No `training_step()` method defined

Trainer
#1

Hi,
I implemented autoencoder from ‘Lightning in 2 steps’ on Colab. I am getting an error No training_step() method defined though I have the training_step() defined. What is the issue.
Below is my code for the Lightnig Module:

class LitAutoEncoder(pl.LightningDataModule):

def init(self):
super().init()
self.encoder=nn.Sequential(
nn.Linear(2828,64),
nn.ReLU(),
nn.Linear(64,3)
)
self.decoder=nn.Sequential(
nn.Linear(3,64),
nn.ReLU(),
nn.Linear(64,2828)
)

def forward(self,x):
embedding=self.encoder(x)
return embedding

def training_step(self,batch,batch_idx):
x,y=batch
x=x.view(x.size(0),-1)
z=self.encoder(x)
x_hat=self.decoder(z)
loss.F.mse_loss(x_hat,x)
self.log(‘train_loss’,loss)

def configure_optimizers(self):
optimizer=torch.optim.Adam(self.parameters,lr=1e-3)
return optimizer

#2

Can you share the colab notebok?

#3
#4

should be

class LitAutoEncoder(pl.LightningModule):
#6

Hi,
Thanks for pointing it out. I had missed it completely.
Where do you set the number of training epochs in trainer?

#7 
Trainer(max_epochs=epochs)
1 Like
#8

Hi,
I want the train and valid losses foe each epoch updates alongwith the progressbar. The valid_loss seem to update only once after the first epoch and then its value does not change. The Colab notebook is attached.

#9

I tried it and it’s changing after every epoch.

#10

Ya, It is changing. It was not changing for the lr I was using earlier. Presently trying with lr scheduling to see if it improves things.