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
This would allow for training using a normal model, and then the model can be partitioned, use conv dot products, use I-GELU, etc after training. This would allow for faster training and fix some bugs associated with training a partitioned model