Fix some parameters and train others

[Pedro Porto Buarque de Gusmão]

Hi,

Is there a way to keep some of the networks layers/parameters fixed and do training only on specific parameters?

Thanks

Pedro

[Francisco Vitor Suzano Massa]

For the modules you want to fix the layers, you can override the accGradParameters function with a void function.
For example, say you have a model like
model = nn.Sequential():add(nn.Linear(2,2)):add(nn.Linear(2,2))

you can fix the first layer by doing

model:get(1).accGradParameters = function() end

Now, if you want to have different learning rates per layer, look at the second post in https://gist.github.com/szagoruyko/1e994e713fce4a41773e

[Pedro Porto Buarque de Gusmão]

Thank you!