Auto-optimized TF-MoDISco is 5-15x faster with byte-identical output
jmschreiber91 · x · 2026-10-14
An auto-optimize-produced PR parallelizes the TF-MoDISco motifs pipeline across processes, rewrites numba kernels in background processes, and reworks clustering, merging and polishing stages. Validated on 4,972 jobs (ISM, DeepLIFT, IG, gradients × 60 models + malformed inputs), it delivers 5-15x speedups and 2-3x memory savings; a 50,000-seqlet benchmark dropped from 1745s to 119.1s with byte-identical output (leidenalg pinned at 0.11.0).
More from coding & agent
- Agent Teams Cost Up to 5.1x More but Barely Beat Single Agents, Vals AI Finds — josh_wills · 2026-10-15
- BaitBench: frontier agents take data-leaky shortcuts in 57% of runs and mostly know they're cheating — gleech · 2026-10-15
- Claude Opus spots bcrypt-per-request API auth behind VPS CPU spikes, load drops 0.9 to 0.4 — DanielLockyer · 2026-10-15
- Burkov: labs are now training coding LLMs on git history, exactly as he predicted 9 months ago — burkov · 2026-10-15
- Cohere research: swarms of small models reach frontier-level software engineering — Cohere_Labs · 2026-10-15
- 12 GitHub repos to supercharge Claude Code: context, parallel sessions, automation — Roger_M_Taylor · 2026-10-15