I am working on a meta-learning project and experimenting with MAML. I am able to get a first-order approximation fine, as it can be implemented without differentiating through gradient descent. When doing “full” MAML, I am running into an error on GPU. Below is a MWE #device = Flux.gpu # Returns a…

Taking the derivative of a scalar loss, that involves a gradient inside, errors on GPU only

Abhinav_Bhatia July 13, 2022, 5:11am 8

1 Like

Topic		Replies	Views
Compilation error in Zygote, Flux and CUDA interaction Machine Learning question , cuda , error , flux , zygote	3	404	January 6, 2023
How to use gradient of neural network as the loss function? Machine Learning question	13	2795	March 23, 2021
Zygote pullback vs gradient New to Julia flux , zygote	2	2197	December 25, 2020
Zygote error in backprop through NN Machine Learning question , error , flux , zygote	4	440	January 21, 2023
Error This intrinsic must be compiled to be called. Flux/Zygote Optimization (Mathematical) flux , zygote , cuarrays	5	1435	August 28, 2021