请勿转载或转发至其他网站。 商业目的也被严格禁止。
Please do not reprint or redistribute it to other sites. Commercial purposes are also strictly prohibited.
我基于Flux 1 Dev,用30多张照片分别通过512x512, 512x768, 768x1024三个尺寸训练了1000步,在权重0.55-1之间表现良好。
Based on Flux 1 Dev, I used more than 30 photos to train 1000 steps with three sizes of 512x512, 512x768, and 768x1024, and performed well between weights 0.55-1.
首次贡献,请各位大师多多支持,欢迎留言。
This is my first contribution, please support me and leave a message.
Description
This version is my first attempt at a portrait, and technical improvements will be made later.
Details
Downloads
123
Platform
CivitAI
Platform Status
Deleted
Created
10/21/2024
Updated
7/7/2025
Deleted
5/23/2025
Trigger Words:
LiQin
Files
flux_lora_face_LiQin_rank16_bf16.safetensors
Mirrors
Huggingface (1 mirrors)
CivitAI (1 mirrors)