This is a small set of PyTorch-based CPU and GPU friendly implementations of common ranking metrics.
Included metrics:
- Precision/Recall
- Average Precision
- Discounted Cumulative Gain
- Normalized Discounted Cumulative Gain
Implementations are tested against scikit-learn
with hypothesis
property-based testing (where possible.)