Migrate workflow=[('train', 1), ('val', 1)] option to mmengine

[EvgeniyBerdyugin]

Thank you very much to the developers for your work. I looked through a lot of documentation, but did not find any analogue of the workflow=[('train', 1), ('val', 1)] option on the latest version of mmengine. I would like that during training, losses would be shown on both the training and test datasets. I put this option in the config or runner, but nothing changes.

[zhouzaida]

Hi, you can configure it in mmengine via train_cfg parameter in Runner (https://mmengine.readthedocs.io/en/latest/get_started/15_minutes.html#build-a-runner-and-run-the-task):

from torch.optim import SGD
from mmengine.runner import Runner

runner = Runner(
    # the model used for training and validation.
    # Needs to meet specific interface requirements
    model=MMResNet50(),
    # working directory which saves training logs and weight files
    work_dir='./work_dir',
    # train dataloader needs to meet the PyTorch data loader protocol
    train_dataloader=train_dataloader,
    # optimize wrapper for optimization with additional features like
    # AMP, gradtient accumulation, etc
    optim_wrapper=dict(optimizer=dict(type=SGD, lr=0.001, momentum=0.9)),
    # trainging coinfs for specifying training epoches, verification intervals, etc
    train_cfg=dict(by_epoch=True, max_epochs=5, val_interval=1),
    # validation dataloaer also needs to meet the PyTorch data loader protocol
    val_dataloader=val_dataloader,
    # validation configs for specifying additional parameters required for validation
    val_cfg=dict(),
    # validation evaluator. The default one is used here
    val_evaluator=dict(type=Accuracy),
)

runner.train()