I saw a Pony Lora and though I will give it a try if I can make a better one.
So here we are, seems to work OK.
Can be very tricky with some Pony models and quickly looks over-trained.

Trained on the base Pony model.

V1a try 0.9-1 strength but one some you might have to go lower.
V1b try 0.6-07 strength.

Decided to share a 2nd epoch merge, might work better.