PyTorch implementation of some attentions for Deep Learning Researchers.

2026 AI/ML internship & new graduate job list updated daily

A beamer template for LAMDA lab at NJU

My learning notes/codes for ML SYS.

Writing AI Conference Papers: A Handbook for Beginners

Qwen3-omni is a natively end-to-end, omni-modal LLM developed by the Qwen team at Alibaba Cloud, capable of understanding text, audio, images, and video, as well as generating speech in real time.

💥 Fast State-of-the-Art Tokenizers optimized for Research and Production

📜 Paper list on decoding methods for LLMs and LVLMs

An automated data pipeline scaling RL to pretraining levels

Methods and Implements of Deep Clustering

Official repo for paper: "GRACE: Generative Representation Learning via Contrastive Policy Optimization"

[NeurIPS'23 Oral] Visual Instruction Tuning (LLaVA) built towards GPT-4V level capabilities and beyond.

An open source implementation of CLIP.

A collection of papers on the topic of ``Computer Vision in the Wild (CVinW)''

[COLM2025] "Weak-for-Strong: Training Weak Meta-Agent to Harness Strong Executors"

An Open-Source Large-Scale Reinforcement Learning Project for Search Agents

Reading notes about Multimodal Large Language Models, Large Language Models, and Diffusion Models

A version of verl to support diverse tool use

LIMI: Less is More for Agency

[COLM 2025] Code for Paper: Learning Adaptive Parallel Reasoning with Language Models

All-in-One Sandbox for AI Agents that combines Browser, Shell, File, MCP and VSCode Server in a single Docker container.

Fast and memory-efficient exact kmeans

Qwen3 is the large language model series developed by Qwen team, Alibaba Cloud.

Qwen3-VL is the multimodal large language model series developed by Qwen team, Alibaba Cloud.

Data Synthesis for Deep Research Based on Semi-Structured Data

Qwen-Image is a powerful image generation foundation model capable of complex text rendering and precise image editing.

Easy Data Preparation with latest LLMs-based Operators and Pipelines.

MiroThinker is open-source agentic models trained for deep research and complex tool use scenarios.