Power-Guided Grouped SRU for Real-Time Causal Audio-Visual Speech Separation - View it on GitHub