Open-source Kev-0.5B: a tiny Qwen2.5-0.5B decision model you can train on a MacBook
sull · x · 2026-09-19
jaredpalmer released Kev-0.5B, a tiny Jev-like decision model built on Qwen2.5-0.5B, with weights on GitHub (121 stars already).
- LoRA adapter plus a readout head on Qwen2.5-0.5B, exposed via a TypeSafe-compatible API
- Packs a document and many typed questions into one sequence; a block-causal mask lets each question attend to the document but not other questions, so all calibrated probabilities come out in a single prefill pass with no decoding
- A pointer head scores each question's options and softmaxes them; the head is trained with cross-entropy against labeled outcomes, so probabilities are learned rather than generated
- Trainable and runnable on a MacBook Pro
More from coding & agent
- Swarms Cloud adds full observability for every agent API execution — KyeGomezB · 2026-09-19
- Microsoft devs launch self-guided GitHub Copilot workshop: local repo to merged PR in 60-90 minutes — 0xkarasy · 2026-09-19
- Groovy's Updated AI Tutorial Covers Ollama4j, Spring AI, Embabel and Micronaut on JDK 25 — therealdanvega · 2026-09-19
- Jev, a typed-reasoning model that answers not writes, hits Venice API beta — 0xAllen_ · 2026-09-19
- Link launches agent wallet for AI shopping in Canada, with per-purchase approvals — jeff_weinstein · 2026-09-19
- SAM demo: secure cross-device agent mesh in 30 seconds, phone sensors become MCP tools — rakyll · 2026-09-19