Get help faster and easier

New user?

Last updated on Sep 11, 2024

Explore strategies and best practices for training custom subject models.

Try it in the app
Create a custom model with images of your brand's assets in a few simple steps.

When training custom models on a specific subject — whether objects or characters — the goal is to identify the subject’s essential features and help the model replicate them across various contexts and positions.

Images

Look for images with the following characteristics when training a subject model:

  • Object consistency: Provide images of the same make and model as your subject while ensuring that the subject doesn't look widely different across images. Avoid mixing multiple colors and ensure a common theme or pattern among images. However, your subject can vary across scenes, poses, clothing, and background.
  • Object focus: Use images of the subject in clear focus without unnecessary distractions. Keep the subject near the center of the image and make sure that it occupies at least 25% of the image's area.
  • Environmental context: Provide images of the subject in different views and contexts, showing it in a variety of lighting conditions. While images with white or transparent backgrounds can be used, it's best to have a mix with more complex surroundings as well.
  • Avoid other objects: Avoid large items in the background or associated with the character. Any large item shown in the images is memorized by the model and will appear in the generated images, similar to the same item in the training dataset. 

  Sample images to use

Four separate panels, each featuring a stylized, cartoonish character against a pink background. The character is a fluffy, round creature with two large eyes and small limbs. This image is to depict the types of images that are suitable and optimal for training custom models in Firefly.
Use a set of images that show the same subject with a consistent style for best results.

  Sample images to avoid

Four panels, each featuring different images. From a small, round creature with blue fur and a larger creature with purple fur and multiple eyes to a Blurred image of a creature with blue fur and two large eyes. This image is to depict the types of images that are not suitable for training custom models in Firefly.
Low-quality images with varying subjects and styles can result in less effective models and hallucinations.

Captions

  • Concisely describe the image and objects distinct from the subject in the caption. This helps the model distinguish the subject from other objects.
  • When captioning, if a key adjective or descriptor co-occurs with the concept and you would like the concept to always be generated matching that adjective, it should be included in the concept to get the best results. For example, you want a custom model to generate gray cats, so your concept is "cat", and your adjective is "gray". If you always want the model to generate a gray cat, you should include that in the concept and make the concept "gray cat".

Concept

  • Ensure that the Concept entered has at least three characters. 
  • Use a proper noun that represents your model's main subject. For example, if your model was trained on your dog, "Spot," enter "Spot" in the Concept field.

More like this

Join our community to connect, learn, and engage

For inspiration, expert tips, and solutions to common issues, visit Discord or the Adobe Firefly Community forum. Connect with our team and fellow users to exchange ideas, share your creations, stay updated with the latest features and announcements, and provide feedback.

Get help faster and easier

New user?