How vkuzina/cassed works
CASSED detects multilabel sensitive-data categories in structured table columns. The main workflow converts standard CSV values and labels into Flair sentence data, then trains a BERT-backed classifier and evaluates it; README also documents using pretrained models. Source excerpts establish preparation and training wiring, but test_model.py and several documented runtime details were not sampled, so those links rely on README evidence.
Main entry points: train_classifier.py Training workflow · test_model.py Model testing · run_config.py Run configuration · prepare_data.py CSV converter
Components 16
Data preparation
- Standard datasets
datasets - Run configuration
config/run_config.py - CSV converter
prepare_data.py - Prepared datasets
datasets/dessi_prepared
Model training
- Training workflow
train_classifier.py - Flair corpus
train_classifier.py - Transformer embeddings
train_classifier.py - Multilabel classifier
extended_flair/extended_text_classifier.py - Sentence splitting
extended_flair/extended_sentence.py - Flair trainer
train_classifier.py
Evaluation and inference
- Trained model
- Model testing
test_model.py - Classifier evaluation
extended_flair/extended_text_classifier.py - Evaluation results
config/run_config.py - Pretrained models
Other components
- Dataset user
Connections 20
- Dataset user → Standard datasets: provides data
README.md - Dataset user → Run configuration: sets paths
README.md - Run configuration → CSV converter: configures paths
prepare_data.py - Standard datasets → CSV converter: reads CSVs
prepare_data.py - CSV converter → Prepared datasets: writes sentences
prepare_data.py - Run configuration → Training workflow: configures training
train_classifier.py - Prepared datasets → Flair corpus: loads splits
train_classifier.py - Training workflow → Flair corpus: creates corpus
train_classifier.py - Training workflow → Transformer embeddings: initializes
train_classifier.py - Training workflow → Multilabel classifier: constructs
train_classifier.py - Multilabel classifier → Sentence splitting: splits sentences
extended_flair/extended_text_classifier.py - Training workflow → Flair trainer: initializes and runs
train_classifier.py - Flair trainer → Multilabel classifier: trains
train_classifier.py - Flair trainer → Trained model: saves model
README.md - Dataset user → Model testing: runs testing
README.md - Model testing → Trained model: loads model
README.md - Model testing → Pretrained models: selects model
README.md - Model testing → Classifier evaluation: evaluates predictions
README.md - Training workflow → Classifier evaluation: evaluates test split
train_classifier.py - Classifier evaluation → Evaluation results: reports metrics
extended_flair/extended_text_classifier.py