Update Adam learning rate after 10 epochs

[cmdrootaccess]

hello,

how can i update the learning rate of adam optimizer after 10 epochs ?
my code is like this

self.lr_decay_epoch = [15,]
if epoch in self.lr_decay_epoch:
     self.lr = self.lr * 0.1
     self.optimizer = Adam(filter(lambda p: p.requires_grad, self.net.parameters()), lr=self.lr, weight_decay=self.wd)

[rohitgr7]

you can use LambdaLR where lambda function can be something like:

lambda epoch: return lr * (0.1 if epoch in self.lr_decay_epoch else 1)

[cmdrootaccess]

can you please show me that in pytorch lightning ?

[rohitgr7]

def configure_optimizers(self):
    opt = # your optimizer
    sched = LambdaLR(opt, lr_lambda=lambda epoch: return lr * (0.1 if epoch in self.lr_decay_epoch else 1))
    return {'optimizer': opt, 'lr_scheduler': sched}

[cmdrootaccess]

thank you so much 😊