You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I tried the pretraining model on the NTU60 following the setup instructions. The training loss is highly oscillating, would you please explain this behavior? the loss starts at 0.265 and goes to ~0.24 minimum, I wonder if this is normal? and what checkpoint should we take for the downstream task?
Thank you.
The text was updated successfully, but these errors were encountered:
Hi, thank you for this great work.
I tried the pretraining model on the NTU60 following the setup instructions. The training loss is highly oscillating, would you please explain this behavior? the loss starts at 0.265 and goes to ~0.24 minimum, I wonder if this is normal? and what checkpoint should we take for the downstream task?
Thank you.
The text was updated successfully, but these errors were encountered: