Inference Logoinference.sh

Variants

Different configurations of the same app.


What are variants?

An app can have multiple variants optimized for different needs:

VariantTypical use
defaultBalanced quality and speed
fastQuick results, lighter model
qualityBest output, more resources
cpuNo GPU required

Choosing a variant

In the app runner, select before running:

code
1Variant: [quality ]2 3 default (8GB VRAM)4 fast (4GB VRAM)5 quality (16GB VRAM)  6 cpu (no GPU)

Why variants matter

Speed vs quality trade-off:

  • fast might take 5 seconds
  • quality might take 30 seconds but look better

Resource requirements:

  • Some variants need more GPU memory
  • cpu variants work without a GPU

Private workers and variants

If using private workers, make sure your hardware matches the variant's requirements.

A quality variant needing 16GB VRAM won't run on an 8GB GPU.


Next

Creating a Flow

we use cookies

we use cookies to ensure you get the best experience on our website. for more information on how we use cookies, please see our cookie policy.

by clicking "accept", you agree to our use of cookies.
learn more.