Bo Pang
Bo Pang
Seems like a problem caused by Apex. Could you please try to run "demo.py" without '--apex' to confirm whether it is caused by Apex?. Evaluation mode does not need so...
And please also upgrade your gcc version to > 5.0 (our version is 5.4.0 20160609).
Right now, the provided model (weight) is trained on pedestrian data. So if you want to track cars, you need to retrain it. The good news is that we are...
TubeTK does not use detection results.
I think here is some problem too. The "dim" parameter for softmax should be 0 in my view of point.
Sorry for the typos. 1. The prediction head is added on f_\theta. 2. argmin -> argmax in Eq.4. 3. SimCLR's augmentation is OK for SMoG.