GPU based HEVC encoding and decoding

Hello All, I built the GPU based video encoding and decoding ffmpeg using this blog.

I am specifically looking for - let’s say I have a video as a pytorch tensor - for example, of shape: 30x3x1920x1080 - 30 frames of a video of size 1920x1080. Can this ffmpeg build takes in pytorch tensor and do hevc compression?

If not this build, is there any other tool that works off of pytorch tensor directly?