Natural Language Processing (NLP) Examples

This folder presents a collection of examples for training and using transformer-based models for NLP tasks. 
The underlying training method utilizes TripleBlind's Blind Learning algorithm to secure the model's IP 
and the privacy of the trainign data. It also allows the training on federated datasets, without 
moving the data outside the owner's infrastructure.

There are two demos in this folder:
1- Sequence Classification
Sequence (or Text) classification is an NLP task that involves assigning categories or labels to a given text sequence.
For example, sentiment analysis is a popular form of text classification where we assign labels, e.g., positive, negative, 
or neutral to a piece of text. This task is helpful for a wide range of applications, from analyzing customer feedback 
to gauging public opinion on social media.

2- Token Classification
Token classification involves assigning a label or category to individual tokens (usually words or subwords) within a text.
This is a more granular form of text classification where each token in the input text is classified into predefined categories.
For example, Named Entity Recognition (NER) is a popular task of token classification. It focuses on identifying and categorizing
named entities such as person names, organizations, locations within a given text. 

Each subfolder contains its own README file with more detailed information about the specific task, 
datasets, and models used in that area of NLP.

