Maintaining uniform AI-generated character designs is useful in a variety of contexts, from creating a cover for a book series, to generating concept art. However, it can be tricky to do – particularly when using an illustrative style.
To make the process easier, we have created this handy guide with everything you need to get you creating consistent characters. The guide features two different methods: without model training and with model training.
Method 1. Without Model Training
Step 1: Start with a Precise Initial Prompt:
To start you’ll need to create a detailed initial prompt. This prompt should elaborate on the your character's primary attributes and establish a base context.
For example you could say: "n illustration of Jason Segen, a male teenage character, short black hair, expressive eyes and large spectacles, wearing a baseball cap, standing in a city street." This initital prompt provides a blueprint for your character's look and initial surroundings.
Step 2: Prompt Diversifications:
Now it’s time to work on making your character appear in new settings and different poses. At this point it’s crucial to preserve a consistent character identity. For example you will need to continue to use the fictional name ‘Jason Segen’ that you began with. This continuity maintains the character's predefined look and traits, even when the activity or setting changes, along with that, it is ideal to keep the visual description of the character consistent across prompts as well.
For example: "an illustration of Jason Segen, a male teenage character, short black hair, expressive eyes and large spectacles, wearing a baseball cap, playing with a friendly cat in a field."
Remember that if using a fixed seed, always vary the wording of your prompt, otherwise you’ll get the same set of images over and over again. However, ensure these variations don’t change your character description blueprint.
If you're still unsatisfied with the results and need a higher level of consistency for your project, we recommend Method 2, creating a consistent character by training a custom model.
Method 2. Training a Custom Model
This second method requires more time and effort, as well as a set of source images,but can yield extremely effective and consistent results when done correctly. It involves training your own fine-tuned model, followed by additional steps to create character coherence.
Because training a custom model can be an in depth process, this guide will, focus more on the specific elements of model training necessary to achieve a consistent character. Click here for a more detailed guide on How to Train a Model.
Step 1: Start Training
Navigate to 'Training and Datasets' section after logging into the Leonardo.Ai app. (Please note that on free plans you are limited to training one model at a time).
Select 'New Dataset' and name it after the character you want to consistently generate (e.g., "Anna" or “John”). This helps the AI to associate that name with a need for repeated generation.
Provide a succinct description for the dataset to outline its use case, for example: “for character design portfolio” or “book covers for book series.”
Step 2: Upload Reference Images
Upload at least 10 reference images to help generate your baseline character. For bettter results, upload 15 or more images. These images should represent the characteristics of the character you wish to create.
You can use stock footage, drawings, or photographs, but it’s important that you always have the legal permission to do so. (Whether that’s the licence to certain images or permission to use a real person’s likeness).
Step 3: Train Your Model
After uploading the selected images, click 'Train Model' to begin the training process.
You can verify the dataset's completion under 'Finetuned Models' > 'Your Models' or under the 'Job Status' section of ‘Training & Datasets’. Once available, it will be listed with the name provided earlier (e.g., "Anna” or “John.”). Note that there is no image preview of your finetuned model until you have started generating images with it.
Step 4: Generate Images
Go to 'Image AI Generation' and select your newly created fine-tuned model from the available options.
To begin image generation, use creative prompts that include the name of your dataset to direct the AI to utilize the trained model.For example: “John Smith wearing a suit and carrying an umbrella, stands in the entrance to a tall office block.”
It is likely that some images still won’t look exactly like what you want. Experiment with prompt structure, incorporating negative prompts as necessary to refine and vary your results while maintaining consistency.
Step 5: Refine and Evaluate
Review your generated images for precision and consistency. Note that in a generation of 4 images, typically, 1-2 images may closely match the desired output, and the rest will not. That’s just part of the creative process!
Adjust the settings to balance creativity with consistency and if you have access to premium services such as Alchemy, consider activating them to enhance image resolution and quality.
Download successful creations and add them to your dataset to continually improve character coherence.
Final Tips:
Remember that when training your model for character consistency, the quantity and quality of your reference images directly influence the model’s performance.
Experimenting and tweaking your prompts and settings is essential to achieve the most best outcome. You’ll find that once the model has been sufficiently trained, producing a stylistically consistent character will get faster and faster!
That does it for our guide to creating a consistent character – we hope it helped! Remember, we're always adding new features and updating old ones, so be sure to check back in for new information and tutorials.
Happy Prompting! 🎨