def pre_train(self, x, y, epochs, batch_size=4, log_base_dir='./logs'):
print(
f"Pre Training for {epochs} epochs bs: {batch_size} storing logs: {log_base_dir}"
)
from framework import train
history, trained_model = train(model=self.model,
x=x,
y=y,
epochs=epochs,
batch_size=batch_size,
log_base_dir=log_base_dir)
self.model = trained_model
loss_fn, [],
logger=logger,
callbacks=[prec_rec])
elif args.action == 'train':
scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau(optim,
'min',
patience=2,
verbose=True,
cooldown=1)
for epoch in range(start_epoch, args.epochs):
train_loss = framework.train(network,
train_loader,
loss_fn,
optim,
epoch,
writer=writer,
early_stop=args.early_stop)
prec_rec = PrecRec(n_thresholds=100)
framework.test(
network,
val_loader,
loss_fn,
[prec_rec],
epoch,
writer=writer,
early_stop=args.early_stop,
)
val_loader,
criteria,
logger=logger,
callbacks=[prec_rec])
f1, thres = prec_rec.best_f1()
print('F1', f1, 'at', thres)
elif args.action == 'train':
scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau(
optim, 'min', patience=2, verbose=True, cooldown=1)
for epoch in range(start_epoch, args.epochs):
train_loss = framework.train(network,
train_loader,
loss_fn,
optim,
epoch,
early_stop=args.early_stop,
logger=logger)
score = framework.test(network,
val_loader,
criteria,
early_stop=args.early_stop,
logger=logger)
scheduler.step(train_loss)
framework.save_checkpoint(epoch,
score,
network,
optim,
path=args.artifacts)