feat: use SWA when finetuning

This commit is contained in:
Hervé BREDIN committed 2023-06-14 22:33:41 +02:00
1 parent 4eb719046b
commit 88b14f6ce1
1 file changed
+7 -1
+7 -1
View File
@@ -36,6 +36,7 @@ from pytorch_lightning.callbacks import (
LearningRateMonitor,
ModelCheckpoint,
RichProgressBar,
StochasticWeightAveraging,
)
from pytorch_lightning.loggers import TensorBoardLogger
from torch_audiomentations.utils.config import from_dict as get_augmentation
@@ -110,7 +111,12 @@ def train(cfg: DictConfig) -> Optional[float]:
# TODO: for fine-tuning and/or transfer learning, we start by fitting
# TODO: task-dependent layers and gradully unfreeze more layers
# TODO: callbacks.append(GraduallyUnfreeze(epochs_per_stage=1))
pass
swa = StochasticWeightAveraging(
1e-2,
swa_epoch_start=2,
annealing_epochs=10,
)
callbacks.append(swa)
checkpoint = ModelCheckpoint(
monitor=monitor,