最有价值项目

AI orbit monitor — papers, models, benchmarks, company news in one dashboard. 40+ sources, LMArena leaderboard, local LLM summary.

Access different AI models in a one place

Open-source Prompt Firewall — deflect up to 95% of redundant LLM traffic before it leaves your infrastructure. Documents: https://isartor-ai.github.io/Isartor/index.html

Run Qwen3.5 on your iPhone. No cloud. No API keys. No subscriptions. 100% on-device.

Trained Neural Networks (LSTM, HybridCNN/LSTM, PyramidCNN, Transformers, etc.) & comparison for the task of Hate Speech Detection on the OLID Dataset (Tweets).

The traditional machine learning models give a lot of pain when we do not have sufficient labeled data for the specific task or domain we care about to train a reliable model. Transfer learning allows us to deal with these scenarios by leveraging the already existing labeled data of some related task or domain. We try to store this knowledge gained in solving the source task in the source domain and apply it to our problem of interest. In this work, I have utilized Transfer Learning utilizing BertForSequenceClassification model. Also tried RobertaForSequenceClassification and XLNetForSequenceClassification models for Fine-Tuning the Model.

Training-free ordinal & sign quantization for compressed nearest-neighbour retrieval over high-dimensional embeddings. Pure Rust, zero system dependencies.

⚠️ DEPRECATED - Use https://github.com/SylphxAI/coderag instead

A modular framework for building and deploying Retrieval-Augmented Generation (RAG) systems with built-in evaluation and monitoring.

Production-grade framework for building multi-agent AI systems. Graph-based orchestration, LLM-agnostic (OpenAI, Google GenAI, Anthropic), 3-layer memory (Redis cache + Postgres + vector store), live agents, parallel tool execution, and native MCP. Ships a full ecosystem: backend, REST API + CLI, TypeScript SDK, and React playground

An open-source vector database

Memcortex is a lightweight semantic memory layer for LLMs using open source embedding model and Weaviate for vector storage and retrieval.

A better version of KU's course catalogue

High-performance, Hardware-First Vector Database engine built in Zig and Go. Sub-millisecond HNSW search with 64-bit metadata filtering and 1.8MB binary footprint.

Chat with Anything 📃

🧬 Self-optimizing vector database using MAP-Elites. 51x faster than ChromaDB, 82x faster than LanceDB, 100% recall. Auto-evolves optimal index configs via Quality Diversity.

OpenAI-compatible bridge that exposes Codex CLI/SDK as /v1/chat/completions with multimodal + JSON schema

This repository provides core code for managing large volumes of video footage, enabling content understanding, automatic tagging, and vector database storage. It integrates multimodal models and LLMs for accurate descriptions and semantic search. A web interface allows visualization.

Ensemble Integration: a customizable pipeline for generating multi-modal, heterogeneous ensembles

The 1st place solution for SIGIR 2020 E-Commerce Workshop Multimodal Product Classification Challenge