How imskr/plant_disease_detection works
Plant Disease Detector is a web app for classifying uploaded plant-leaf images. A user selects an image in the browser, submits it for analysis, and sees the predicted disease class. The sampled server code loads a FastAI learner and uses it for inference; the README documents a broad set of plant and disease classes. Training notebooks and deployment guides are present, but are not part of the runtime workflow described here. The graph reflects the sampled Starlette app; the separate Flask example is omitted.
Main entry points: server.py Starlette server · index.html Upload page · client.js Upload client · style.css Page styles
Components 8
Browser experience
- Upload page
app/view/index.html - Upload client
app/static/client.js - Page styles
app/static/style.css
Inference service
- Starlette server
app/server.py - Image prediction
app/server.py
Model assets
- ResNet34 model
app/models/export_resnet34_model.pkl
Other components
- User
- Model download
Connections 9
- User → Upload page: opens
- Starlette server → Upload page: serves
- Upload page → Upload client: loads
- Upload page → Page styles: loads
- Upload client → Starlette server: posts image
- Starlette server → Image prediction: dispatches
- Image prediction → ResNet34 model: uses learner
- Starlette server → Upload client: returns result
- Model download → Starlette server: downloads if absent