What's next for Rife

[sachaaaaaaaaaaaaaaaaaaaaaaaaaaa]

Hello everyone

Like the rest of the community who follow your work, we can't wait to see the new version of Rife.

As for many I use SVP 4 Pro to see anime more fluid through your technology which is a real revolution even if many do not understand it.

Congratulations on your work with StepFun, it must have changed your mind about doing other things, I hope they give you the means to develop.

As far as I'm concerned, you've got gold in your hands with Rife, and the competition is starting to realize it.

And the proof of the pudding is that others are using your algorithm to improve it, like GMFSS or MultiPassDedup, and others are taking other paths, like MAMBA-VFI or GIMM.

I hope this isn't too difficult for you.

I can't wait to see a new version released, every day I check the repo for the slightest update.

Waiting for an update that will revolutionize everything once again, with a model that doesn't need scene change detection and all in real time maybe.

You should set up a Discord to discuss this with your community, because lots of people admire your work and are trying to push forward the future of image interpolation, which is possible today

Keep up the good work

Sach_Tv

@routineLife1
@TNTwise
@david4958606
@AkarinVS
@98mxr
@GSeanCDAT

[david4958606]

Sorry but I don't know why I have been mentioned in this issue. I am neither a contributor nor a user of RIFE. Sorry again that I cannot offer any assistance.

[hzwer]

Hello, in the past two years, I have rarely kept up with the work in the CV research field and have gradually shifted my focus to LLM research. I'm not sure if I will continue to improve RIFE unless I have a particularly interesting new technology I want to try. Or maybe one day I'll take on a CV-related project or have a new intern.

[netExtra]

Hello, in the past two years, I have rarely kept up with the work in the CV research field and have gradually shifted my focus to LLM research. I'm not sure if I will continue to improve RIFE unless I have a particularly interesting new technology I want to try. Or maybe one day I'll take on a CV-related project or have a new intern.

That's a shame but I understand. Is there some way to actually improve Rife and reduce artifacts? Even if it is at the cost of size and performance, it would be nice to see what could be possible.

[netExtra]

Hello, in the past two years, I have rarely kept up with the work in the CV research field and have gradually shifted my focus to LLM research. I'm not sure if I will continue to improve RIFE unless I have a particularly interesting new technology I want to try. Or maybe one day I'll take on a CV-related project or have a new intern.

By the way. Rife v4.26 does seem to be a little bit better for artifacts than v4.25. I don't remember this being true and my last comment suggests the opposite. The difference is very small and maybe I haven't come across the artifacts I saw previously.