search

ekosman / AnomalyDetectionCVPR2018-Pytorch

Pytorch version of - https://github.com/WaqasSultani/AnomalyDetectionCVPR2018
172 stars 52 forks source link

[Video_demo.py] Can I load multiple feature extractor's weights? #177

Closed Dev-ori closed 5 months ago

Dev-ori commented 1 year ago

Hi, thanks for nice work here.

Now, I have extracted features from multiple feature extractor models.

And I trained them with one AD model(FC layer).

But with your 'Video_demo.py', I can only load 1 feature extractor model weight.

Do you think that i could modify the code to run with multiple feature extractor model?

I'm looking for the solutions to fix this but as far, i couldn't so. hope that you may could see if it is possible.

Questions Summarize:

  1. Can I load multiple feature extractor model weights ?

  2. Or IF I don't "have to" load the feature extractor weights, then How about the concatenated features to loaded.

thank you for reading.

Best regards.

ekosman commented 1 year ago

In order to help, I need to understand the new framework better. Could you maybe draw it?