So once again, sorry for the massive delay as I was distracted by Stable Diffusion and other stuff.
Anyway - Here's, finally, a new Flowframes version.
It adds the latest RIFE 4.6 model to all implementations (NCNN/VS/CUDA), fixes some important bugs and UX issues.
The output format selection has been completely remade, replacing the entire settings tab. Instead, you can now pick the output container, codec, quality and color format independently, without needing to open the settings.
Image sequence output now also works with RIFE-NCNN-VS.
Some bugs with the queue have been fixed - For example, when queueing up image sequences, you are now prompted to set their frame rate the moment you add them - Not during queue processing which would make the batch processing feature pointless.
RIFE-NCNN-VS no longer shows an error if you try to interpolate inputs that have a resolution that's not divisible by 2 (can happen with GIFs or frames), instead it automatically adds padding.
Full changelog on Discord and in Flowframes.