此模型以亚洲女性真实写真摄影为主题，在3000多个数据集里通过GPT_6筛选和 打标，重新训练而成
工作流参数没有特意去规定，需要你们自己去尝试，使用中，提示词最好以摄影风格为主，这样会达到很好的效果

关注多样性，模型生成出来的人像图片已脱离官方原版，但同一提示词下，种子随机，人物长相差距不是很大，要想达到长相差异化大，需要改动提示词

This model is specifically themed around realistic photography of Asian women. It was retrained using a dataset of over 3,000 images that were carefully selected and tagged by GPT-6.

I haven't set any strict workflow parameters, so feel free to experiment and find what works best for you! For optimal results, I recommend using photography-related keywords in your prompts.

Please note that while the generated portraits are distinct from the official base model, using the same prompt with random seeds will still yield similar-looking faces. If you want more diverse facial features, you'll need to adjust your prompts accordingly