Training an Object Model
Object models learn to recognize and accurately recreate specific items—perfect for products, branded items, or unique objects you want to feature in generated images.
Ideal Use Cases
Product photography: Generate your products in any setting
Marketing materials: Create variations for campaigns
E-commerce: Generate lifestyle images with your products
Brand assets: Feature branded items consistently
Image Requirements
Number of Images: 10-20 photos
Image Guidelines:
Object clearly visible and in focus
Multiple angles (front, side, top, 3/4)
Clean backgrounds preferred
Consistent lighting across images
High resolution for detail capture
Object fills significant portion of frame
What to Include
Good training images:
Full product shots from various angles
Close-up details of distinctive features
Neutral backgrounds (white, gray)
Consistent scale across images
Professional-quality photography
Avoid:
Cluttered backgrounds
Partial views of the object
Heavily edited or filtered images
Mixed products in same training set
Blurry or low-resolution photos
Training Process
Navigate to Models > Create New Model
Select Object as the model type
Upload your training images
Name your object model
Click Start Training
Wait approximately 15-30 minutes
Credit Cost
Training an Object model costs 400 credits.
Using Your Object Model
Place your trained object in any scene:
> "[trigger word] on a wooden table, morning sunlight, lifestyle photography" > "[trigger word] being held by a hand, minimalist background" > "[trigger word] in an outdoor setting, nature background, product shot"
Tips for Product Photography
Consider the context: Think about where your product would naturally appear
Match lighting styles: Describe lighting that suits your product
Use lifestyle scenarios: Place products in real-world situations
Maintain brand consistency: Keep generated images aligned with your brand