Log in
Register
VAMCompanion

Plugins + Scripts VAMCompanion

Download [<1 MB]
Overview Updates (1) History Discussion
mmn123 said:
Most of it works on the RX 9070 XT, but as VR-user I could run into VRAM-shortages ... I'll write, if I have found good settings ... Only thing which still is only CPU is XTTS ... But there should be a solution (but maybe no totally easy one at the moment. It is only a matter of development time, though, people are working on it).

On both my mobile APU and the RX 9070 XT there was nothing much at release but it gets better every day ...
Click to expand...

Yeah, I'm a VR user, too & a bit wary about the 16Gb VRAM on the 9070XT. It's a sweet gaming card, but my guess would be that in order to run VaM in VR AND run GPU-accelerated nontrivial local models, 24-32Gb would be basically mandatory?

Maybe your setup is simply the best for Radeon users - don't try to make the 9070XT do something that it's not optimized for (and doesn't have the software ecosystem for), and just outsource to a networked secondary PC/laptop's CPU (Probably also easier to get networks to run on CPU?)

Only bummer for me is that I'd need another SSD to set up my old PC - I didn't think buy a spare one before the storage madness started ... :cry:
 
Case said:
Yeah, I'm a VR user, too & a bit wary about the 16Gb VRAM on the 9070XT. It's a sweet gaming card, but my guess would be that in order to run VaM in VR AND run GPU-accelerated nontrivial local models, 24-32Gb would be basically mandatory?

Maybe your setup is simply the best for Radeon users - don't try to make the 9070XT do something that it's not optimized for (and doesn't have the software ecosystem for), and just outsource to a networked secondary PC/laptop's CPU (Probably also easier to get networks to run on CPU?)

Only bummer for me is that I'd need another SSD to set up my old PC - I didn't think buy a spare one before the storage madness started ... :cry:
Click to expand...
Yeah, I can feel the pain, buying a ssd now ... There is one thing I found out: that the 16 GB VRAM were maxed out at first because of some strange AI integration in AMD Adrenalin (which loaded a LLM into the VRAM). Now it doesn't seem to do it anymore. Meaning: Maybe with better control of VRAM usage one could load a small one for inference in VAMcompanion. But it needs further experimentation ...
 
Nice plugin but i got an issue. Root camera show permission denied after I make a request and log spamming like this.
[15:05:49] Continuous agent request failed: 400 Client Error: Bad Request for url: http://127.0.0.1:1234/v1/chat/completions
[15:05:49] RunManager: failed #78 reason=continuous agent failed
[15:05:49] RunManager: finished #78 reason=tracked LM request finished
Camera refresh failed: image "pyimage2129" doesn't exist

Any idea?
 
sckvair said:
Nice plugin but i got an issue. Root camera show permission denied after I make a request and log spamming like this.
[15:05:49] Continuous agent request failed: 400 Client Error: Bad Request for url: http://127.0.0.1:1234/v1/chat/completions
[15:05:49] RunManager: failed #78 reason=continuous agent failed
[15:05:49] RunManager: finished #78 reason=tracked LM request finished
Camera refresh failed: image "pyimage2129" doesn't exist

Any idea?
Click to expand...
Does this happen when you launch the program for the first time, or after you start the server, stop it, and then start it again?
Also, are you using a vision-capable model? If the model does not support vision, an error will occur, and you need to disable the cameras.
 
Last edited:
Dlesser said:
Does this happen when you launch the program for the first time, or after you start the server, stop it, and then start it again?
Also, are you using a vision-capable model? If the model does not support vision, an error will occur, and you need to disable the cameras.
Click to expand...
Oh you right, i need vision model! Sorry about that , I need image-to-text model right?
1784984708553.png
 
Didnt work anywhere. Its working without camera vision so its something with screenshots maybe. Also found this in logs

[16:37:32] camera_2_published_to_request_cache=true; camera_2_published_to_preview_cache_error=[WinError 5] Permission Denied: 'G:\\Installed\\VirtAMate\\AddonPackages\\plugins\\Dlesser Plugins\\VAMCompanion\\camera_preview_cache\\camera_2.tmp' -> 'G:\\Installed\\VirtAMate\\AddonPackages\\plugins\\Dlesser Plugins\\VAMCompanion\\camera_preview_cache\\camera_2.png'
Camera refresh failed: image "pyimage107" doesn't exist

But this is not a spam. Spam is like i show before. Have no idea what is wrong here
 
UPD: fixed with downloaded google models from your screenshot. I downloaded vision model from huggin before and it doesnt work for me.
1784988236050.png


Thanks for your attention! This is very intresting tool <3
 
btw can i disable answering for every screenshots request?

p.s. So iw tested plugin and it looks perspective. Right now there is annoying thing with spamming replies when I even not chatting and also models didnt know well whats going all. So I tested this two models and both of them know a little bit things about environments but dont know what happened. I wear red bikini at my person and ask direct question- "which color of bikini?". And still dont get right answer lol. Maybe Im doing something wrong idk. All cameras show environments and persons. I even make mirror in front of person but still no luck. Anywhere i hope author make improvements and this will be amazing plugin! Good luck!
 
