You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi, I just want to share my experience when I developed KAN that the scale parameter seems quite important (but that was in the very beginning of the KAN project, so I could be hallucinating). Would love to hear your experimental results! Great initiative, would love to see a more efficient KAN implementation (with good features maintained).
The text was updated successfully, but these errors were encountered:
________________________________
From: Blealtan ***@***.***>
Sent: Friday, May 3, 2024 8:53:31 AM
To: Blealtan/efficient-kan ***@***.***>
Cc: Ziming Liu ***@***.***>; Author ***@***.***>
Subject: Re: [Blealtan/efficient-kan] Necessity of the scale parameter? (Issue #1)
Thank you for your information. It's added back now with a controlling option, but I haven't yet done any experiments on that.
—
Reply to this email directly, view it on GitHub<#1 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AFTV5B2YZSM7RC32UZAU54DZANGAXAVCNFSM6AAAAABHEVQEVSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAOJSGU4DIOBXGA>.
You are receiving this because you authored the thread.Message ID: ***@***.***>
Hi, I just want to share my experience when I developed KAN that the scale parameter seems quite important (but that was in the very beginning of the KAN project, so I could be hallucinating). Would love to hear your experimental results! Great initiative, would love to see a more efficient KAN implementation (with good features maintained).
The text was updated successfully, but these errors were encountered: