How to handle pretrained models without training them

[staniPetrox]

Hey gang!

I have written an Encoder model and a decoder model and I want to train them separately.

class Decoder(pl.LightningModule):
    def __init__(self, encoder_model):#visualize_latent):
        super().__init__()
        self.encoder_model = encoder_model

However, when I give my Decoder an Encoder hyperparameter, how do I make sure it will not be trained?
I actually asked that question before, but it wasn't answered for a while and I do not find sufficient support in the docs.

Thanks in advance!

[staniPetrox]

Ok, I found out from other forums that one should use .freeze(), in this case: self.encoder_model.freeze()