How walid-mk/qiskitfallfest_emsi_quantumafrica works
This repository is a hands-on crop-classification challenge: participants explore supplied satellite-feature data, prepare a training-only feature view, compare classical and quantum/hybrid models, evaluate validation predictions, and submit a fitted model. The sampled Python modules implement checked data loading, model interfaces, a reference SVM, metrics, reproducible training-size experiments, and opt-in IBM inference. The README also documents the notebook-led exploration and saved result comparisons. Notebook internals were not sampled, so those links reflect documented capabilities rather than inspected calls.
Main entry points: Qiskit_QuantumAfrica_Crop_Challenge.ipynb Challenge notebook · coverage.py Training-size experiments · models.py QNN and classical control · data.py Data loader checks
Components 14
Challenge workflow
- Challenge notebook
notebooks/Qiskit_QuantumAfrica_Crop_Challenge.ipynb
Data and preparation
- Crop data release
data - Data loader checks
challenge/data.py - Feature preparation
notebooks/Qiskit_QuantumAfrica_Crop_Challenge.ipynb - PCA feature views
data/views
Modeling
- RBF SVM reference
challenge/classical.py - QNN and classical control
challenge/models.py - Submission interface
challenge/submission.py
Evaluation and inference
- Training-size experiments
challenge/coverage.py - Classification metrics
challenge/metrics.py - Saved predictions and runs
results - Opt-in hardware inference
challenge/hardware.py
Other components
- Participant
- IBM quantum backend
Connections 15
- Participant → Challenge notebook: works through
README.md - Challenge notebook → Crop data release: explores
README.md - Data loader checks → Crop data release: loads and validates
challenge/data.py - Challenge notebook → Feature preparation: prepares features
README.md - Feature preparation → PCA feature views: uses prepared views
README.md - Training-size experiments → Data loader checks: loads training data
challenge/coverage.py - Training-size experiments → RBF SVM reference: fits reference
challenge/coverage.py - Training-size experiments → QNN and classical control: fits alternatives
challenge/coverage.py - Submission interface → QNN and classical control: inherits classifier
challenge/submission.py - Training-size experiments → Classification metrics: scores predictions
challenge/coverage.py - QNN and classical control → Classification metrics: validates probabilities
challenge/models.py - Challenge notebook → Saved predictions and runs: compares saved results
README.md - Challenge notebook → Submission interface: documents interface
README.md - QNN and classical control → Opt-in hardware inference: provides trained circuit
- Opt-in hardware inference → IBM quantum backend: runs inference
challenge/hardware.py