search

vt-vl-lab / flownet2.pytorch

Off-the-shelf FlowNet module in PyTorch-0.3.0
Apache License 2.0
118 stars 37 forks source link

correlation function runs very slow #4

Open sczhou opened 6 years ago

sczhou commented 6 years ago

Hi, I'm trying to train FlowNet2 separately. It's very slow when I train the flownetC, It's about 5.5s a batch (batch size = 8) on TiTan X, I don't know whether there are some problems, I want to know how fast you train, and could you give some advice? Thanks for your help!

Yuliang-Zou commented 6 years ago

What's the input size and how do you load the image? Large resolution input indeed takes a lot of time.

sczhou commented 6 years ago

@ Yuliang-Zou Thanks for your reply, My input size is 320*448 that just as the Flownet ICCV15 paper's setting and I load the image by dataloader that load the batch images per iteration. My FlownetS just costs 0.3s/batch, but the FlownetC costs 5s/batch. as u know, in origin paper, the flownetC is just twice the time of flownetS, so are there some problems? and how about your training time? Thanks!