def cli_main():
pl.seed_everything(1234)
# ------------
# args
# ------------
parser = ArgumentParser()
parser = pl.Trainer.add_argparse_args(parser)
parser = LitClassifier.add_model_specific_args(parser)
parser = MNISTDataModule.add_argparse_args(parser)
args = parser.parse_args()
# ------------
# data
# ------------
dm = MNISTDataModule.from_argparse_args(args)
# ------------
# model
# ------------
model = LitClassifier(args.hidden_dim, args.learning_rate)
# ------------
# training
# ------------
trainer = pl.Trainer.from_argparse_args(args)
trainer.fit(model, datamodule=dm)
# ------------
# testing
# ------------
# todo: without passing model it fails for missing best weights
# MisconfigurationException, 'ckpt_path is "best", but ModelCheckpoint is not configured to save the best model.'
result = trainer.test(model, datamodule=dm)
pprint(result)
def cli_main():
pl.seed_everything(1234)
# ------------
# args
# ------------
parser = ArgumentParser()
parser = pl.Trainer.add_argparse_args(parser)
parser = LitClassifier.add_model_specific_args(parser)
parser = MNISTDataModule.add_argparse_args(parser)
args = parser.parse_args()
# ------------
# data
# ------------
dm = MNISTDataModule.from_argparse_args(args)
# ------------
# model
# ------------
model = LitClassifier(args.hidden_dim, args.learning_rate)
# ------------
# training
# ------------
trainer = pl.Trainer.from_argparse_args(args)
trainer.fit(model, datamodule=dm)
# ------------
# testing
# ------------
result = trainer.test(model, datamodule=dm)
pprint(result)
def main(args: Namespace) -> None:
# ------------------------
# 1 INIT LIGHTNING MODEL
# ------------------------
model = GAN(lr=args.lr, b1=args.b1, b2=args.b2, latent_dim=args.latent_dim)
# ------------------------
# 2 INIT TRAINER
# ------------------------
# If use distributed training PyTorch recommends to use DistributedDataParallel.
# See: https://pytorch.org/docs/stable/nn.html#torch.nn.DataParallel
dm = MNISTDataModule.from_argparse_args(args)
trainer = Trainer.from_argparse_args(args)
# ------------------------
# 3 START TRAINING
# ------------------------
trainer.fit(model, dm)
return acc
def accuracy(self, logits, y):
# currently IPU poptorch doesn't implicit convert bools to tensor
# hence we use an explicit calculation for accuracy here. Once fixed in poptorch
# we can use the accuracy metric.
acc = torch.sum(torch.eq(torch.argmax(logits, -1), y).to(torch.float32)) / len(y)
return acc
def validation_epoch_end(self, outputs) -> None:
# since the training step/validation step and test step are run on the IPU device
# we must log the average loss outside the step functions.
self.log("val_acc", torch.stack(outputs).mean(), prog_bar=True)
def test_epoch_end(self, outputs) -> None:
self.log("test_acc", torch.stack(outputs).mean())
def configure_optimizers(self):
return torch.optim.Adam(self.parameters(), lr=self.hparams.learning_rate)
if __name__ == "__main__":
dm = MNISTDataModule(batch_size=32)
model = LitClassifier()
trainer = pl.Trainer(max_epochs=2, ipus=8)
trainer.fit(model, datamodule=dm)
trainer.test(model, datamodule=dm)
# ------------------------
# 2 INIT TRAINER
# ------------------------
# If use distributed training PyTorch recommends to use DistributedDataParallel.
# See: https://pytorch.org/docs/stable/nn.html#torch.nn.DataParallel
dm = MNISTDataModule.from_argparse_args(args)
trainer = Trainer.from_argparse_args(args)
# ------------------------
# 3 START TRAINING
# ------------------------
trainer.fit(model, dm)
if __name__ == "__main__":
cli_lightning_logo()
parser = ArgumentParser()
# Add program level args, if any.
# ------------------------
# Add LightningDataLoader args
parser = MNISTDataModule.add_argparse_args(parser)
# Add model specific args
parser = GAN.add_argparse_args(parser)
# Add trainer args
parser = Trainer.add_argparse_args(parser)
# Parse all arguments
args = parser.parse_args()
main(args)