Its not working with fp8_e4m3fn.

[Dervlex]

When using the LoRA with the BF16 model but set the weidht_dtype to fp8_e4m3fn the output looks sadly bad.
Any idea to fix this ?

Default:

FP8

[X-niper]

Hi, fp8_e4m3fn base model is now supported, please check the solution here https://github.com/ModelTC/Qwen-Image-Lightning?tab=readme-ov-file#-using-lightning-loras-with-fp8-models

[caffeinism]

Hi, fp8_e4m3fn base model is now supported, please check the solution here https://github.com/ModelTC/Qwen-Image-Lightning?tab=readme-ov-file#-using-lightning-loras-with-fp8-models

Is there no alternative to the 8-step method?