A Deep Learning Based Voice Analytics toolkit
-
Updated
Jul 6, 2021 - Python
A Deep Learning Based Voice Analytics toolkit
A reinforcement learning-based system designed to detect and prevent jailbreak attempts in AI models, ensuring safe and controlled model behavior under adversarial conditions.
A comprehensive AI-powered mental health tracking and wellness platform built with Streamlit and BERT-based emotion classification.
Custom Italian-language chatbot that chats about weather, news, finance, manages TODOs, and even tells jokes.
Sentiment Analysis comparing SVM & Transfomer
AI-powered mental wellness platform providing empathetic conversational support, emotion detection, crisis awareness, RAG-based responses, personalized interventions, and mood tracking.
A classification project that detects hate speech in social media comments using fine-tuned transformer models, zero-shot learning, and baseline classifiers on the UC Berkeley D-Lab dataset.
AI-powered music recommendation web app that analyzes user mood from natural language using Hugging Face DistilBERT and recommends music through the Spotify Web API, built with React, Django REST Framework, Firebase Authentication, Firestore, Tailwind CSS, and Shadcn UI.
Architecting a high-speed, semantic security firewall using fine-tuned DistilBERT to detect adversarial prompt injections within aerospace procurement workflows.
The system takes invoice text as input (extracted from documents such as PDFs) and predicts the most appropriate category.
AI-powered grievance management system using NLP/ML for multilingual complaint translation, DistilBERT classification, urgency detection, and MiniLM similarity analysis. Built with Next.js, FastAPI, PostgreSQL, Redis, Celery, and JWT. Dockerized locally and deployed on Azure Containers. Archived; live demo available.
A small example showing how to fine-tune DistilBERT (TF version)
Lilypad sentiment analysis module
Case prático para vaga de estágio em desenvolvimento na AutoU. A proposta era desenvolver uma aplicação web que utilizasse IA para classificar emails e sugerir uma resposta automática
This is our 5th Semester EDAI project
LLM transformer Classifier with DistilBERT model
This is a production ready DistilBERT Sentiment Analysis model for service reviews designed to work as a low cost market research tool with the nuiance of an actual market researcher.
Pairfect 💞 is an AI-powered relationship app that analyzes couple photos and personalities to reveal compatibility, create romantic art, poems, and playlists. Powered by GPT-4o and HuggingFace, it transforms emotions into art, helping couples visualize their bond through intelligent, creative storytelling. ❤️
Sarcasm detection on news headlines — TF-IDF baseline vs. fine-tuned DistilBERT (91.9% accuracy), with detailed error analysis on model failure modes. FastAPI + React, deployed live.
To associate your repository with the distilbert-model topic, visit your repo's landing page and select "manage topics."