Skip to content

Samples of multi-class text classification with Differential Privacy Tensorflow 2.0

Notifications You must be signed in to change notification settings

sharmi1206/differential-privacy-tensorflow

Repository files navigation

differential-privacy-tensorflow (Reference https://github.com/tensorflow/privacy)

Multi class text classification using bbc text (bbc-text.csv)

Dataset available at https://www.kaggle.com/yufengdev/bbc-text-categorization/data

Multi class/label text classification using text comment

Dataset available at https://www.kaggle.com/c/jigsaw-toxic-comment-classification-challenge/data

TO DO

  1. Download the corressponding datasets and place in the data folder

  2. Download glove.6B.100d.txt from https://www.kaggle.com/terenceliu4444/glove6b100dtxt and place it inside embedding/glove.6B/ folder

Multi class text classification using product reviews

https://www.kaggle.com/selener/multi-class-text-classification-tfidf/data

Following are the differential privacy sample codes with tenskrflow2 and differential privacy

imdb_dpsgd_keras.py -- Movie Reviews Binary Classification (Ref https://github.com/tensorflow/privacy)

multi_image_mnist_dpsgd_keras.py - Multi-class IMAGE classification with MNIST, fashion_MNSIT, and CIFAR datasets (Ref https://github.com/tensorflow/privacy)

multitext_comment_toxicity_dpsgd_keras.py - Multi-class multi-label text classification with toxic comment dataset available from Kaggle

multitext_dpsgd_CNN_keras.py - Multi-class text classification with bbc-text.csv using CNN

multitext_dpsgd_LSTM_keras.py - Multi-class text classification with bbc-text.csv using LSTM

multitext_dpsgd_LSTM_CNN_keras.py - Multi-class text classification with bbc-text.csv using LSTM & CNN

multitext_product_complains_dpsgd_LSTM_CNN.py - Multi-class text classification of product reviews using LSTM & CNN

About

Samples of multi-class text classification with Differential Privacy Tensorflow 2.0

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages