diff options
author | metamuffin <metamuffin@disroot.org> | 2023-09-20 21:40:05 +0200 |
---|---|---|
committer | metamuffin <metamuffin@disroot.org> | 2023-09-20 21:40:05 +0200 |
commit | d0ce1d9134968a15d37135622138f6b8b7667454 (patch) | |
tree | 29e40aa04df7a1b7b48811ab7582b52964af1dc6 /src/embedders | |
parent | 8b65d87640e7367cd88e72c320e4370f4ba471a6 (diff) | |
download | embeddings-sort-d0ce1d9134968a15d37135622138f6b8b7667454.tar embeddings-sort-d0ce1d9134968a15d37135622138f6b8b7667454.tar.bz2 embeddings-sort-d0ce1d9134968a15d37135622138f6b8b7667454.tar.zst |
replace cosinesim with L2-norm euclidean
Diffstat (limited to 'src/embedders')
-rw-r--r-- | src/embedders/vecmetric.rs | 13 |
1 files changed, 6 insertions, 7 deletions
diff --git a/src/embedders/vecmetric.rs b/src/embedders/vecmetric.rs index 0c63911..1bda3a8 100644 --- a/src/embedders/vecmetric.rs +++ b/src/embedders/vecmetric.rs @@ -19,15 +19,14 @@ impl VecMetric for ManhattenDistance {} impl MetricElem for CosineSimilarity { fn dist(&self, other: &Self) -> f64 { - let x = self - .0 + let len_a = self.0.iter().map(|x| x.powi(2)).sum::<f32>().sqrt(); + let len_b = other.0.iter().map(|x| x.powi(2)).sum::<f32>().sqrt(); + self.0 .iter() .zip(other.0.iter()) - .map(|(a, b)| *a * *b) - .sum::<f32>(); - let mag_a = self.0.iter().map(|x| x.powi(2)).sum::<f32>(); - let mag_b = other.0.iter().map(|x| x.powi(2)).sum::<f32>(); - (x / (mag_a * mag_b).sqrt()) as f64 + .map(|(a, b)| (*a / len_a - *b / len_b).powi(2)) + .sum::<f32>() + .sqrt() as f64 } } impl MetricElem for EuclidianDistance { |