Too low number of IDs with NNs: reverting to linear classifer

[joycelyny]

We have a dataset with two groups that have empty results files when opening the report.tsv, and we think it may be because the linear classifer was used. Is there a way to override this manually?