Chatterbox, Resemble AI's production-grade open source TTS model

What’s the compute requirements, or like throughput min and max?