---
license: apache-2.0
tags:
  - crop-disease-detection
  - computer-vision
  - pytorch
  - efficientnet
  - agriculture
  - plant-diseases
datasets:
  - plant-village
  - dhan-shomadhan
  - custom
language: en
---

# Crop Disease Detection Model (EfficientNet-B3)

## Overview

This model detects and classifies crop diseases using computer vision and deep learning. Built on **EfficientNet-B3** and trained on a curated dataset of 13,000+ images, the model can recognize 17 disease classes across **five major crops**:

- **Corn** (Common Rust, Gray Leaf Spot, Northern Leaf Blight, Healthy)  
- **Potato** (Early Blight, Late Blight, Healthy)  
- **Rice** (Brown Spot, Leaf Blast, Neck Blast, Healthy)  
- **Wheat** (Yellow Rust, Brown Rust, Healthy)  
- **Sugarcane** (Red Rot, Bacterial Blight, Healthy)

✅ **Accuracy:** 94.8%  
✅ **Precision:** 95.4%  
✅ **Recall:** 94.5%

The model contributes to **SDG 2 – Zero Hunger**, **SDG 12 – Responsible Consumption**, and **SDG 13 – Climate Action** by enabling early intervention and sustainable agriculture practices.

---

## Model Details

- **Architecture**: EfficientNet-B3 (pretrained on ImageNet)
- **Classifier Head**: Replaced with `Linear(1536 → 17)`
- **Framework**: PyTorch
- **Total Parameters**: ~10.7M
- **Training**:
  - 5-fold cross-validation
  - Early stopping (best at epoch 29)
  - Augmentation & normalization

---

## How to Use

> 💡 This model requires preprocessing consistent with training (image resizing, normalization). For ready-to-use prediction.

### Inference Example (PyTorch)
```python
import torch
from torchvision import transforms
from PIL import Image
import requests
from huggingface_hub import hf_hub_download

# Download the model file from Hugging Face
model_path = hf_hub_download(repo_id="VisionaryQuant/5_Crop_Disease_Detection", filename="best_crop_disease_model.pt")

# Load the model (make sure your architecture matches)
model = torch.load(model_path, map_location=torch.device('cpu'))
model.eval()

# Preprocess input image
image = Image.open("your_crop_image.jpg").convert("RGB")
transform = transforms.Compose([
    transforms.Resize((300, 300)),
    transforms.ToTensor(),
    transforms.Normalize(mean=[0.485, 0.456, 0.406],
                         std=[0.229, 0.224, 0.225])
])
input_tensor = transform(image).unsqueeze(0)

# Run inference
with torch.no_grad():
    logits = model(input_tensor)
    probs = torch.nn.functional.softmax(logits, dim=1)
    predicted_idx = torch.argmax(probs, dim=1).item()

# Map class index to label
idx2label = {0: "Corn___Common_Rust", 1: "Corn___Gray_Leaf_Spot", ..., 16: "Sugarcane___Healthy"}  # Add full mapping
print("Prediction:", idx2label[predicted_class])
```

## Real-World Applications
- Smart Farming: Disease detection via mobile/drones

- Scalable Monitoring: Surveying across large farmlands

- Yield Optimization: Early diagnosis = lower crop loss

## Citation
If you use this model, please cite it as:

**BibTeX:**
```
@misc{5cropdiseasedetection2025,
  title  = {Crop Disease Detection using EfficientNet-B3},
  author = {Abdullahi Olalekan Abdulmumeen},
  year   = {2025},
  url    = {https://huggingface.co/VisionaryQuant/5_Crop_Disease_Detection}
}
```

**APA:**
```
Abdulmumeen, A. O. (2025). Crop disease detection using EfficientNet-B3 [Model]. Hugging Face. https://huggingface.co/VisionaryQuant/5_Crop_Disease_Detection
```

## Contact & Credits
Developed by Abdullahi Olalekan Abdulmumeen <br/>
For the NaijaFarmConsultAI 3MTT Knowledge Showcase project