mlx-decision runs decision models locally on Apple Silicon, up to 6x faster than PyTorch
armbues · reddit · 2026-10-25
An open-source Python package, mlx-decision, runs and fine-tunes decision models (Cloudflare's Clef, Perplexity's pplx-decider, d3, H2O-Lightning, LiquidAI d1) locally on Apple Silicon via MLX.
Key points:
- Decision models take a message/document/screenshot plus typed questions and return probabilities per allowed answer in one pass—useful for ticket triage, sentiment scoring, screenshot QA
- Verified token-for-token against official PyTorch code; typically 1.4–6x faster on MPS
- On M5 Pro: d1-3B at 0.06s per request; pplx-decider-flash (4B) fits a 16GB MacBook Air at 3GB in 4-bit
- Model outputs can train other models: d1-3B fine-tuned from 86.6% to 96.4% on TREC in 4 minutes
- Includes a local web Studio for comparisons, dataset evaluation (accuracy, confusion matrix, calibration), and fine-tuning; Python, CLI, HTTP, and MCP interfaces
- MIT licensed, pip install mlx-decision
More from coding & agent
- Qwen Code Desktop v0.26.1 ships daemon token persistence and durable context compression — github-actions[bot] · 2026-10-25
- Two sandboxed OpenAI agents reportedly exfiltrated data via CPU cache timing side channel — Puzzleheaded-King584 · 2026-10-25
- 76% of PM job postings want AI skills: a visual glossary of 15 agent terms — aakashgupta · 2026-10-25
- Claude-tested bots for weeks, then 219 real players flipped this game's difficulty curve — idlerunner00 · 2026-10-25
- Benchmarking Microsoft's Knowledge Layer: Foundry IQ multi-hop RAG pipeline dissected — adnan_hashmi · 2026-10-25
- Microsoft launches Agentic DevOps Engineer Expert certification via Exam AZ-401 — adnan_hashmi · 2026-10-25