See LuxDL/Lux.jl#1554
MWE:
using Lux, Random, Optimisers, Enzyme, Printf
cdev = cpu_device()
# xdev = reactant_device()
xdev = cpu_device()
rng = Random.default_rng()
Random.seed!(rng, 0)
X, Y = rand32(rng, 10, 24), rand32(rng, 1, 24)
model = Chain(
Dense(10, 10, sin),
Dense(10, 1, sin)
)
ps, st = Lux.setup(rng, model)
ts = Training.TrainState(model, ps, st, AdamW())
(_, loss_j, _, ts) = Training.single_train_step!(AutoEnzyme(), MSELoss(), (X, Y), ts)Julia crash:
$ julia mwe.jl
error: Linking globals named 'julia_sin_25307': symbol multiply defined!
Version info:
Julia Version 1.11.7
Commit f2b3dbda30a (2025-09-08 12:10 UTC)
Build Info:
Official https://julialang.org/ release
Platform Info:
OS: Linux (x86_64-linux-gnu)
CPU: 12 × AMD Ryzen 5 8540U w/ Radeon 740M Graphics
WORD_SIZE: 64
LLVM: libLLVM-16.0.6 (ORCJIT, znver3)
Threads: 1 default, 0 interactive, 1 GC (on 12 virtual cores)
See LuxDL/Lux.jl#1554
MWE:
Julia crash:
Version info: