ComfyUI Extension: ComfyUI-Prompt-MZ

ComfyUI-Prompt-MZ Introduction

ComfyUI-Prompt-MZ is an extension designed to enhance your experience with prompt words in AI art generation. It leverages the capabilities of llama.cpp to provide various nodes that help in beautifying prompt words and performing image interrogation, similar to the functionality offered by clip-interrogator. This extension is particularly useful for AI artists who want to refine their prompts and improve the quality of their generated images.

How ComfyUI-Prompt-MZ Works

ComfyUI-Prompt-MZ works by integrating with the ComfyUI framework and utilizing llama.cpp to process and enhance prompt words. The extension provides nodes that can beautify prompt words, making them more effective for generating high-quality images. Additionally, it includes nodes for image interrogation, which can analyze an image and generate descriptive prompts based on its content. This helps in creating more accurate and detailed prompts for AI art generation.

ComfyUI-Prompt-MZ Features

Beautifying Prompt Words

The extension includes nodes that can take your input prompts and enhance them by making them more descriptive and aesthetically pleasing. This can help in generating better and more detailed images.

Image Interrogation

Similar to clip-interrogator, this feature allows you to analyze an image and generate prompts based on its content. This is useful for creating prompts that are closely aligned with the visual elements of the image.

Model Selection

ComfyUI-Prompt-MZ provides nodes that allow you to select different models for processing your prompts. You can choose from local models or preset models, giving you flexibility in how you want to process your prompts.

Batch Generation

The extension supports batch generation of prompt words, making it easier to process multiple prompts at once. This is particularly useful for artists who need to generate a large number of prompts quickly.

Custom Instructions

You can add custom instructions to your prompts, allowing for more personalized and specific prompt generation. This feature is useful for artists who have specific requirements for their prompts.

ComfyUI-Prompt-MZ Models

Florence-2-large

This model is used for image interrogation and can generate detailed prompts based on the content of an image. It is useful for creating prompts that are closely aligned with the visual elements of the image.

Qianwen 2.0

A preset model that can be used for generating high-quality prompts. It is designed to provide detailed and descriptive prompts for AI art generation.

LLava-v1.6-vicuna-13b

This model is used for generating prompts and can be customized with different settings to achieve the desired output.

Phi-3-mini

A smaller model that can be used for generating prompts quickly. It is useful for artists who need to generate a large number of prompts in a short amount of time.

What's New with ComfyUI-Prompt-MZ

Recent Changes

• 2024-06-22: Added Florence-2-large image interrogation model node.
• 2024-06-20: Added nodes to select local ollama models.
• 2024-06-05: Added Qianwen 2.0 preset model.
• 2024-06-05: Optional chat_format, post-processing after image interrogation.
• 2024-06-04: Added some preset models.
• 2024-06-04: Added universal node, support manual selection of models.
• 2024-05-30: Added ImageCaptionerConfig node to support batch generation of prompt words.
• 2024-05-24: Display the generated prompt words in the current node after running.
• 2024-05-24: Compatible with Zhipu API.
• 2024-05-24: Use A1111 weight scaling, thanks to ComfyUI_ADV_CLIP_emb.
• 2024-05-13: Added OpenAI API node.
• 2024-04-30: Support for custom instructions.
• 2024-04-30: Added llava-v1.6-vicuna-13b.
• 2024-04-30: Added translation.
• 2024-04-28: Added Phi-3-mini node.

Troubleshooting ComfyUI-Prompt-MZ

Common Issues and Solutions

Module 'llama_cpp' has no attribute 'LLAMA_SPLIT_MODE_LAYER'

Solution: Upgrade the llama_cpp_python version to the latest version. You can download and install it from llama-cpp-python releases.

Failed to load shared library LLama.dll

Solution: Switch the CUDA version to 12.1. If you are using the Qiuye launcher, go to Advanced Settings -> Environment Maintenance -> Install PyTorch -> Select the version with CUDA 12.1.

llama_cpp_python-0.2.63-cp310-cp310-win_and64.whl returned nonzero exit status

Solution: Ensure a stable internet connection. If necessary, use a VPN or manually install llama_cpp_python.

Learn More about ComfyUI-Prompt-MZ

For additional resources, tutorials, and community support, you can visit the following links:

• ComfyUI GitHub Repository
• llama.cpp GitHub Repository
• ComfyUI_ADV_CLIP_emb GitHub Repository These resources provide detailed documentation, examples, and community forums where you can ask questions and get support from other AI artists and developers.