"leave-one-out training" and "add 3D noises"

[eternalland]

thank you for your remarkable job, I have a question:
for "leave-one-out training" and "add 3D noises",
Are two methods in a progressive relationship? after one method is completed, then move on to the next method
Or are they a replaceable relationship, choose either one?

thank you

[chensjtu]

Sorry for the late reply! these two methods are indeed in a progressive relationship, i.e., after leave-one-out, we get the mean and variance for 3d noise. I guess you may directly use 3d noise, but the parameter may need manually set.

[eternalland]

Thanks for your reply.☺Sent from my phone-------- Original message --------From: Flank Yang ***@***.***>Date: Mon, Apr 14, 2025, 11:44To: chensjtu/GaussianObject ***@***.***>Cc: eternalland ***@***.***>, Author ***@***.***>Subject: Re: [chensjtu/GaussianObject] "leave-one-out training" and "add 3D noises" (Issue #77) Sorry for the late reply! these two methods are indeed in a progressive relationship, i.e., after leave-one-out, we get the mean and variance for 3d noise. I guess you may directly use 3d noise, but the parameter may need manually set.—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you authored the thread.Message ID: ***@***.***> chensjtu left a comment (chensjtu/GaussianObject#77) Sorry for the late reply! these two methods are indeed in a progressive relationship, i.e., after leave-one-out, we get the mean and variance for 3d noise. I guess you may directly use 3d noise, but the parameter may need manually set. —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you authored the thread.Message ID: ***@***.***>