Promptmaster-mistral-7b

Model Card for Model ID

<!-- Provide a quick summary of what the model is/does. -->

This model is a finetune of Mistral with my custom dataset to create a model specialized in Stable Diffusion prompts.

Direct Use

The model was fine tuned to respond to the following prompt format:

<|im_start|>system\nYou are a prompt generation AI. your task is to take a user input for a stable difusion prompt and output and expand the supplied prompt in a stable difusion format to provide better output. Do not deviate from the format. Do not output anything other than a stable diffusion prompt.<|im_end|>\n<|im_start|>user\n{input}<|im_end|>\n<|im_start|>assistant\n

simply replace {input} with your basic prompt of choice and the model will return an expanded prompt with added (mostly) appropriate keywords and formatting.

Bias, Risks, and Limitations

Look, not going to beat around the bush here. A LOT of stable diffusion generation out there publicly available is... not safe for work. This model was trained on data that includes many terms that some people may find offensive. It does have a bias towards outputting nudity and worse even when not specificly prompted. If you so much as mention the word 'woman' in your prompt there is a a very high probability the output will include some terms that would make a sailor blush.

You have been warned.