Model card for vit_small_plus_patch16_dinov3.ft_plantdoc_384

Overview

This model classifies diseases from plant images.

• Dataset size: 8000 images
• Number of classes: 39
• Architecture: DINOv3 ViT Small (384)

Metrics

• mAP: 0.91
• Accuracy: 0.83

Model

Model Details

• Model Type: Image classification
• Backbone: vit_small_plus_patch16_dinov3.lvd1689m
• Model Stats:
  • Params (M): 28.7
  • GMACs: 8.1
  • Activations (M): 21.8
  • Image size: 384 x 384
• Papers:
  • DINOv3: https://arxiv.org/abs/2508.10104
  • An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale: https://arxiv.org/abs/2010.11929v2- Original: https://github.com/facebookresearch/dinov3
• Dataset:
  • PlantDoc (8000 images)
• Pretrain Dataset: LVD-1689

Model Usage

Built with:

import torch
import timm

# create model
model = timm.create_model(
    "vit_small_plus_patch16_dinov3.lvd1689m",
    pretrained=False,
    num_classes=39,
)

# load weights
state_dict = torch.load("model.bin", map_location="cpu")
model.load_state_dict(state_dict)