Overview / Usage

This project addresses the challenge of assessing text stress levels for users. By leveraging a TensorFlow NLP model trained on Reddit posts data, it provides a practical tool for individuals to gauge the emotional intensity of text content. This work has real-world applications in helping users manage stress, anxiety, and emotional well-being by offering a quick and accessible stress assessment for written content. It can be utilized in production as a feature in various applications, such as mental health apps, social media platforms, or communication tools, to empower users to make more informed decisions regarding the emotional impact of their text interactions.

Methodology / Approach

The data is collected from a kaggle dataset, which already had the labeled data with the text taken from different subreddits. This data is cleaned as explained in the video, then works from the input text are tokenized and are given to an nlp model with lstm layers. Then the model is stored as tflite file and is placed in pythonanywhere website server and rest api is used to communicate b/w the app made in the flutter and the model stored in the cloud.

Repository

https://github.com/anna-sai-nikhil/Intel-oneapi-stresstracker