Last edited:
sckvair said:
btw can i disable answering for every screenshots request?

p.s. So iw tested plugin and it looks perspective. Right now there is annoying thing with spamming replies when I even not chatting and also models didnt know well whats going all. So I tested this two models and both of them know a little bit things about environments but dont know what happened. I wear red bikini at my person and ask direct question- "which color of bikini?". And still dont get right answer lol. Maybe Im doing something wrong idk. All cameras show environments and persons. I even make mirror in front of person but still no luck. Anywhere i hope author make improvements and this will be amazing plugin! Good luck!
Click to expand...
I have already noticed an issue where the agent speaks on every single pass. I did not notice it before because the responses were slow, but it became obvious after I loaded a fast model that completes each pass in just a few seconds. I will fix this.
 
sckvair said:
UPD: fixed with downloaded google models from your screenshot. I downloaded vision model from huggin before and it doesnt work for me.
View attachment 609265

Thanks for your attention! This is very intresting tool <3
Click to expand...
Gemma 4 E2B is a very basic model. It will make a lot of mistakes and behave very unintelligently. If possible, I recommend downloading Gemma 4 12B QAT instead.
 
I tried out a bit more with my two computer setup ... Problem: I cannot use an external and physically seperate xtts server in my network. It does work in my browser on my VAM and VAMCompanion machine and I can output test wav files without any trouble. but although I have given VAMCompanion the correct IP and port of my xtts-server there never seem to be any POST or GET requests in the logs of it ...

It was not very fun to set it up for ROCm, but I got it running. Now only the connection is missing ...
 
Last edited:
mmn123 said:
I tried out a bit more with my two computer setup ... Problem: I cannot use an external and physically seperate xtts server in my network. It does work in my browser on my VAM and VAMCompanion machine and I can output test wav files without any trouble. but although I have given VAMCompanion the correct IP and port of my xtts-server there never seem to be any POST or GET requests in the logs of it ...

It was not very fun to set it up for ROCm, but I got it running. Now only the connection is missing ...
Click to expand...
I haven’t added network support for either XTTS or LM Studio in VAMCompanion. It could be implemented so the app connects over IP to LM Studio and an XTTS server running on another computer, similar to OpenClaw or Hermes.

I’m not ready to work on that feature right now because I already have several other scripts in development. The VAMCompanion download includes the full source code. If you’d like to implement this functionality yourself, I can make you a co-author of the resource so you can upload your improved version here.

I also tested an AMD Ryzen AI Max+ 395 system with 128 GB of RAM/VRAM. Setting up ROCm was a pain for me too. Even though I could allocate up to 96 GB as VRAM, I eventually sold the machine and switched back to NVIDIA GPUs.
 
Dlesser said:
I haven’t added network support for either XTTS or LM Studio in VAMCompanion. It could be implemented so the app connects over IP to LM Studio and an XTTS server running on another computer, similar to OpenClaw or Hermes.

I’m not ready to work on that feature right now because I already have several other scripts in development. The VAMCompanion download includes the full source code. If you’d like to implement this functionality yourself, I can make you a co-author of the resource so you can upload your improved version here.

I also tested an AMD Ryzen AI Max+ 395 system with 128 GB of RAM/VRAM. Setting up ROCm was a pain for me too. Even though I could allocate up to 96 GB as VRAM, I eventually sold the machine and switched back to NVIDIA GPUs.
Click to expand...
Ah okay, then the reason is clear and thank you, also for the other plugins, too. I will have a look in the source code and maybe this could be a project, where it would be interesting for me to put some serious time into it (after all many useful skills intersect here with the aesthetics of -ahem - an human biologic activity we all apreciate for its sheer vital necessity - cough, cough. In short: It could be cool.

But I am still a beginner ... If some testing and adding functionality should lead to more confidence and maybe a working solution, I'll write here ...
 
mmn123 said:
Ah okay, then the reason is clear and thank you, also for the other plugins, too. I will have a look in the source code and maybe this could be a project, where it would be interesting for me to put some serious time into it (after all many useful skills intersect here with the aesthetics of -ahem - an human biologic activity we all apreciate for its sheer vital necessity - cough, cough. In short: It could be cool.

But I am still a beginner ... If some testing and adding functionality should lead to more confidence and maybe a working solution, I'll write here ...
Click to expand...
Right now, I’m working on a much bigger problem: getting the model to stop talking on every processing cycle and behave naturally, like a real person. That’s much more complicated than adding support for a remote server.

If you’re a beginner, I recommend using Codex or Claude Code. They can help you install the necessary software and packages, set up a virtual environment, write code, and handle many other development tasks.
 
Dlesser said:
Right now, I’m working on a much bigger problem: getting the model to stop talking on every processing cycle and behave naturally, like a real person. That’s much more complicated than adding support for a remote server.

If you’re a beginner, I recommend using Codex or Claude Code. They can help you install the necessary software and packages, set up a virtual environment, write code, and handle many other development ta
Click to expand...
Sounds good, I will try that (as a good start into plugin development).
 
You must log in or register to reply here.
Back
Top Bottom