Question about framework

[diimdeep]

Could someone share more info

• model sizes / quality
• realtime factor
• is it using VAD? meaning audio with a lot of silence processed faster
• is it using only GPU or APU ?