Train Faster, Perform Better: Modular Adaptive Training in Over-Parameterized Models

Add code
May 13, 2024

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: