ayansinha/false-positives-scancode-bert-base-uncased-L8-1

The BERT model was pretrained on BookCorpus, a dataset consisting of 11,038 unpublished books and English Wikipedia . It was fine-tuned on Scancode Rule texts, specifically in the context of sentence classification . It has the same biases, but as the task it is . is a very specific field(license tags vs false positives) without those intended biases, it's safe to assume those . don't apply at all here . The classes aren't balanced. The errors have lower confidence scores using thresholds on confidence scores . This . makes it a perfect classifier as the classification task is comparatively easier . The . accuracy is very easily achieved every . time, though more learning,