I’m trying to use Enzyme to autodiff a function with respect to a neural network’s parameters. I’m running into this error: No augmented forward pass found for jl_genericmemory_slice at context: %93 = call nonnull "enzyme_type"="{[-1]:Pointer}" {} addrspace(10)* @jl_genericmemory_slice({} addrsp…

I think you would just do adtype = AutoEnzyme(; mode=set_runtime_activity(Forward)) or adtype = AutoEnzyme(; mode=set_runtime_activity(Reverse)), depending on your preference

Error using Enzyme to autodiff wrt Flux neural net parameters

General Usage

wsmoses December 17, 2024, 11:30pm 4

This is just saying we haven’t added that 1.11 intrinsic support yet (and should).

The canonicalization failed one is weird, can you file an issue with that. That shouldn’t happen ever.

1 Like

Topic		Replies	Views
Autodiff of vector inputs with Enzyme.jl (and possibly Optimization.jl) Optimization (Mathematical) question , optimization , sciml , enzyme	9	1132	August 16, 2023
What's the state of Automatic Differentiation in Julia January 2023? General Usage autodiff	41	12431	June 21, 2024
Reliability of Enzyme.jl General Usage question , diffeq , autodiff	11	2114	October 23, 2022
Enzyme with Const() on a vector throws an error Optimization (Mathematical) autodiff , enzyme	20	537	April 10, 2025
Lux, ComponentArrays and flat parameters : computing the gradient works with Zygote but not with Enzyme New to Julia enzyme	16	1785	May 14, 2024

Error using Enzyme to autodiff wrt Flux neural net parameters

Related topics