This LoRA is trained on Pony and utilizes 1,247 images of Moona Hoshinova for training. The process focused on clothing adaptability, avoiding image repetition to enhance flexibility in outfit generation.

What can be generated:

• Default costume with moona hoshinova \(1st costume\) trigger.

• Basically any clothing

What cannot be generated (accurately):

• Any other Moona costume (due to lack of dataset)

===

Regarding Early Access

This model is in early access to help offset the server rental costs for training. All buzz collected goes directly towards these expenses. The model will still be freely available in 3 days if the cost-recovery target isn't met. Your support helps make continued development possible. Thank you for understanding!