Add Fantasy Strategy returning Exact GP #2040
GirishMKulkarni
started this conversation in
General
Replies: 2 comments
-
It is probably possible to convert an exact GP model back to an approximate GP, but this is not implemented at the moment. This would probably require a research effort, and not something that we will be looking to implement in the near future. |
Beta Was this translation helpful? Give feedback.
0 replies
-
Dear Geoff Pleiss,
I contacted you regarding an online version of the get_fantasy_model function for approximate GP. You mentioned that it will require some research efforts. I have been thinking about it and I think I might have a solution for it. Please see a very short description of the solution attached. Kindly let me know if you think this is a right way and can be pursued further. I will be happy to discuss/explore the details with you.
Best regards,
Girish
Dr. Girish Kulkarni
Postdoctoral Researcher
Mathematisches Institut der Heinrich-Heine-Unviersität Düsseldorf
Universitätsstr. 1
Building: 25.22.
Floor/Room: 03.51
D - 40225 Düsseldorf
…On Jun 29 2022, at 7:33 pm, Geoff Pleiss ***@***.***> wrote:
It is probably possible to convert an exact GP model back to an approximate GP, but this is not implemented at the moment. This would probably require a research effort, and not something that we will be looking to implement in the near future.
—
Reply to this email directly, view it on GitHub (#2040 (comment)), or unsubscribe (https://github.com/notifications/unsubscribe-auth/ADGLO27AOPTCECUGAB7EQM3VRSCHJANCNFSM5YXLD6BA).
You are receiving this because you authored the thread.
|
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
I want to add new data sequentially to an already trained approximate GP model. I used the
get_fantasy_model
method in ApproximateGP to add the first new data point, I get the Exact GP model in return. I want to know your thoughts on changing the return type. How can get the Approximate model even after adding new data? I need this because I want to have online training with one data point at a time.Beta Was this translation helpful? Give feedback.
All reactions