Omost文生图模型

这个模型由ControlNet作者发布,通过Llama3模型生成Canvas代码,再用Diffusion模型生成最终的图片。模型支持全局和局部描述,局部描述支持location、offset和area三个参数,

作者还提出了Prompt Prefix Tree,用来生成更多的prompts,

参考资料

  • https://mp.weixin.qq.com/s/XSj4LKebD_QIxN3ruKjcJA
  • https://huggingface.co/spaces/lllyasviel/Omost
  • https://huggingface.co/lllyasviel/omost-llama-3-8b/tree/main
  • https://civitai.com/models/139562/realvisxl-v40-turbo

Leave a Reply

Your email address will not be published. Required fields are marked *