Flux:
Use 1.0 strength.
If an image is blurry switch the seed to fix.
If the lora has no effect update your comfyui. If that doesn't work try a clean install. (comfy just fixed flux lora yesterday)
Note: V1 is affected by a bug in the trainer. V2 should work better.
Prompts/ComfyUI workflow embedded in images.
PDXL:
There was a comment that asked for me to train this in pony. Here you go! PDXL is already pretty capable so its a very likely a little redundant.
I recommend to use [0.6-1.0] weights.