comfyui_tagger

comfyui_tagger
★ 6

标签生成批量处理图像理解保存与触发词
使用 Florence2 VLM 批量提取、生成并优化图像 prompt 标签,支持目标检测、图像描述、分割和 OCR,能添加触发词并保存到文件,便于管理与复用。
💡 批量从图片生成与优化 prompt 标签并保存备用
🍴 1 Forks💻 Python🔄 2024-09-03
📦
网盘下载
复制链接后前往夸克网盘下载
https://pan.quark.cn/s/e98a62d17551
📦 requirements.txt
transformers>=4.39.0
matplotlib
timm
pillow
>=
10.2.0
📄 README

comfyui_tagger

comfyui_tagger is a tool designed to streamline the process of generating and managing prompt tags for Stable Diffusion models. This tool offers several key features:

  • Batch reverse engineering of prompt tags
  • Saving prompt tags to files
  • Adding trigger words at the beginning of text
  • Using Florence 2 to generate prompt words such as flux, sd3, kolors, etc.
  • Features

  • Batch Reverse Engineering of Prompt Tags
  • Process multiple inputs to extract and refine prompt tags efficiently.
  • Saving Prompt Tags to Files
  • Save generated or refined prompt tags to text files for easy management and future use.
  • Adding Trigger Words at the Beginning of Text
  • Prepend specific trigger words to the start of each prompt tag, enhancing the effectiveness of the prompts.
  • Using Florence 2 to Generate Prompt Words
  • Utilize Florence 2 to generate advanced prompt words like flux, sd3, kolors, etc., improving the quality and diversity of generated images.