callbacks

import tensorwaves.optimizer.callbacks

Collection of loggers that can be inserted into an optimizer as callback.

class Loadable

Bases: ABC

abstract static load_latest_parameters(filename: Union[Path, str]) → dict

class Callback

Bases: ABC

Interface for callbacks such as CSVSummary.

See also

Callbacks

abstract on_optimize_start(logs: Optional[Dict[str, Any]] = None) → None

abstract on_optimize_end(logs: Optional[Dict[str, Any]] = None) → None

abstract on_iteration_end(iteration: int, logs: Optional[Dict[str, Any]] = None) → None

abstract on_function_call_end(function_call: int, logs: Optional[Dict[str, Any]] = None) → None

class CallbackList(callbacks: Iterable[Callback])

Bases: Callback

Class for combining Callback s.

Combine different Callback classes in to a chain as follows:

>>> from tensorwaves.optimizer import Minuit2
>>> optimizer = Minuit2(
...     callback=CallbackList([TFSummary(), YAMLSummary("fit_result.yml")])
... )

property callbacks: List[Callback]

on_optimize_start(logs: Optional[Dict[str, Any]] = None) → None

on_optimize_end(logs: Optional[Dict[str, Any]] = None) → None

on_iteration_end(iteration: int, logs: Optional[Dict[str, Any]] = None) → None

on_function_call_end(function_call: int, logs: Optional[Dict[str, Any]] = None) → None

class CSVSummary(filename: Union[Path, str], function_call_step_size: int = 1, iteration_step_size: Optional[int] = None)

Bases: Callback, Loadable

Log fit parameters and the estimator value to a CSV file.

on_optimize_start(logs: Optional[Dict[str, Any]] = None) → None

on_optimize_end(logs: Optional[Dict[str, Any]] = None) → None

on_iteration_end(iteration: int, logs: Optional[Dict[str, Any]] = None) → None

on_function_call_end(function_call: int, logs: Optional[Dict[str, Any]] = None) → None

static load_latest_parameters(filename: Union[Path, str]) → dict

class TFSummary(logdir: str = 'logs', step_size: int = 10, subdir: Optional[str] = None)

Bases: Callback

Log fit parameters and the estimator value to a tf.summary.

The logs can be viewed with TensorBoard via:

tensorboard --logdir logs

on_optimize_start(logs: Optional[Dict[str, Any]] = None) → None

on_optimize_end(logs: Optional[Dict[str, Any]] = None) → None

on_iteration_end(iteration: int, logs: Optional[Dict[str, Any]] = None) → None

on_function_call_end(function_call: int, logs: Optional[Dict[str, Any]] = None) → None

class YAMLSummary(filename: Union[Path, str], step_size: int = 10)

Bases: Callback, Loadable

Log fit parameters and the estimator value to a tf.summary.

The logs can be viewed with TensorBoard via:

tensorboard --logdir logs

on_optimize_start(logs: Optional[Dict[str, Any]] = None) → None

on_optimize_end(logs: Optional[Dict[str, Any]] = None) → None

on_iteration_end(iteration: int, logs: Optional[Dict[str, Any]] = None) → None

on_function_call_end(function_call: int, logs: Optional[Dict[str, Any]] = None) → None

static load_latest_parameters(filename: Union[Path, str]) → dict