Shortcuts

RuntimeInfoHook

class mmengine.hooks.RuntimeInfoHook[source]

A hook that updates runtime information into message hub.

E.g. epoch, iter, max_epochs, and max_iters for the training state. Components that cannot access the runner can get runtime information through the message hub.

after_test_epoch(runner, metrics=None)[source]

All subclasses should override this method, if they need any operations after each test epoch.

Parameters
  • runner (Runner) – The runner of the testing process.

  • metrics (Dict[str, float], optional) – Evaluation results of all metrics on test dataset. The keys are the names of the metrics, and the values are corresponding results.

Return type

None

after_train_iter(runner, batch_idx, data_batch=None, outputs=None)[source]

Update log_vars in model outputs every iteration.

Parameters
  • runner (Runner) – The runner of the training process.

  • batch_idx (int) – The index of the current batch in the train loop.

  • data_batch (Sequence[dict], optional) – Data from dataloader. Defaults to None.

  • outputs (dict, optional) – Outputs from model. Defaults to None.

Return type

None

after_val_epoch(runner, metrics=None)[source]

All subclasses should override this method, if they need any operations after each validation epoch.

Parameters
  • runner (Runner) – The runner of the validation process.

  • metrics (Dict[str, float], optional) – Evaluation results of all metrics on validation dataset. The keys are the names of the metrics, and the values are corresponding results.

Return type

None

before_run(runner)[source]

Update metainfo.

Parameters

runner (Runner) – The runner of the training process.

Return type

None

before_train(runner)[source]

Update resumed training state.

Parameters

runner (Runner) – The runner of the training process.

Return type

None

before_train_epoch(runner)[source]

Update current epoch information before every epoch.

Parameters

runner (Runner) – The runner of the training process.

Return type

None

before_train_iter(runner, batch_idx, data_batch=None)[source]

Update current iter and learning rate information before every iteration.

Parameters
  • runner (Runner) – The runner of the training process.

  • batch_idx (int) – The index of the current batch in the train loop.

  • data_batch (Sequence[dict], optional) – Data from dataloader. Defaults to None.

Return type

None

Read the Docs v: latest
Versions
latest
stable
Downloads
pdf
html
epub
On Read the Docs
Project Home
Builds

Free document hosting provided by Read the Docs.