Replacing 3D convolution with any other convolution (depthwise seperable or group convolution)

mzolfaghari / ECO-efficient-video-understanding

Code and models of paper " ECO: Efficient Convolutional Network for Online Video Understanding", ECCV 2018

MIT License

437 stars 96 forks source link

Open trived76 opened 4 years ago

trived76 commented 4 years ago

Hi @mzolfaghari , Thanks very much for this great work. It is not like an issue but just a general question.

Could you please suggest if there is any replacement option for 3D convolution (could I use depthwise or group convolution by manipulating them?

I look forward to hearing from you. Thanks.