TriFind

A multimodal AI system that allows text, image, and voice search — powered by CLIP, FAISS, Whisper, and SpeechT5.

Developed in Python using Hugging Face Transformers and Gradio for the interface.

Features

Text & Image Embeddings using OpenAI’s CLIP model

Speech-to-Text with Whisper (via Hugging Face API)

Text-to-Speech with Microsoft SpeechT5

Vector Search powered by FAISS

Sentiment Feedback System to collect and analyze user impressions with DistilBERT

Interactive Gradio Interface for seamless multimodal queries