search

talosh / flameTimewarpML

Flame Machine Learning Timewarp. based on arXiv2020-RIFE
MIT License
34 stars 9 forks source link

fails to render after fresh install #96

Open avlaanderen opened 2 months ago

avlaanderen commented 2 months ago

Hey

We're using flameTimeWarpML (0.4.4 build 15) in a shared env. We run the installer from the shared folder, and all installs well but...

It pip's the latest release of Numpy (2.0) and 1.0 is required. Lucky for me, I have a few systems running, and working, so copied the /home/flame/flameTimewarpML/miniconda3 folder over to the non-working version, and voila... we render.

Also, I find the manual install confusing... preferably we do it manually in a shared env of course. If source and destination is specified more explicit (or explained) it might be easier....

OS: Rocky Linux 8.7 DKU: DKU_18.2.2 Flame: 2023.3.2, 2024.2.2

Thanks!

alatteri commented 2 months ago

Use .0.4.5, its way easier. You just drop the package in a central defined hooks folder. No need to initialize all that stuff.

talosh commented 2 months ago

I've started using v0.4.5 dev 003 in production at work and though I stumble upon some little issues here and there to be fixed it generally gives me a better result, so I would say yes pleas try it and it likely to be an update soon in order to fix some minor issues