Relax! Flux is the ML library that doesn't make you tensor
21st century AD
Please do not feed the models
Geometric Deep Learning for Flux
Computer vision models for Flux
Read ONNX graphs in Julia
A flexible neural net training library inspired by fast.ai
Parameterise all the things
Distributed computation of differentiation pipelines to use multiple workers, devices, GPU, etc. since Julia wasn't fast enough already
I heard you like compile times
Flux Website