david-thrower / cerebros-core-algorithm-alpha

The Cerebros package is an ultra-precise Neural Architecture Search (NAS) / AutoML that is intended to much more closely mimic biological neurons than conventional neural network architecture strategies.

Other

27 stars 4 forks source link

Add use cases: GPT2 encoder for phishing email detection #125

Closed sashakolpakov closed 7 months ago

sashakolpakov commented 8 months ago

Use a GPT2 tokenizer/preprocessor/encoder combo from keras_nlp on the phishing email database.

david-thrower commented 8 months ago

Checklist (Maintainer to dos):

Clean up cicd-requirements (only packages that are not already in requirements.txt) ✅
Add back the tests that were commented out (except maybe BERT text classification) ✅
Create an issue for hyperparam tuning for GPT text classification ... see if we can get this to 100% recall; 98% precision
Add to URS for commercial UX platform, a text classification module based on GPT
Replace "_" with "-" in file paths (Merge and create issue for later)