Tomadoc - Tomato Disease Classifier
This model is a fine-tuned version of google/vit-base-patch16-224 optimized for identifying diseases in tomato plant leaves. It was trained to serve as the backbone core for the Tomadoc application.
Model Description
- Model Type: Vision Transformer (ViT)
- Task: Image Classification (10 Tomato classes)
- Base Model: google/vit-base-patch16-224
- Dataset: PlantVillage (Tomato leaf subset containing healthy and diseased leaves)
Training & Hardware Details
- Hardware Used: NVIDIA T4 GPU (Google Colab)
- Precision: Mixed Precision (FP16)
- Learning Rate: 2e-5
- Batch Size: 32
- Epochs: 5
Intended Uses & Limitations
This model is intended to classify leaf images of tomato plants into their respective healthy or diseased categories. It performs best when provided with clear, centered images of single leaves under decent lighting.
Evaluation Results
The model was evaluated on a dedicated test split from the filtered dataset:
- Final Test Accuracy: 0.9961
