VectraYX-Vision-1B: Open-Sourced Sub-2B Spanish Cybersecurity VLM
Juan S. Santillana · hf · 2026-08-12
Released VectraYX-Vision-1B, a sub-2B parameter Spanish/LATAM cybersecurity vision-language model.
- Architecture: Couples a frozen SigLIP encoder to a Spanish decoder via an MLP.
- Findings: Introduces a NoPE (No Positional Encoding) ablation for visual attention. Despite functional pipelines, the model reports near-zero visual grounding capability.
- Availability: Open-source weights are available, along with planned remediation steps.
More from Models
- Nous Research Teases Next Free Model Release, Asks Community for Picks — NousResearch · 2026-08-12
- ChatGPT o3 Glitching: Users Report Incomplete Replies and Cutoff Outputs — Aine_123 · 2026-08-12
- Study Finds No Reasoning Distillation Traces in Kimi K3, Cites Data Contamination — bookwormengr · 2026-08-12
- 3B-Param OCR Model Runs Entirely in Your Browser via Rust & WASM — doodlestein · 2026-08-12
- Community Urges Arena to Overhaul Image and Video Model Benchmarks — PAstynome · 2026-08-12
- Qwen 27B Beats 35B in Coding on 32GB GPUs, But Local Models Still Can't Self-Check — WSTangoDelta · 2026-08-12