Can Edge-Deployable Vision-Language Models Identify Species?
Evaluation of 2-8B VLMs (Qwen3-VL, Gemma3) against BioCLIP on camera-trap species ID shows all models degrade sharply on field imagery.
The study tests whether edge-deployable 2-8B vision-language models carry genuine taxonomic knowledge, comparing Qwen3-VL 2B/4B/8B and Gemma3 4B against the 300M specialist BioCLIP on a 96-species task across clean iNaturalist photos and six LILA.science camera-trap collections. All models degrade 9.6-26.6 percentage points on field imagery, and BioCLIP outperforms every VLM by 33.2-59.2 points on an expanded 200-image sample, suggesting specialized data rather than scale drives the gap. Under open-set prompting, 5.9-9.6% of responses are syntactically valid but taxonomically nonexistent species names, with fabrication rankings replicating across evaluation sets.
- All tested models degrade 9.6-26.6 points between clean photos and camera-trap imagery.
- BioCLIP (300M) beats every 2-8B generalist VLM by 33.2-59.2 percentage points.
- Degradation reflects image legibility shifts, since BioCLIP's domain gap matches the best VLM's.
- 5.9-9.6% of open-set responses are fabricated but taxonomically plausible species names.
- Organizations
- LILA.science
From the original article
1 min read at arxiv.orgCamera traps often run in the field on edge hardware with limited or no connectivity, making small, locally-deployable vision-language models (VLMs) -- not frontier-scale ones -- the practically relevant class to evaluate for species identification.
By William Zhou · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.