e-means: k-means yang lebih cepat, kini di FAISS 1.16.0

Kiriman publik pada 11 November 2026 menunjuk ke e-means, algoritma pengelompokan yang kini ada di FAISS v1.16.0. Kiriman itu menyebut pencarian dan jawaban AI yang lebih cepat, penggunaan RAM dan VRAM yang lebih rendah, ketersediaan sekarang, serta penggunaan pada server standar dan perangkat keras lokal tanpa peningkatan yang mahal.

Halaman proyek yang independen dan utas pengumuman adalah sumber angka di bawah. Kutipan dikutip secara harfiah.

“I’m excited to open-source e-means: a modern, robust k-means algorithm developed at Meta FAIR. A replacement of Lloyd's algorithm, e-means delivers a ~10× speedup in both epochs and wall-clock time while achieving lower loss.”

Dari pembukaan halaman proyek e-means, e-means.github.io.

“Across a large range of scales and datasets, e-means reaches a test loss below Lloyd's at 25 epochs, in 1.2 to 3 epochs. Measured end to end, that is a 10–45× wall-clock reduction on GPU against the state-of-the-art Lloyd implementation (Flash-KMeans [2]), and 8–15× on CPU against Faiss [1]. Speed and quality usually trade against each other in k-means, while e-means achieves both better speed and loss.”

Dari bagian «An order of magnitude less epochs» di e-means.github.io.

“Introducing e-means, an algorithm beating k-means in 10x fewer epochs. With order of magnitude wall-clock speedups, e-means is released in FAISS and needs no hyperparameter tuning beyond k. It is robust to various scales and datasets.”

Dari kiriman pengumuman Swann Bessa, 10 November 2026.

Swann Bessa dkk., e-means: outperforming k-means (in one-tenth the epochs). Halaman proyek: e-means.github.io. Artikel: arXiv:2611.09501.

Halaman proyek · Artikel di arXiv · Kode di FAISS · Postingan di X · @swann_bessa · Semua berita