Hugging Face releases Smol2Operator: a full recipe for training SmolVLM2-2.2B into a GUI agent
Original title:Smol2Operator: Post-Training GUI Agents for Computer Use
AI summary
Hugging Face released Smol2Operator, which uses two-stage SFT to train SmolVLM2-2.2B-Instruct, a model without GUI grounding ability, into a GUI agent, raising ScreenSpot-v2 accuracy from about 0% to 41% and then to 61%.
Key Takeaways
- Hugging Face released Smol2Operator,which uses two-stage SFT to train SmolVLM2-2.
- 2B-Instruct
- a model without GUI grounding ability,into a GUI agent,raising ScreenSpot-v2 accuracy from about…
Quick Facts
- Published
- Updated
- Source
- Hugging Face 社区博客(混合发现)
- Category
- Papers & Research
- AI score
- 62
原文给出从数据统一到两阶段微调的完整训练配方和结果数字,且全部开源,可复现到自己的 GUI Agent 项目中。
Related entities
Source & copyright
This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.
How to Cite
When citing this article, keep the title, date, and permalink, and credit Orange News as the source.
Orange News: Hugging Face releases Smol2Operator: a full recipe for training SmolVLM2-2.2B into a GUI agent. September 23, 2025. https://new.nat6.net/en/news/bkrtf15iggxqduorkkxnrfjls/