Cua-S1 Form Model: Repo, Weights and Dataset Links Released
aigclink · x · 2026-09-21
Follow-up post providing the three entry points for Cua's cua-s1-forms: the GitHub repo (cua-s1 under trycua/cua, with model, synthetic-data, training and eval code), model weights, and dataset. Per the repo, Cua-S1 studies small specialist computer-use models; the first checkpoint cua-s1-form-v0 targets form-oriented UI tasks and is early-stage research — the source-only release ships no weights, datasets or binaries, which must be downloaded separately.
Related event: Ex-OpenAI Researcher Open-Sources 706K-Param Form-Filling Model(2 posts)→
More from Research
- What AI means for mathematicians: seven predictions extrapolated from software — cgarciae88 · 2026-09-21
- Stanford RL method fixes VLA latency, lifting robot success from 42% to 97% with 10 minutes of data — burny_tech · 2026-09-21
- Dev Trains Continual-Learning Model From Scratch on 8GB VRAM Laptop, Now at 530M Params — Another__one · 2026-09-21
- Prove2Me: Claude agents wrote 13M lines of Lean in 11 days to formalize Fermat's Last Theorem — liuzhuang1234 · 2026-09-21
- ICLR Review Policy Raises Concerns: 3 Paper Credits Means Mandatory Reviewing — Striking-Warning9533 · 2026-09-21
- EvoOntology: a self-evolving ontology layer bridges the agent-data gap — RUC-DataLab · 2026-09-21