A suggestion: I'd swap llava for Florence-2 for your open set text description. ... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

rocauc on Oct 8, 2024 | parent | context | favorite | on: Video Surveillance with YOLO+llava

A suggestion: I'd swap llava for Florence-2 for your open set text description. Florence-2 seems uniformly more descriptive in its outputs.

Eisenstein on Oct 8, 2024 | [–]

They are using Ollama which is based on llama.cpp; florence is not supported on that backend.

jerpint on Oct 8, 2024 | [–]

I found grounding-dino better than Florence and faster

netdur on Oct 8, 2024 | [–]

I found YOLOS to be faster and better, bot real time but 22k objects under half second

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact