Skip to content

A simple bootstrap for training custom models on AffectNet dataset on a system with multiple GPUs

License

Notifications You must be signed in to change notification settings

ali-sedaghi/AffectNet-Trainer

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

AffectNet-Trainer

A simple bootstrap for training custom models on AffectNet dataset on a system with multiple GPUs using Mirrored Distributed Strategy.

Dataset

For downloading AffectNet, only Lab Managers, Professors or Students using their academic email account should fill out the request form HERE.

Dataset must be inside data folder of the project.

Adding Models

For adding custom keras models, add a new python file inside models folder and use it inside main.py.

Keras callbacks will save weights and logs inside checkpoints folder.

About

A simple bootstrap for training custom models on AffectNet dataset on a system with multiple GPUs

Resources

License

Stars

Watchers

Forks

Languages