Use PEFT or Full-parameter to CPT/SFT/DPO/GRPO 600+ LLMs (Qwen3.6, DeepSeek-R1, GLM-5.1, InternLM3, Llama4, ...) and 300+ MLLMs (Qwen3-VL, Qwen3-Omni, InternVL3.5, Ovis2.5, GLM4.5v, Gemma4, Llava, Phi4, ...) (AAAI 2025).

AdalFlow: The library to build & auto-optimize LLM applications.

A persistent, unified memory layer for all your AI agents (e.g. Claude Code, Codex), backed by Markdown and Milvus.

Train embedding and reranker models for retrieval tasks on Apple Silicon with MLX

Production-ready RAG Framework (Python/FastAPI). 1-line config swaps: 6 Vector DBs (Weaviate, Pinecone, Qdrant, ChromaDB, pgvector, MongoDB), 5 LLMs (Gemini, OpenAI, Claude, Ollama, OpenRouter). OpenAI-compatible API. 2100+ tests.

Small Language Model Inference, Fine-Tuning and Observability.

An easy-to-use python toolkit for flexibly adapting various neural ranking models to target domain.

A library integrating embedding and reranker models from OpenAI, SentenceTransformers etc for semantic search in vector database.

1st Place Solution for Eedi - Mining Misconceptions in Mathematics Kaggle Competition

An OpenAI Compatible API which integrates LLM, Embedding and Reranker. 一个集成 LLM、Embedding 和 Reranker 的 OpenAI 兼容 API

Unified LLM API client library for Python. Simple API for Chat, Embedding, Rerank, and Tokenizer. OpenAI-compatible with streaming support and unified usage tracking.

NeuralCache is a drop-in reranker for Retrieval-Augmented Generation (RAG) that learns which context the model actually uses.

Langchain integration package with LocalAI. It provides embeddings and reranker support

A Python cli-command tool for creating reports for any Google query.

FastAPI Local RAG Q&A Demo

The method of re-ranking involves a two-stage retrieval system, with re-rankers playing a crucial role in evaluating the relevance of each document to the query. RAG systems can be optimized to mitigate hallucinations and ensure dependable search outcomes by selecting the optimal reranking model.

A comprehensive RAG FastAPI service that handles document uploads and retrievals, built with Python. Uses PyMuPDF for document processing, turbopuffer for vector storage, OpenAI for models, and cohere for reranking.

This is RAG Modules Repo. This includes various modules in the RAG ecosystem.

Multi-Objective Recommender System

聚焦电商场景咨询意图杂、PDF 商品和规则手册信息提取难以及核心业务办理依赖人工等痛点，搭建全栈电商智能客服中台。系统采用"意图拦截分流 + Agent 工具调度"架构，构建起多模态文档解析与结构化入库、意图驱动的 Agent 自主调度业务工具与 RAG 知识检索并经混合召回与交叉编码器重排生成回答、四维自动化量化评测的完整闭环链路，显著提升复杂咨询环境下的响应速度与回复精度。