extendes attention algorithm with masking

JakubSchwenkbeck · Dec 7, 2024 · f80fa3a · f80fa3a
1 parent e3c8aff
commit f80fa3a
Showing 1 changed file with 0 additions and 1 deletion.
diff --git a/src/attention/scaled_dot_attention.rs b/src/attention/scaled_dot_attention.rs
@@ -11,7 +11,6 @@ pub fn scaled_dot_product_attention(
 ) -> Array3<f32> {
     let scores = scaled_dot_product(q, k, v.clone(), mask);
     let sm_scores = softmax_3d(&scores);
-    // TODO implement masking
     tensor_product(&sm_scores, &v)
 }