parts_and_value)

transform()

Transformation either setting self.output or any other value in state.

Parameters:engine –

Returns:

class ignite_framework.transitions.base_transitions.BaseTransition(name)

Base class for all transitions.

name = 'BaseTransition'

Transitions

class ignite_framework.transitions.conditional_counters.EveryNumthValue(name, conditional_value_ref, threshold_or_num_calls, caller_refs='', min_or_max='max')

condition()

” Modify set.conditional_set_value based on user conditions. For triggering the callbacks list of the state variable ‘self.n_conditional_counts` the self.conditional_set_value has to be increased by ‘+1`.

false_transform()

get(*name_or_parts)

Get method for parameterized attribute names, especially for state object containers, but not only.

Main reason is to simply avoid unhandy commands, e.g. getattr(state.transitions, ‘{}_{}’.format(name, suffix) commands and simply write state.transitions.get(name, suffix).

get_bso_names()

n_conditional_counts

StateVariable is the basic state variable without any additional type checking or setting conditions.

Features:

• gets/sets/calls a _value and gets/sets/calls a list of callbacks which are callables triggered when _value is set

• overloaded suffixes: [‘_callbacks’, _ref’, ‘_every’, ‘_once’]

  • _ref: returns StateObjectsReference-instance of itself
  • _callbacks: returns list of callbacks (functions called when value is set)

• set-value type dependent __set__() overload:

  • CALLABLE_TYPES & tuple`([<CALLABLE_TYPES>, <dict>]): appended to its `_callbacks-list
  • any other type: sets its _value

• state shortcut auto-generation

name = 'EveryNumthValue'

set(*name_or_parts_and_value)

true_transform()

class ignite_framework.transitions.conditional_counters.OnceNumthValue(name, conditional_value_ref, threshold_or_num_calls, caller_refs='', min_or_max='max')

condition()

” Modify set.conditional_set_value based on user conditions. For triggering the callbacks list of the state variable ‘self.n_conditional_counts` the self.conditional_set_value has to be increased by ‘+1`.

false_transform()

get(*name_or_parts)

Get method for parameterized attribute names, especially for state object containers, but not only.

Main reason is to simply avoid unhandy commands, e.g. getattr(state.transitions, ‘{}_{}’.format(name, suffix) commands and simply write state.transitions.get(name, suffix).

get_bso_names()

n_conditional_counts

StateVariable is the basic state variable without any additional type checking or setting conditions.

Features:

• gets/sets/calls a _value and gets/sets/calls a list of callbacks which are callables triggered when _value is set

• overloaded suffixes: [‘_callbacks’, _ref’, ‘_every’, ‘_once’]

  • _ref: returns StateObjectsReference-instance of itself
  • _callbacks: returns list of callbacks (functions called when value is set)

• set-value type dependent __set__() overload:

  • CALLABLE_TYPES & tuple`([<CALLABLE_TYPES>, <dict>]): appended to its `_callbacks-list
  • any other type: sets its _value

• state shortcut auto-generation

name = 'OnceNumthValue'

set(*name_or_parts_and_value)

true_transform()

class ignite_framework.transitions.early_stopping.EarlyStopping(name, patience, loss_function, output_var_ref, started_var_ref=<ignite_framework.framework.framework.StateObjectsReference object>, completed_var_ref=<ignite_framework.framework.framework.StateObjectsReference object>)

Firing a early stopping event each time the early stopping conditions are reached, but without stopping the training. Therefore the optimum checkpoint time can be verified from a post=training viewpoint. As saving checkpoints of e.g. model, tokenizer, vocabs, configs etc.

condition()

Any condition returning a bool :param engine:

Returns:True or False

current_loss

StateVariable is the basic state variable without any additional type checking or setting conditions.

Features:

• gets/sets/calls a _value and gets/sets/calls a list of callbacks which are callables triggered when _value is set

• overloaded suffixes: [‘_callbacks’, _ref’, ‘_every’, ‘_once’]

  • _ref: returns StateObjectsReference-instance of itself
  • _callbacks: returns list of callbacks (functions called when value is set)

• set-value type dependent __set__() overload:

  • CALLABLE_TYPES & tuple`([<CALLABLE_TYPES>, <dict>]): appended to its `_callbacks-list
  • any other type: sets its _value

• state shortcut auto-generation

minimum_loss

StateVariable is the basic state variable without any additional type checking or setting conditions.

Features:

• gets/sets/calls a _value and gets/sets/calls a list of callbacks which are callables triggered when _value is set

• overloaded suffixes: [‘_callbacks’, _ref’, ‘_every’, ‘_once’]

  • _ref: returns StateObjectsReference-instance of itself
  • _callbacks: returns list of callbacks (functions called when value is set)

• set-value type dependent __set__() overload:

  • CALLABLE_TYPES & tuple`([<CALLABLE_TYPES>, <dict>]): appended to its `_callbacks-list
  • any other type: sets its _value

• state shortcut auto-generation

n_early_stops

Iteration counter is a state variable, only allowing ‘’int’’ with increasing values by step size ‘+`1`.

n_false_transforms

Iteration counter is a state variable, only allowing ‘’int’’ with increasing values by step size ‘+`1`.

n_true_transforms

Iteration counter is a state variable, only allowing ‘’int’’ with increasing values by step size ‘+`1`.

name = 'EarlyStopping'