Spaces:
Build error
A newer version of the Streamlit SDK is available: 1.57.0
title: Toxic Tweets
emoji: 🤢
colorFrom: yellow
colorTo: orange
sdk: streamlit
app_file: app.py
pinned: false
Toxic Tweets
Developing a Language Model to classify toxic 🤢 tweets using HugginFace, Streamlit and GitHub.
Jules Blount 31430956
Milestone 1
To begin I should mention I already have a home server with docker already installed with multiple containers running.
The tutorial I followed to install docker on my server is located Here
Docker runtime environment verification:
Python prompt from Python container:
Milestone 2
For milestone-2, I was tasked to develop a streamlit app that allows the user to enter a text, select a pretrained model and get the sentiment analysis of the text using HuggingFace transformers library and HuggingFace Spaces.
Streamlite app is located here
Milestone 3
For milestone-3, I was challenged to build a multi-headed model that’s capable of detecting different types of of toxicity like threats, obscenity, insults, and identity-based hate better than Perspective’s current models. The classifier was be developed using a pretrained language model of my choice, I chose DistilBERT.
Streamlite app is located here
Model development and training can be found in the toxic_comments notebook here