Train Models Contrastively in Pytorch

Radient turns many data types (not just text) into vectors for similarity search, RAG, regression analysis, and more.

Self-adaptive Planning Agent。自适应规划代理的多模态检索增强生成技术。

Anaya is a Content Engine that specializes in analyzing and comparing multiple PDF documents. It uses Retrieval Augmented Generation (RAG) techniques to effectively retrieve, assess, and generate insights from the documents.

Multimodal RAG and comparisons between language models. (Project for Deep Learning Module at the FHSWF)

Multimodal RAG powered by Milvus, Visualized BGE model, and GPT-4o.

Summer of Code 2024

How to build a Multi-Modal retrieval system using LlamaIndex. Wikipedia Text embedding index: Generate GPT text embeddings from OpenAI for texts Wikipedia Images embedding index: CLIP embeddings from OpenAI for images

Star Stream: A powerful, custom-built streaming platform for immersive event broadcasting. Features include live video streaming, audio transcription, AI-powered video Q&A, and real-time chat – all in one seamless solution.