Also using svd method? - Githubissues

ethanhe42 / channel-pruning

Channel Pruning for Accelerating Very Deep Neural Networks (ICCV'17)

https://arxiv.org/abs/1707.06168

MIT License

1.08k stars 310 forks source link

Closed Wenstery closed 7 years ago

Wenstery commented 7 years ago

From the code, VGG16 acceleration factor can be 3X-4X, not only by using channel-pruning method but also used svd ? Is it true?

And What about the effect of ssd mobilenet ?

ethanhe42 commented 7 years ago

It use SVD for spatial decomposition and channel decomposition, please refer to paper for details.
haven't studied.

Wenstery commented 7 years ago

Oh,the paper hasn't described the svd steps, thanks a lot.

ethanhe42 commented 7 years ago

details are in:

spatial decomposition (Speeding up Convolutional Neural Networks with Low Rank Expansions)
channel decomposition (Accelerating Very Deep Convolutional Networks for Classification and Detection)