ComfyUI Node: Danbooru Tag Generator

Danbooru Tag Generator:

The PromptDanTagGen node, also known as the Danbooru Tag Generator, is designed to assist AI artists in generating detailed and contextually relevant tags for their artwork. This node leverages a language model to predict and generate tags based on various input parameters such as artist name, characters, copyrights, special tags, and general descriptions. By providing a comprehensive set of inputs, the node can produce a rich set of tags that enhance the discoverability and categorization of artwork. The main goal of this node is to streamline the tagging process, making it easier for artists to manage and organize their digital creations effectively.

Danbooru Tag Generator Input Parameters:

model

This parameter specifies the path to the pre-trained language model that will be used for generating tags. The model is crucial as it determines the quality and relevance of the generated tags.

artist

This is a string input where you can specify the name of the artist. It helps in generating tags that are specific to the artist's style and works. The default value is an empty string.

characters

This string input allows you to list the characters featured in the artwork. Including character names can help in generating more accurate and relevant tags. The default value is an empty string.

copyrights

This string input is used to specify any copyrights associated with the artwork. This information is important for generating tags that respect intellectual property rights. The default value is an empty string.

special_tags

This string input allows you to include any special tags that you want to be considered in the tag generation process. These tags can be specific attributes or themes related to the artwork. The default value is an empty string.

general

This is a multiline string input where you can provide a general description of the artwork. This description helps in generating a broader set of tags that capture the overall essence of the artwork. The default value is an empty string.

blacklist

This string input allows you to specify any tags that should be excluded from the generated tags. This is useful for avoiding irrelevant or inappropriate tags. The default value is an empty string.

rating

This parameter allows you to specify the rating of the artwork. The available options are "safe", "sensitive", "nsfw", and "nsfw, explicit". The rating helps in generating tags that are appropriate for the specified content rating.

target

This parameter specifies the target category for the generated tags. The available options are based on predefined target keys, which help in tailoring the tags to specific categories or themes.

width

This integer input specifies the width of the artwork in pixels. It helps in generating tags that are relevant to the artwork's dimensions. The default value is 1024, with a minimum of 256 and a maximum of 4096, adjustable in steps of 32.

height

This integer input specifies the height of the artwork in pixels. Similar to the width, it helps in generating tags that are relevant to the artwork's dimensions. The default value is 1024, with a minimum of 256 and a maximum of 4096, adjustable in steps of 32.

escape_bracket

This boolean input determines whether to escape brackets in the generated tags. This is useful for ensuring that special characters are handled correctly. The default value is False.

temperature

This float input controls the randomness of the tag generation process. A higher temperature value results in more diverse tags, while a lower value produces more focused tags. The default value is 1.35, adjustable in steps of 0.05.

Danbooru Tag Generator Output Parameters:

output

This output parameter returns the final set of generated tags as a string. These tags are designed to be comprehensive and relevant to the provided input parameters, enhancing the discoverability and categorization of the artwork.

llm_output

This output parameter provides the raw output from the language model, which includes the generated tags before any post-processing. This can be useful for understanding the model's predictions and making further adjustments if needed.

Danbooru Tag Generator Usage Tips:

• Ensure that you provide detailed and accurate input parameters to get the most relevant and useful tags.
• Use the blacklist parameter to exclude any tags that you find irrelevant or inappropriate for your artwork.
• Adjust the temperature parameter to control the diversity of the generated tags. A higher value can produce more creative tags, while a lower value can result in more focused and specific tags.
• Utilize the escape_bracket option if your artwork includes special characters that need to be handled correctly in the tags.

Danbooru Tag Generator Common Errors and Solutions:

Model not found

• Explanation: The specified model path is incorrect or the model is not available.
• Solution: Verify the model path and ensure that the model is correctly installed and accessible.

Invalid input parameters

• Explanation: One or more input parameters are not provided in the correct format or are out of the acceptable range.
• Solution: Check the input parameters for correct format and values. Refer to the input parameter descriptions for guidance.

Tag generation failed

• Explanation: The language model encountered an issue during the tag generation process.
• Solution: Ensure that all required input parameters are provided and try adjusting the temperature or other settings. If the issue persists, check the model and tokenizer for any compatibility issues.