Question for @DeepLabCut about using previously trained network

@DeepLabCut

Is it possible to use a previously trained network to analyze new videos after the anaconda powershell has been closed?

I trained a network and it was in the middle of testing a batch of videos when the computer lost power and the powershell crashed. I would like to be able to start the process again at the testing phase instead of having to retrain the network, but am unsure if this is possible or how to do it using python. Any tips you can give would be appreciated, thanks!

Matt

Absolutely.

For training: Check out box 2 in the protocol paper on how to easily edit the pose_cfg.yaml file to point to the weights you already have stored.

For analysis: just run the same call, the config file tells the network which weights to use!

image