Fine-tuning Qwen2.5 0.5B on 5,000+ personal WhatsApp messages to clone your texting style on a MacBook
alexcovo_eth · x · 2026-11-13
A developer fine-tuned Qwen2.5 0.5B on 5,000+ of their own WhatsApp messages so the small model mimics their texting style — all on a MacBook.
Recipe: 4-bit quantization, LoRA fine-tuning with mlx-lm on Apple Silicon. A cheap, reproducible way to build a personal "digital twin" for messaging.
More from coding & agent
- Hermes plugin hackathon drew 1508 plugins and 5985 agent tools in one week — Teknium · 2026-11-13
- Nous Research's Hermes App Ships an Official TouchDesigner Plugin for Agents — RileyRalmuto · 2026-11-13
- STOC 2027 Posts All 1,665 Submissions; AI Formalizes 1,496 into Lean 4 Statements — ChenhaoTan · 2026-11-13
- User leaves Fable 5.2 running overnight on Blender and Unreal, wakes up to finished game environment assets — FinanceYF5 · 2026-11-13
- New research tackles the agent experience problem: knowledge trapped in weights — khademinori · 2026-11-13
- YC F26's Acyclic Labs open-sources infra to run 10,000-agent swarms in one session — ycombinator · 2026-11-13