cvlab-stonybrook / Scanpath_Prediction

Predicting Goal-directed Human Attention Using Inverse Reinforcement Learning (CVPR2020)
MIT License
103 stars 22 forks source link

how to test trained models to generate scanpath? #19

Closed zhangkp1997 closed 2 years ago

ouyangzhibo commented 2 years ago

For now, you may evaluate the model using the validation set. Also, feel free to contact Greg (gregory.zelinsky@stonybrook.edu) to request the access to the test set.