Discofork
Repository brief

facebookresearch/faiss

Read the upstream summary on the left, browse the cached forks below it, and load each fork comparison into the right-hand panel.

Cached analysis
cached 2026-03-30T16:06:28.811Z
1mo ago

facebookresearch/faiss

Faiss is a widely used library for efficient similarity search and clustering of dense vectors. It is active, heavily forked, and actively maintained, with recent commits focused on API updates, validation, SIMD dispatch, and thread-count handling.

GitHub
Loading tags...
Stars39,549
Forks4,311
Default branchmain
Last pushed2026-03-30T03:30:01Z
Recommended shortcuts

Jump straight into Discofork's strongest cached fork picks, or open a compare view in one click.

Forks

Choose a fork to inspect

10 of 10 fork briefs
Selected

Choose this fork if your main goal is QuickerADC-style product-quantization speedups and you can live with a stale, heavily diverged codebase. Choose upstream FAISS if you want current maintenance, broader feature coverage, and lower integration risk.

Prefer this fork if you want the learned adaptive early termination research stack and its reproducibility assets. Prefer upstream if you need an actively maintained, current Faiss for production or integration work.

Prefer this fork only if Windows/cross-platform portability is the primary requirement and you can absorb maintenance debt. For most adopters wanting current Faiss behavior, upstream is the better default because this fork is significantly behind and appears stale.

Prefer this fork only if you need an older, self-contained Faiss snapshot with generated wrappers/docs and are willing to give up current upstream fixes and API evolution. For active development or production use, upstream is the better choice.

Choose this fork only if you need its specific local changes and can afford to own long-term merge and validation work. If you want current Faiss behavior, performance fixes, and API compatibility, upstream is the safer default.

Choose this fork only if the extra CPU/SIMD optimization work is the main requirement and you can absorb ongoing maintenance. If you want current Faiss features, validation fixes, and a lower-risk upgrade path, upstream is the better default.

Choose this fork if your priority is the extra IVF/binary-index functionality and merge-oriented workflows; stick to upstream if you want the latest Faiss fixes, broader compatibility, and less maintenance risk.

Prefer this fork only if you need its specific IO/API and compatibility changes. For most adopters, upstream is the safer default because this fork is materially behind and highly divergent.

Prefer this fork if you need its custom SIMD/quantization/validation work and can absorb long-term merge debt. Prefer upstream if you want the current Faiss API, active maintenance, and the broadest supported workflow surface.

Choose this fork only if its RAFT/CAGRA and benchmark-oriented additions are specifically valuable. If you want the safest, most current Faiss for production or long-term maintenance, upstream is the better default.