sobalgi / ds222_assignment2 Goto Github PK

Logistic regression classifier implementation with L2 regularization for DBPedia dataset with 50 classes

Simply clone this repository via

git clone https://github.com/sobalgi/DS222_assignment2.git

• Python with NumPy
• tensorflow
• h5py

This consists of 2 parts each included in separate folders.

cd part1_logreg_local
./train_n_log_accuracies.sh

To reproduce or run the In-Memory local implementation of logistic regression and generate logs.

‘part1_logreg_local’ folder has the code files and generated log files for In-Memory local implementation of logistic regression with L2 regularization.

cd part2_logreg_ps
./train_n_log_accuracies.sh

‘part2_logreg_ps’ folder has the codes for ASP, BSP and SSP SGD implementation of L2 regularized logistic regression and appropriate logs.

The detailed report is also included as ‘DS222_as2_report.pdf’