search

amuamushu / adv_avod_ssn

Deep Sensor Fusion for Single Source Robustness
https://ayushmore.github.io/2022-03-07-improving-robustness-via-adversarial-training/
0 stars 0 forks source link

Try out TA's Suggestions to improving inference. #23

Closed amuamushu closed 2 years ago

amuamushu commented 2 years ago

Talked to a TA about improving runtime of existing code. His suggestions are

  1. make sure the model is not updating the weights --> pytorch: model.train() is off
  2. input in the inference doesn't need gradient computations --> pytorch: input.requires_grad(False)
  3. other direction: increase batch_size as much as memory fits

Make sure nvidia-smi shows percentage when it's running.

_Originally posted by @amuamushu in https://github.com/amuamushu/adv_avod_ssn/issues/6#issuecomment-1029469546_

amuamushu commented 2 years ago

won't do bc of time constraint