
I decided to train a lora on pdxl for my oc, Ascella. This was trained on ponydiffusion v6 so I have no clue if this works well on any other sdxl model.
Human:
Here are some relevant tags: ascella, 1girl, tank top, leggings, skirt, watch, jacket
If hair is giving issues try: undercut, ponytail
Anthro v2: (Better anthro version, Trained on AutismMix)
Here are some relevant tags: ascella, white tank top, black leggings, blue pleated skirt, ponytail, long hair
Anthro v1: (came out overcooked)
Here are some relevant tags: ascella, tank top, leggings, skirt, ponytail, messy hair
If colors aren't correct try putting orange fur in negatives or put purple fur in positives
描述:
Made with better dataset images with better tagging. Quality overall should be way better. Design is also more accurate to the intended one. Also trained on AutismMix.
She still has two hairstyles, ponytail and long. Long hair/hair down is default so tag ponytail to get that style.
训练词语: ascella,ponytail
名称: Ascella.safetensors
大小 (KB): 223097
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success