ComfyUI Extension: Searge-LLM for ComfyUI v1.0

Searge-LLM for ComfyUI v1.0 Introduction

ComfyUI_Searge_LLM is an extension designed to enhance the capabilities of ComfyUI by leveraging the power of a language model. This extension acts as a prompt generator or prompt improver, transforming your initial text-to-image prompts into more detailed and refined versions. This can be particularly useful for AI artists who want to generate high-quality prompts without needing to manually craft every detail.

Custom node Searge-LLM for ComfyUI

How Searge-LLM for ComfyUI v1.0 Works

At its core, ComfyUI_Searge_LLM uses a language model to process and enhance text prompts. Think of it as a smart assistant that takes your basic idea and expands it into a more comprehensive and detailed prompt. For example, if you input a simple prompt like "a sunset over the mountains," the extension can transform it into a more vivid and descriptive prompt, such as "a breathtaking sunset casting golden hues over the rugged mountain peaks, with a serene lake reflecting the vibrant colors."

The extension works by utilizing a pre-trained language model, which has been trained on vast amounts of text data. This model understands the nuances of language and can generate text that is coherent and contextually relevant. By feeding your initial prompt into this model, ComfyUI_Searge_LLM can generate an improved version that is more likely to produce high-quality images when used in text-to-image generation tasks.

Searge-LLM for ComfyUI v1.0 Features

Prompt Generation and Improvement

• Text Input: You provide a basic text prompt.
• Model Selection: Choose from different language models to process your prompt.
• Token Limit: Set the maximum number of tokens (words or characters) for the generated text.
• Instructions: Provide specific instructions to guide the language model in generating the prompt. For example, you can instruct it to "Generate a prompt from prompt".

Advanced Options

• Temperature: Controls the randomness of the text generation. Lower values make the output more predictable, while higher values introduce more creativity and variability.
• Top-p: Also known as nucleus sampling, this parameter limits the model to considering only the top p% of tokens with the highest probabilities.
• Top-k: Limits the number of highest probability tokens considered for each step of the generation.
• Repetition Penalty: Adjusts the likelihood of repeating tokens, helping to avoid repetitive text. These features allow you to customize the text generation process to suit your specific needs, whether you want more creative and varied outputs or more consistent and predictable ones.

Searge-LLM for ComfyUI v1.0 Models

The extension currently supports the Mistral-7B-Instruct-v0.3 model, which is a powerful language model designed for generating high-quality text. This model is available in the GGUF format and can be downloaded from the HuggingFace repository.

• Mistral-7B-Instruct-v0.3: This model is well-suited for generating detailed and contextually relevant prompts. It can handle a wide range of text inputs and produce high-quality outputs that are ideal for text-to-image generation tasks.

Troubleshooting Searge-LLM for ComfyUI v1.0

Common Issues and Solutions

1. Missing llama-cpp Error:

• Solution: Open a command line interface in the ComfyUI_windows_portable/python_embeded directory and run the following commands:

python -m pip install https://github.com/oobabooga/llama-cpp-python-cuBLAS-wheels/releases/download/cpu/llama_cpp_python-0.2.89+cpuavx2-cp311-cp311-win_amd64.whl
python -m pip install https://github.com/oobabooga/llama-cpp-python-cuBLAS-wheels/releases/download/textgen-webui/llama_cpp_python_cuda-0.2.89+cu121-cp311-cp311-win_amd64.whl

2. Errors Related to llama-cpp:

• Solution: Manually install llama-cpp in the Python environment used for ComfyUI. Uninstall any existing llama-cpp packages and then run:

python -m pip install llama-cpp-python

Frequently Asked Questions

• "Can you add [FEATURE] to this node?"
• You can post your feature request on the project's GitHub issue tracker. The author may consider implementing it in a future update if it aligns with the project's goals.

Learn More about Searge-LLM for ComfyUI v1.0

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

• HuggingFace Model Repository
• llama-cpp-python GitHub Repository
• ComfyUI GitHub Repository These resources provide valuable information and support to help you get the most out of ComfyUI_Searge_LLM. Whether you're looking for detailed documentation, community forums, or additional tutorials, these links will guide you to the right place.