Transforming 16bit input to 8 bit output using inference.

[darkcristle]

UPDATE: Issue was ffmpeg was storing images as png. Changed the extension to jpg in the command, this has fixed the issue.

Hello, first off my apologies if this is obvious, I am not technically inclined.

I managed to get the pytorch version working as well as the ncnn version. I noticed a large speed discrepancy with pytorch being slower at a large factor.

The reason for this seems to be that the ncnn simply downgrades all 16bit input to 8bit and then exports 8bit frames. Is there a way to do the same using the inference instead of downconverting using another software first? Thanks!

I am using the 4x anime video model.