Method __init__

models/dit.py:138–140

(self, num_heads)

class AttentionModule(nn.Module):
    def __init__(self, num_heads):
        super().__init__()
        self.num_heads = num_heads
        
    def forward(self, q, k, v):
        x = flash_attention(q=q, k=k, v=v, num_heads=self.num_heads)