You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
For co2 efficiency, can you please release some pertained models? (of what you have of course...)
I'm assuming you already have a human pose estimator, so it would just be a waste to every person to train a new one.
The text was updated successfully, but these errors were encountered:
Hi, Thanks for your interest. We don't have any pretrained human model at the moment, but we have had several people request this and will likely add one in a coming update. If you're interested in helping make it happen then we would welcome a PR.
As far as I know there generally aren't large pose datasets for animals that would be worth releasing as a pretrained model, except maybe the Caltech Birds dataset (although hopefully this will change in the future). Generally models used for animal pose estimation are trained on small datasets for specific experimental setups, and therefore are not much use outside of that specific domain. We have a pretrained fly model in our data repo used for the example notebooks. https://github.com/jgraving/DeepPoseKit-Data/tree/master/datasets/fly but this is not much use outside of that specific recording setup.
For co2 efficiency, can you please release some pertained models? (of what you have of course...)
I'm assuming you already have a human pose estimator, so it would just be a waste to every person to train a new one.
The text was updated successfully, but these errors were encountered: