mr-z-2697 / ddfi-rife Goto Github PK

Hi

I am new about this, just have that question, how can i convert a video keeping the audio?

Thanks for this amazing scrip.

does version 1.3.1 use cuda no instead of nncn-Vulkan?
librife.dll
because it gives me an error and my video card only supports vulkan
the other version that uses the
RIFE.dll
works perfectly but this new one doesn't work

[pipe:: Invalid data found when processing input]

My guess is that this project removes duplicated frames and replaces them with interpolated ones? (creating a necessary number of interpolated frames in place of duplicated frames to make the motion actually smoother?)

Is it trying to achieve something like this?

hello I'm not able to run the application every time it appears

ddfi2.py: error: the following arguments are required: -i/--input

could you please send a complete command line as an example
I'm using version: ddfi_R_2-v1.0.1

Could you create a google colab notebook for this project? I'd really appreciate it. And could you add an option to choose rife model 2.3? Thanks!

There seems to be an issue on the very first stage of the process.
At first, the generated script parse.py does not work, I've even tried to just load the mere video source and the script does not output properly anything.
Then, I reused a .py vapoursynth script; step by step I saw that every previous instruction outputted properly, until the instruction props2csv seems to not output anything. Naturally, it is a function that should just evaluate every frame and not process the frames as it was any video.
Yet, checking infos_running.txt, there are not output... so, I tried to pointing culprits:

1. First candidate: xvs is outdated and no longer provides proper outputs for props2csv
2. Second candidate: ddfi does not send the proper instructions to props2csv
3. Third candidates: my whole environment: locally installed Vapoursynth R61, python 3.10.10 and Windows 11