How do I remove a pooling layer and it's subsequent layers from a pretrained network?

[VHK459]

I have defined a custom model, which has a graph mean pooling layer. I would like to remove it along with its subsequent layers. Is there a way to do this in pytorch geometric?

[rusty1s]

You could use model hooks to get intermediate outputs before the pooling layer. Alternatively, you can use torch.fx to modify the model on the fly.