UMAP 0.6 docs: how to embed million-scale datasets, tuned via 3M arXiv papers
leland_mcinnes · x · 2026-11-11
UMAP 0.6 ships a large-scale tutorial embedding all 3.07M arXiv papers (title+abstract embedded as 4096-dim vectors by F2LLM-v2 8B). It covers the parts that matter at million scale: nearest-neighbor search, embedding initialization, the layout optimizer, repulsionstrength, and hard negative sampling — and why the new pipeline (compatibilitylayout=False) suits very large datasets. Timings on a 64-core, 500GB RAM machine.
Related event: UMAP 0.6 released with major speedups for million-scale datasets(5 posts)→
More from Research
- Mobile AI Factory: agentic pipeline builds on-device SLMs that beat 5-shot frontier models — kalyan_kpl · 2026-11-11
- Amazon paper: multi-agent setups burn 1.4-4.7x tokens for small gains; single agent is the default — rohanpaul_ai · 2026-11-11
- MIT's Buehler builds "recursive meta-intelligence": hundreds of AI agents inhabit self-created worlds to study material failure — ProfBuehlerMIT · 2026-11-11
- Explosion launches Ellf directory: 52 NLP templates plus multi-agent annotation workflow — _inesmontani · 2026-11-11
- STABLE Optimizer Accepted at NeurIPS 2026: Taming Continual Learning Forgetting at the Optimizer Level — v_lomonaco · 2026-11-11
- From Bahdanau to FlashAttention: a timeline of attention's 8-year evolution — cneuralnetwork · 2026-11-11