A repo of small vision models built for hacking, not chasing SOTA
mervenoyann · x · 2026-11-18
Hugging Face's Merve Noyan highlighted a vision model repository whose philosophy is explicitly not training the next SOTA model — it's for getting started with vision, training your first object detector, or equipping your agent with small task-specific models. The author says the pipeline below has been tested across different domains and works well, and mentions combining it with Qwen3.5; limitations will be shown at an upcoming talk.
More from coding & agent
- dottxt tool-call index shows LFM3-1.7B broken tool calls jump from 7.19% to 15.96% across providers — remilouf · 2026-11-18
- Pocket Folio adds API key + CLI so AI agents can auto-fill your card binders — cnakazawa · 2026-11-18
- Merge Adds Triggers to Agent Handler: Agents Fire Automatically on Emails and PRs — shensi · 2026-11-18
- Frontier models race jets through the Top Gun canyon: Astra wins, Opus refuses to fly — abhi3188 · 2026-11-18
- Running a literature research project inside Grok bot, from question to grant aim — MaxUnfried · 2026-11-18
- Incredibuild spins up Islo, an enterprise 'software factory' for AI agents, with $15M from Insight Partners — Shruti_0810 · 2026-11-18