search

miaozhang0525 / iDARTS

codes for ICML2021 paper iDARTS: Differentiable Architecture Search with Stochastic Implicit Gradients
8 stars 2 forks source link

Hi, where is the part related to neumann approximation located in the code? #1

Open pipilurj opened 3 years ago

miaozhang0525 commented 3 years ago

Please check the _hessian_vector_product function in the iDARTS/NAS_Bench201_codes/exps/algos/iDARTS_V2.ipynb, where we give a simple example with 2 terms approximation.

pipilurj commented 3 years ago

Thank you very much!

On Wed, Aug 18, 2021 at 10:55 PM MiaoZhang0525 @.***> wrote:

Please check the _hessian_vector_product function in the iDARTS/NAS_Bench201_codes/exps/algos/iDARTS_V2.ipynb, where we give a simple example with 2 terms approximation.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/MiaoZhang0525/iDARTS/issues/1#issuecomment-901184246, or unsubscribe https://github.com/notifications/unsubscribe-auth/AIS6CUYB53ZQ6IOKWG77LHLT5PCWLANCNFSM5CMDML7Q . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email .

-- Best Regards, Renjie PI

pipilurj commented 3 years ago

I see, did you try to approximate the inverse with more steps? For example, 100? Does the error still converge?