NA on precision and roc_auc

[joonan30]

Hi Kathy,

Thanks for the nice tool. I am trying to train a model with my own ATAC-seq data. It's about 49 datasets so I updated the config file to n_genomic_features: 49. When I checked the train result selene_sdk.train_model.train.txt , most of loss results are 0.

loss
0.0009951249016627846
0.0
0.0
0.0
0.0
0.0
0.0

Then, the validation result file contains NA for average_precision and roc_auc are 0.

loss    average_precision       roc_auc
0.0     NA      NA
0.0     NA      NA
0.0     NA      NA

Do you know what would cause this?

Thanks,

[kathyxchen]

Hi, can you email me at chen.kathleenm@gmail.com with your configuration file information? Happy to help try to reproduce the error and troubleshoot. Thanks!

[joonan30]

Thanks Kathy, I just sent it.

[esignor]

Good morning,
I'm sorry to disturb you, but I am encountering a similar issue when using the sei-framework ([https://github.com//issues/30]).
May I know if and how you managed to solve it?
Thank you very much for your help!

Eleonora