# SLA

## Query execution times

* 3D Face Reconstruction Service: 2.9 sec per request in queue
* 3D Hair Reconstruction Service: 
  * In ***volume*** mode: 39.6 sec per request in queue
  * In ***card*** mode: 9.9 sec per request in queue
* GLB Constructor Service: 4.6 sec per request in queue

## Classifier accuracy

* *gender classification*: **0.961324** for adult faces
* *glasses detection*: **0.984178**