Self-attention-based diarization model. Args: n_speakers (int): Number of speakers in recording in_size (int): Dimension of input feature vector n_heads (int): Number of attention heads n_units (int): Number of units in a self-attention block
(self, n_speakers, in_size, n_heads, n_units, n_layers, dim_feedforward=2048, dropout=0.5, has_pos=False)
source not stored for this graph (policy: none)
nothing calls this directly
no test coverage detected