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
No, we only support channel pruning for convolutional layers. For fully-connected layers, you can use SVD to decompose it into two smaller fully-connected layers for compression and acceleration.
Does Pocketflow support channel prune for fully_connect layer ?
The text was updated successfully, but these errors were encountered: