Are VLMs only good for free-text? Structured extraction support is lacking
tenkei_01 · reddit · 2026-10-24
A developer working VLMs into CV pipelines notes most models lack structures for classification and extraction/attribution tasks — many don't even offer JSON schema or structured responses. He asks whether most people use VLMs purely for generation, and how others handle structured output gaps in production.
More from Models
- Scale CEO teases new models as users say Muse Wonder Full is underrated — alexandr_wang · 2026-10-24
- Canva Craft trained on millions of templates and 25,000 designer-evaluated image pairs — jacob_top · 2026-10-24
- $34.72 of Kimi K3.1 Pro tokens builds a playable Spider-Man Manhattan in 8 hours — sven_ai · 2026-10-24
- Cisco engineer replicates free Antares security model at ~80% of official score — shashib · 2026-10-24
- Kimi K3.1 vs Opus 5.5: 80% Cheaper and 4x Faster on Same Task — sven_ai · 2026-10-24
- Claim: Open-source Kimi K3.1 Pro beats OpenAI's GPT-6-Astra, unverified — sven_ai · 2026-10-24