search

tyronechen / genomenlp

https://genomenlp.readthedocs.io/en/latest/
MIT License
5 stars 3 forks source link

[BUG] Training step breaks because of type assumptions on class label #4

Closed tyronechen closed 1 year ago

tyronechen commented 1 year ago

Describe the bug When running train, if the dataset object has previously encoded class labels, it will break.

This is because it tries to coerce this to a class and fails.

To Reproduce Run train with any data

Expected behavior Training to be conducted as normal

Suggested fix Added a condition to skip type change if it is correct

Screenshots NA

tyronechen commented 1 year ago

Fixed in 2.5.0