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
When using Keras 3.8.0/3.7.0 and tensorflow 2.18.0 to train on multiple gpus using MirroredStrategy, you typically get a NANs after a few training steps. Downgrading to Keras 3.6.0 fixes this. Maybe there is a bug in these versions? Here is a minimal example I posted on stack overflow a while back:
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
When using Keras 3.8.0/3.7.0 and tensorflow 2.18.0 to train on multiple gpus using MirroredStrategy, you typically get a NANs after a few training steps. Downgrading to Keras 3.6.0 fixes this. Maybe there is a bug in these versions? Here is a minimal example I posted on stack overflow a while back:
https://stackoverflow.com/questions/79285532/multi-gpu-training-in-tensorflow-results-in-nans
Beta Was this translation helpful? Give feedback.
All reactions