I finally made verity fully local work.
I spend 8 hours to run qwen model for verity, whisper model for speech to text and text to speech model to work fully localy on my pc. Its fast and pretty inteligent. Tutorial soon.
My pc stat's:
Laptop rtx 5060 8gb vram - only 8 bit models can run on it.
64 gb tam ddr5 5200mt/s
Ryzen 7 8450 HX
Needed like around 10 GB.