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
Think about how to apply the algorithm to Layernorm.
Why
Parameterize alpha of SmoothQuant.
Current implementation fixes alpha value of SmoothQuant. But, best alpha may vary depending on the models. Let's parameterize it. Users would be able to set their best alpha then.
Think about how to apply the algorithm to Layernorm.
Layernorm decomposes into Mul and Add. Therefore, if the circle to apply is optimized first, such mul and add are fused into other ops, which makes it impossible to apply SmoothQuant. Therefore, we need to consider the order of the sequences.
The text was updated successfully, but these errors were encountered:
What
Let's support more features for
fm-equalize
.Why
Current implementation fixes alpha value of SmoothQuant. But, best alpha may vary depending on the models. Let's parameterize it. Users would be able to set their best alpha then.
Layernorm decomposes into Mul and Add. Therefore, if the circle to apply is optimized first, such mul and add are fused into other ops, which makes it impossible to apply SmoothQuant. Therefore, we need to consider the order of the sequences.
The text was updated successfully, but these errors were encountered: