search

he-y / soft-filter-pruning

Soft Filter Pruning for Accelerating Deep Convolutional Neural Networks
https://arxiv.org/abs/1808.06866
374 stars 74 forks source link

How does this method accelerate training and predicting? #12

Closed jimmyliu2013 closed 5 years ago

jimmyliu2013 commented 5 years ago

Thank you for your excellent paper. But I don't understand why setting filters to zero can make it faster, since you still need to do the matrix multiplication. Could you please give me a brief explanation? Thank you very much!

he-y commented 5 years ago

Please refer to the released pruned model and instruction for the model. You can find the construction of pruned model here.

jimmyliu2013 commented 5 years ago

Thank you for your time. So we don't accelerate the training process, but after the training is finished, we remove all the zero-filters from the model so that the model is smaller and inference is faster. Am I right?

he-y commented 5 years ago

Yes, our paper is about accelerating the inference instead of the training.