|
Even if the was an unconcerned model, current video models create around 20secs, there was a model released in the last week that can handle 60sec, but again, these model have censorship and you would have to run these kind of things locally, so you would need the hardware. LTX 2.3 drop in the last week.
There was a model last month that would cascade, you insert a reference image to start and you use the last frame to start the process again.
But at the moment... like the guy said above.
|