Not the author, but I normally wouldn't expect one. Flux requires totally different tagging compared to 1.5/PDXL/Illustrious (where as if you make a tagged training data set for one, you've got data set that will generally work for all three of those, minus some minor edge cases the involve heavy use of renamed/deprecated/merged/new booru tags owing to them being based off differently aged dumps), is 4x as expensive to train, and there's already a pair of models to replicate Syd Mead's style for Flux.