Oracle bone script decipherment via human-workflow-inspired deep learning

A lightweight LMM-based Document Parsing Model

Monkey (LMM): Image Resolution and Text Label Are Important Things for Large Multi-modal Models (CVPR 2024 Highlight)

Reference PyTorch implementation and models for DINOv3

Detect Anything via Next Point Prediction (Based on Qwen2.5-VL-3B)

AI-assisted Deciphering Oracle Bone Script

An open-source implementaion for fine-tuning Qwen-VL series by Alibaba Cloud.

Contexts Optical Compression

A simple, easy-to-hack GraphRAG implementation

A modular, primitive-first, python-first PyTorch library for Reinforcement Learning.

A comprehensive survey of Composed Multi-modal Retrieval (CMR), including Composed Image Retrieval (CIR) and Composed Video Retrieval (CVR), etc.

Efficient Retrieval Augmentation and Generation Framework

Source code and dataset for ICMR'24 paper "Component-Level Oracle Bone Inscription Retrieval" (Best Paper Candidate)

Source code for AAAI'25 paper "Component-Level Segmentation for Oracle Bone Inscription Decipherment"

Enhanced ChatGPT Clone: Features Agents, MCP, DeepSeek, Anthropic, AWS, OpenAI, Responses API, Azure, Groq, o1, GPT-5, Mistral, OpenRouter, Vertex AI, Gemini, Artifacts, AI model switching, message…

Community maintained hardware plugin for vLLM on Ascend

A simple, locally hosted Web Search MCP server for use with Local LLMs

VeOmni: Scaling Any Modality Model Training with Model-Centric Distributed Recipe Zoo

Deep Research Agent CognitiveKernel-Pro from Tencent AI Lab. Paper: https://arxiv.org/pdf/2508.00414

Fine-tuning & Reinforcement Learning for LLMs. 🦥 Train OpenAI gpt-oss, DeepSeek, Qwen, Llama, Gemma, TTS 2x faster with 70% less VRAM.

"RAG-Anything: All-in-One RAG Framework"

An Efficient and User-Friendly Scaling Library for Reinforcement Learning with Large Language Models

Train transformer language models with reinforcement learning.

verl: Volcano Engine Reinforcement Learning for LLMs

EasyR1: An Efficient, Scalable, Multi-Modality RL Training Framework based on veRL

Few-Shot Cross-Lingual Stance Detection with Sentiment-Based Pre-Training

llm & rl