Tomadoc Advanced Tomato Ripeness Classifier
This model is a fine-tuned version of google/vit-base-patch16-224 designed to identify the exact ripening stage of a tomato. It was trained on the extensive FreshCheck dataset to provide highly accurate agricultural insights for the Tomadoc application.
Model Description
- Model Type: Vision Transformer (ViT)
- Task: Image Classification
- Classes: Unripe, Ripe, Overripe, Damaged
- Base Model: google/vit-base-patch16-224
Training & Hardware Details
- Dataset: FreshCheck Fruit Ripeness Classification Dataset
- Hardware Used: NVIDIA T4 GPU
- Precision: Mixed Precision (FP16)
- Evaluation Strategy: Epoch-based
- Learning Rate: 2e-5
- Epochs: 5
Evaluation Results
- Final Test Accuracy: 0.9983
- Downloads last month
- 5