PlantVLM-Edge enables quantized vision-language models for on-device crop diagnosis
Original title: PlantVLM-Edge: Quantized Vision-Language Models for On-Device Crop Health Diagnosis
4-bit quantization retains 98.6% answer quality while reducing memory by 49%, running in 0.12s for first token on edge devices.
PlantVLM-Edge is a quantized VLM framework for conversational crop health diagnosis on resource-constrained devices. It combines domain adaptation, post-training quantization, and offline serving. Validated on Qwen3-VL-2B and LLaVA-1.5-7B with 4-bit AWQ, achieving near full-precision quality with low memory usage of 2.01 GiB.
- 4-bit AWQ retains 98.6% of full-precision answer quality
- reduces weight memory of 2B backbone by 49%
- first token in about 0.12 s
- completes 64-token response in 0.78 s
- AI models
- Qwen3-VL-2BLLaVA-1.5-7B
From the original article
1 min read at arxiv.orgTimely and accurate crop health diagnosis is essential for sustainable agricultural production, yet accurate diagnosis often requires expert knowledge that is not readily available. Recent vision-language models (VLMs) have demonstrated remarkable capabilities in visual understanding, reasoning, and natural-language interaction, creating new opportunities for intelligent crop health assessment.
By Gaurab Wagle · 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.