model-optimization
model-optimization copied to clipboard
Is it possible to implement post-training pruning?
MacOS
- TensorFlow version (you are using): 2.3
- Are you willing to contribute it (Yes/No): No
Motivation
The document just shows how to prune the model during training. Each time, we need to call model.fit. During the process, other weights will be modified. We just want to set some weights to zero. In the meantime, don't change other weights.
The feature have been implemented in Pytorch.
Currently we are working on this @fredrec has more details
So pruning doesn't work if we are not able to retrain the model after adding the prune_low_magnitude layer ?
Hello guys, any news on the topic ?