Hi, its a problem with the non - flux models, its something we are actively working on. 

Based AI

I wonder if maybe youre leaving the tab and the sockets are not running. Because the socket needs to be active to get a stream from the backend. Its possible when that connection drops it never re-establishes. If this is true, then we need a backup polling strategy or something.

Michael Aubry

How often are you talking about? I try to test the generator every hour for non Flux models. I dont run into this often. Anything youre doing that i can follow to recreate the issue?

In playing around today, I found that the 'model' you choose makes a big difference.  Example - 'FLUX Schnell' seems to work better than 'Juggernaut XL' and 'RealVisXL'

brucevn

Andrew said this happens when you sit in the /generate page for long time and do many generations? I designed it mostly to be quick here then the rest is done in the editor. But is seems ppl like sitting on the /generate page long time. So I will insepct and fix. Is this also your way of doing thing?

It is just me or does this lock up a lot?

Performance

BasedLabs Support

Discuss AI video generation and related topics.

AI Video

Discuss AI art generation and related topics.

AI Art

Get help with BasedLabs and our products.

aiwars

A thoughtfully curated gallery of aesthetic wallpapers and aesthetic backgrounds for your devices. Browse through minimalist, vintage, and coquette aesthetic designs, featuring unique aesthetic pictures that range from nature-inspired to modern abstract art.

aesthetics

A community space to share and discover your next anime PFP, Discord PFP, or funny PFP. Browse trending cool PFP and cute PFP collections, share your favorite aesthetic PFP designs, and get inspiration for your next profile picture