Can be used with what Program/tool?
The models are all pytorch, onnx, and ncnn, so you can run them in command line or you can import them into a GUI that supports them.
can you name a GUI tool that you recommend that support those modules and are processing videos?
The tool that was mentioned in the post (chaiNNer). SVFI also supports importing models.
The Open Model DB site suggests VSGAN-TensorRT-docker, enhancr, ComfyUI, and AnimeJaNaiConverterGui (onnx only) in addition to their GUI.
what modules you tested and found to do well (from OpenModelDB)?
Stay away from enhancr.
The author dumped the project a year back.
Just use chainner.
EDIT:
Seems a good alternative.
I was impressed with TNTwise, it’s very simple and easy to use (and was designed on Linux ).
I also noticed on the repo that one of their featured models is based on Proteus. I wonder how it compares to the Topaz version of Proteus. OpenModelDB has a model that’s based on Gaia HQ and trained for film.
@Akila I tried several of the VHS and film models. There’s a 2x model called VHS upscale and Denoise Film that did a pretty good job.
I have UniFab, upscale it’s not good I don’t use it and HDR conversion gives me destroyed vids! I’m a old DVDFab ripp/clone user, this core product from them is good/ok, but Unifab ist not
so I tried this module as soon as you published the OpenDB site and for me the results where unacceptable, it added so much noise. that is why I asked you if you have good modules.
Or maybe you know how to disable/reduce the noise it adds.
If you’re using chaiNNer, you can add a denoise node to the chain.
This is a great way to make people lose respect for you, guys. No one likes unscrupulous companies.
Take a lesson from all of the antitrust cases going on right now.
For my part I imported models of open model db in svfi and its fails, it does not work and they say to click on risk mode in advanced options and I can’t find it anywhere. How make open model db models work on svfi if anyone knows thank you
I followed the instructions and it worked perfectly fine.
SVFI only supports adding Compact, SPAN, ATD, and ONNX models.
How did you proceed to make it work? I created the Compact folder with the Models subfolder and I put the pytorch models but it does this to me all the time…And I don’t know where to click the risk mode.
This model don’t work 4x LexicaDAT2_otf - OpenModelDB
That’s not a Compact model. Nor is it ATD, SPAN, or ONNX.
You just choose the model and launch the encoding.
Hey all, just wanted to revive this thread.
Is there a strong contender for TVAI in OpenModelDB?
Is it worth setting up a GUI and trying out some of those models?
As another contribution, I recently found this:
The free version only allows for 10 sec videos and from what I was able to test, it does a heavy AI re-generation of the video. Not ideal for what many of us se TVAI for but could still be worth following the progress of this tool.
yes but often it is just online by uploading it. I will definitely not upload my stuff anywhere and it’s not practical if you have large amounts of data.
Krea is amazing and if you’re looking for one place to do all creative image and video generation & editing then it’s the best solution out there.
It’s nowhere near what Topaz is doing for existing images and videos though, and for now just not in the same marketplace.
Like 99% of TVAI users here we don’t want our footage uploaded to a remote server to be processed.
I’m saying it’s best for creative generation aka text-to-image and text-to-video.