0% found this document useful (0 votes)
7 views

Create Character

Uploaded by

elitelife029
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
7 views

Create Character

Uploaded by

elitelife029
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Creating Consistent Characters with Leonardo AI

Fixed Seed Images Method

1. Access Image Generator


● Navigate to the image generator within the left sidebar menu under the "User Tools"
section.
● Click on "Image Generation."
2. Configure Settings
● Generate four AI images at a time.
● Use the Photo Real and Alchemy pipelines if you have a paid Leonardo account.
● Set image dimensions to 512 x 768 pixels (or as preferred).
● If you plan to use these pic to train your model Leonardo prefers you to use either
512X512 or 768X768
3. Enter Initial Prompt
● Type a descriptive prompt, e.g., "female Instagram model with white hair, tan skin, blue
eyes, athletic build."
● Click "Generate."
4. Copy Image Seed
● Click the three-dot icon in the upper right corner of the generated image set.
● Select "Copy Seed."
5. Enable Fixed Seed
● Scroll down to the bottom of the left-hand sidebar and click on "Show Advanced
Features."
● Enable the "Use Fixed Seed" feature and paste the copied seed.
6. Edit Prompt for New Setting
● Edit the initial prompt to include new settings, e.g., "black dress in a hotel lobby."
● Click "Generate" to create images with the consistent character in a new setting.
7. Repeat for Different Settings
● Keep the fixed seed in place and modify the prompt for different scenarios, e.g., "sitting at
a candle-lit table in a fancy restaurant downtown" or "at a beach party during golden hour
in Southern California."
● Click "Generate" for each new prompt.

Trained Model Method

1. Download Training Images


● Download multiple images (preferably up to 40) of the character you want to train the
model on.
2. Create a New Dataset
● Go to the Leonardo home dashboard.
● Click on "Training and Datasets" in the left-hand sidebar menu.
● Click "New Dataset."
● Name your dataset and give it a description.
● Click "Create Dataset."
3. Upload Images
● In the dataset editor, upload the images you downloaded.
● Click the "Train Model" button.
4. Configure Model Training
● Name the model and provide a description.
● Set the training resolution to match your image dimensions (e.g., 512 x 512).
● Select the base model (e.g., Stable Diffusion version 1.5, 512 x 512).
● Choose a category for your model (e.g., General or Photography).
● Enter an instance prompt (e.g., "female Instagram model").
● Click "Start Training."
5. Monitor Training Process
● Monitor the job status in the "Job Status" section.
● Wait for an email notification indicating that the training is complete.
6. Use Trained Model
● Return to the image generator and turn off the Photo Real pipeline.
● Select the custom model you trained.
● Configure any additional settings (e.g., image dimensions, style).
● Enter a prompt including the instance prompt (e.g., "female Instagram model jogging in
the woods").
● Click "Generate."
7. Fine-Tune Model
● Experiment with changing the model category if needed.
● Use the fixed seed from the original image data set within the trained model for
consistency.
● Generate images and adjust prompts as necessary.

By following these steps, you can create consistent characters using Leonardo AI with both the Fixed
Seed Images method and the Trained Model method.

You might also like