visionjo / facerec-bias-bfw

Source code and notebooks to reproduce experiments and benchmarks on Bias Faces in the Wild (BFW).
BSD 3-Clause "New" or "Revised" License
45 stars 9 forks source link

Five fold experiments (accuracy) #18

Closed visionjo closed 4 years ago

visionjo commented 4 years ago

Created script to perform a 5-fold experiment based on accuracy.

Metrics class was implemented, which provides an interface for confusion stats, and calculations based on these stats.

Updated setup.py

Added utils to find_best_thresholds with interface that allows for scoring function and comparison operating to be specified as optional input arguments

Fold 1: 0.9567425341620287 Fold 2: 0.9580187810716259 Fold 3: 0.9558251402784467 Fold 4: 0.9585394632863159 Fold 5: 0.9525168810245931

---

Avg. Acc: 0.9563