FunctionLab / selene

a framework for training sequence-level deep learning networks
BSD 3-Clause Clear License
373 stars 87 forks source link

Bugfixes for EvaluateModel and MatFileSampler #168

Closed kathyxchen closed 3 years ago

kathyxchen commented 3 years ago

Reference Issues/PRs

Fixes bugs in EvaluateModel use_features_ord functionality (evaluating a model on only a subset of ordered targets) and MatFileSampler, which reset the pointer to what row to sample too early (i.e. supposed to do it when you've sampled all the rows in the matrix)

What does this implement/fix? Explain your changes.

See above

What testing did you do to verify the changes in this PR?

Ran EvaluateModel using MatFileSampler input data on 2 different cases, 1 subsetting chromatin features and 1 using the full set of features predicted by the model.