Additional data to train approximate gp model

[GirishMKulkarni]

Hi,

I would like to add new data to the already trained approximate GP model. Is there a way in gpytorch to update the hyperparameters without recalculating again incorporating the additional data? For Exact GP models, there are get_fantasy_data() and set_train_data() methods.
But how can I achieve this for Approximate GP?
(A similar issue for Exact GP was answered here #645)

Thanks in advance!

[wjmaddox]

Currently, this is somewhat impossible without re-training but this exact question is what #1874 will do for you once we merge it in. That PR is basically complete at this stage.

[GirishMKulkarni]

Dear Wesley, Thank you so much for the update. Looking forward to the adittion. Best regards, Girish

…

On Apr 24 2022, at 4:56 pm, Wesley Maddox ***@***.***> wrote: Currently, this is somewhat impossible without re-training but this exact question is what #1874 (#1874) will do for you once we merge it in. That PR is basically complete at this stage. — Reply to this email directly, view it on GitHub (#1993 (comment)), or unsubscribe (https://github.com/notifications/unsubscribe-auth/ADGLO22KOUB7R6423DEJNKLVGVOLVANCNFSM5UGKAHTA). You are receiving this because you authored the thread.