"mixed_precision fp16" not working for Flux

Why doesn't mixed_precision fp16 work for Flux?
In SDXL it was possible to train in fp16 and the result was quite normal (probably, didn't compare).

If specify mixed_precision fp16 for Flux, the training goes normally and the training speed is **5 times faster** than with bf16, but as a result LoRa turns out to be "empty", it does not make any changes when connected.

By the way, how right or wrong is it to train in FP16? Do the gains in speed and resources justify the losses?

Start by reproducing Flux training with mixed_precision set to fp16, then compare the resulting LoRA with a bf16 run and the reported SDXL behavior. Done means identifying why the fp16 LoRA is empty or ineffective and documenting or correcting the behavior, with a reproducible comparison.

Written by the indexing model from the issue text.