Optimizing Pretrained Convolutional Neural Networks for Tomato Leaf Disease Detection

Table 2

Distribution of the field-based dataset for training, validation, and evaluation.

Type of tomato leaf disease	Training	Validation	Testing
Healthy	2790	796	398
Late blight	3393	968	485
Septoria leafspot	403	115	58
Bacterial speck	336	96	48
Bacterial spot	739	211	106
Early blight	2991	854	427