Support ensemble in RIFE?

[banjaminicc]

Running the 4.x models with ensemble is known to provide better results(usually with less warping) at the cost of speed and can be quite useful in certain scenarios
Is it possible to implement it in vsmlrt.py or will we have to build our own model then run it with vsmlrt?

[WolframRhodium]

The implementation is possible. We want to achieve a balance between many high-quality built-in models and binary size, and we are busy on other projects right now, so this may not be currently implemented.

[WolframRhodium]

done, you need to update vsmlrt.py and download ensemble models here

[banjaminicc]

Thanks for the insanely quick update haha
cheers!