Updated training data, no longer using images generated with detailer loras
Updated the training data, hopefully will now work in conjunction with other loras better. Best used at .8 strength.
