[Feature Request]: Template for LTX2.3 FLF2V with x2 Upscaler

### Description

Hello,

the current FLF2V template does not do a 2nd pass with upscale, so I'm limited to short ~10sec vids b/c of VRAM.

I tried implementing my own, I looked on youtube, but none them do it perfectly, they fail to properly trim the latent at the end (the last few frames always get washed and blurred). I dont know what it is, but it is deep with how the model works, it seems that "LTXVCropGuides" needs to be used implemented somehow, (or updated?)

### Additional information

_No response_

No file, test, or entry point is named. Start by locating the current FLF2V template and reviewing how LTXVCropGuides relates to the final latent frames and second upscale pass. Done means a working LTX2.3 FLF2V template that supports the requested upscale workflow without washed or blurred ending frames.

Written by the indexing model from the issue text.