Official source code of the INTERSPEECH 2023 paper: "Audio-Visual Speech Separation in Noisy Environments with a Lightweight Iterative Model" (AVLIT) - View it on GitHub