Copy and Pasted from Reddit, thought it may be of interest to some on here. Please use responsibly.
1) Download 20-30 images of a person from the internet. The images should be of fairly high quality, with consistent face, age, and body features. The images can be NSFW.
2) Visit [Login or Register to view], and create an account or sign in with your email. It is free.
3) Go into your settings and change the slider so as to allow 'view inappropriate content'.
4) Select 'train model', and upload the images to create your LORA model. Ensure you select the private setting. This should usually take around half an hour, or less, to create.
5) Once this is done, you can select 'Create' to start creating images (based on FLUX). You can also upload image prompts for image2image creation.
6) Add your LORA model by clicking on the plus sign, add the trigger word, write your prompt, and click 'generate'. The images will take anything between 5 seconds and a minute to generate. You can also add in more LORA models (3 in total) created by other users, or import them from Civitai website. These include NSFW models.
For example, if you are a Yankees fan like me (home town) and wanted to create an image of a woman standing in a kitchen, your prompt could be:-
Alternatively, if you have a fast gaming GPU like an RTX4090 and have plenty of memory, then you could download Pinokio AI in order to run AI locally, and within that you can download Flux, etc (or any one of many AI programs), and get creating.
1) Download 20-30 images of a person from the internet. The images should be of fairly high quality, with consistent face, age, and body features. The images can be NSFW.
2) Visit [Login or Register to view], and create an account or sign in with your email. It is free.
3) Go into your settings and change the slider so as to allow 'view inappropriate content'.
4) Select 'train model', and upload the images to create your LORA model. Ensure you select the private setting. This should usually take around half an hour, or less, to create.
5) Once this is done, you can select 'Create' to start creating images (based on FLUX). You can also upload image prompts for image2image creation.
6) Add your LORA model by clicking on the plus sign, add the trigger word, write your prompt, and click 'generate'. The images will take anything between 5 seconds and a minute to generate. You can also add in more LORA models (3 in total) created by other users, or import them from Civitai website. These include NSFW models.
For example, if you are a Yankees fan like me (home town) and wanted to create an image of a woman standing in a kitchen, your prompt could be:-
Photo real, realistic, kitchen setting, sunny day, full length front profile, woman facing camera, black hair tied back in a bun, busty, short height, slim petite frame, wearing a New York Yankees sweater, wearing slim jeans,
Alternatively, if you have a fast gaming GPU like an RTX4090 and have plenty of memory, then you could download Pinokio AI in order to run AI locally, and within that you can download Flux, etc (or any one of many AI programs), and get creating.