| Hunyuan | 3D personalized character models

This LORA is trained based on the basic model Tencent Hybrid-DiT, welcome everyone to try different things

Full body effects work better

Recommended large model: Tencent Hybrid-DiT

Example prompting words:

1girl, solo, stuffed toy, stuffed animal, bubble tea, teddy bear, black hair, overalls, jewelry, hat, holding, smile, earrings, plaid, cup, full body, holding cup, open mouth, white background, simple background, socks, twin braids, blush, shoes, shirt, shorts, black shirt

Negative prompting words:

ng_deepnegative_v1_75t, (badhandv4:1.2), EasyNegative, (worst quality:2),

Sampling method: Euler a, DPM++2M Karras

Iteration steps: 35

Recommended size: 576*1024 (other sizes are welcome to try)

Prompting word guidance coefficient (CFG): 7.0