A suggest would be to disable either the upscale box or the standard VFI/save combo by default. Running two rounds of VFI is quite slow. If you're aiming at newer users, that might help it get a better initial impression.

As a side note: FILM-VFI looks great but is quite slow. (If you have the CUDA support for it, GIMM is faster and produces identical results.)

is using i2v lightning lora better than the t2v one?

What Loras were used in your example video? I ran them using your workflow and did not achieve the effect shown in the example

how would you turn this into a image to video:>