search

TengdaHan / TemporalAlignNet

[CVPR'22 Oral] Temporal Alignment Networks for Long-term Video. Tengda Han, Weidi Xie, Andrew Zisserman.
MIT License
113 stars 4 forks source link

About pre-training hardware #2

Closed tomchen-ctj closed 2 years ago

tomchen-ctj commented 2 years ago

Hi,

Thank you for your solid work and code releases. May I ask how many GPUs are used in the basic TAN pre-training? (Sorry if I missed that description in your paper)

Thanks!

TengdaHan commented 2 years ago

Hi, the basic TAN pre-training is on frozen visual features, so it requires fewer GPUs. I mostly use only one GPU (but with 48gb gpu-memory) for one experiment.

tomchen-ctj commented 2 years ago

I see, thank you!