Its not working with fp8_e4m3fn. #32
Description
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
{{ message }}
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