
There are many Emiru LoRA's out there (to be fair, most of them aren't for Pony), and my aim isn't to create the best one, but to have a better suit at different situations than those that exist. The model was made with OneTrainer, with ~100 images running 10 epochs twice, instead of 20 epochs, creating a nearly inflexible idea of what it should represent.
This was trained to have a higher 'strength', or 'density', so it pops even under biased Checkpoints. Check the images attached to the post for more information on their generation methods.
描述:
训练词语:
名称: EmiPonyXLdense.safetensors
大小 (KB): 111881
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success