Skip to content

Latest commit

 

History

History
13 lines (8 loc) · 545 Bytes

README.md

File metadata and controls

13 lines (8 loc) · 545 Bytes

Emotion in gaze trainer

  • Author: Andrei Alexandru
  • Faculty: Faculty of Electronics, Telecommunications and Information Technology
  • Master program: TAID

This code implements a trainer for a multi-task network which performs gaze estimation and emotion classification. It was developed for my Dissertation project.

The hyperparams for an experiment are available and can be modified using the config from configs/efficientnet_train.yaml. After the desired params are set, one has to run script python tools/train.py.