
- A hand-drawn comic look
- Watercolor or oil painting styles
- Fashion lookbooks
- Filmic tones (e.g., Wes Anderson, Noir, etc.)
- Brand style guidelines
🧭 How It Works
- Go to your Atlabs dashboard
- Click “Models” from the top navigation bar.
- Click the “Train New Model” button at the top-right corner
- Choose “Style Model”
- Give your model a name
- Give your model a trigger word (any unique alphanumeric word works here; for e.g.: ATL_BAR1 )
- Upload 2–150 reference images in the style you want
- Click “Generate Captions”. Our AI will analyse the images and write the appropriate captions for you.
- “Start Training” to begin
- A hand-drawn comic look
- Watercolor or oil painting styles
- Fashion lookbooks
- Filmic tones (e.g., Wes Anderson, Noir, etc.)
- Brand style guidelines
Don’t mix drastically different styles. Pick a clear aesthetic.
🖼 Using Your Style Model
Once trained, your style model will appear under “Visual Style” whenever you:- Generate a video from script or audio
- Regenerate a visual in a frame
- Apply style to an image prompt
📌 Notes
- You can train multiple style models per account.
- Each model is private to your workspace unless shared.
- You can delete or archive old models anytime. Just ping us on live chat for any help.
🧠 Pro Tips
- Add consistent background or framing to improve style extraction.
- Mix with platform characters for stylized storytelling.
🛠 FAQ
Can I retrain or update a style model later?Not yet. You can delete and re-upload a better version. What’s the difference between character vs. style models?
Character = Face + Identity.
Style = Aesthetic + Visual feel. Can I use a style model with platform characters?
Yes! Apply it like a filter across all visuals.
Need help? Reach out to us at support@atlabs.ai.