EthoML / VAME

Variational Animal Motion Embedding - A tool for time series embedding and clustering
https://ethoml.github.io/VAME/
GNU General Public License v3.0
12 stars 2 forks source link

Train Issue with aligned data #94

Open La2872 opened 1 month ago

La2872 commented 1 month ago

When I am using a data set that is already aligned and I change the config file to egocentric_data = true it still required pose_erf index and raise an error when I try to use create_trainset. The way I solved it (at least I hope) is to first change to egocentric_data = true when I use pose_to_numpy and then when I create_trainset I need to change in the configrations to egocentric_data = false. I am trying to train the model but there are still problems with the results folder that isn't created as wanted.

luiztauffer commented 1 month ago

hi @La2872 , thank you for reporting this. Could you provide a sample from your dataset to help us debug it?

La2872 commented 1 month ago

Hi! I used fly data that. will give a sample from this. the data is egocentric aligned.

It doesn't let me upload the file because it is too big. How can I send it?

Also, I would appreciate the help for the use of the egocentric alignment for data that is not egocentric aligned

La2872 commented 1 month ago

https://drive.google.com/drive/folders/1wuTQJ8KHhXIeMgiquGWiNGxgDDYQtqbd?usp=sharing

This is the link to the data.

Another issue that I have is that when I train the model the results folder is being deleted.

Thanks for the help.

luiztauffer commented 1 week ago

hi @La2872 sorry for the delayed reply, I've been fixing other issues with VAME, now I can take a look at the aligned data errors. Could you please share give me access to the folder you shared above?