Late Interaction Models Training & Retrieval

Advancing the frontier of efficient AI

Code for paper "Process-Level Trajectory Evaluation for Environment Configuration in Software Engineering Agents"

The Batched API provides a flexible and efficient way to process multiple requests in a batch, with a primary focus on dynamic batching of inference workloads.

Advancing search on top of AI agents

Get up and running with Kimi-K2.5, GLM-5, MiniMax, DeepSeek, gpt-oss, Qwen, Gemma and other models.

Youtu-Embedding is an industry-leading, general-purpose text representation model developed by Tencent Youtu Lab.

S.U.P.E.R.M.A.N. optimizes the macOS software update experience.

[ICLR 2026] Youtu-GraphRAG: Vertically Unified Agents for Graph Retrieval-Augmented Complex Reasoning

🏡 GitHub Pages template for personal academic homepage

A simple yet powerful agent framework that delivers with open-source models

A Survey of Efficient Attention Methods: Hardware-efficient, Sparse, Compact, and Linear Attention

[COLM 25] Phased Training for LLM-powered Text Retrieval Models Beyond Data Scaling

An extensive list of Shields.io badges.

BrowseComp-Plus: A More Fair and Transparent Evaluation Benchmark of Deep-Research Agent (ACL 2026 Main)

Embedding Atlas is a tool that provides interactive visualizations for large embeddings. It allows you to visualize, cross-filter, and search embeddings and metadata.

《动手学大模型Dive into LLMs》系列编程实践教程

Fused Qwen3 MoE layer for faster training, compatible with Transformers, LoRA, bnb 4-bit quant, Unsloth. Also possible to train LoRA over GGUF

嗨！thesis！哈尔滨工业大学毕业论文LaTeX模板

Python based web automation tool. Powerful and elegant.

An open-source cross-platform alternative to AirDrop

Tips for releasing research code in Machine Learning (with official NeurIPS 2020 recommendations)

CVPR and NeurIPS poster examples and templates. May we have in-person poster session soon!

[EMNLP 2025] TokenSkip: Controllable Chain-of-Thought Compression in LLMs

Microsoft PowerToys is a collection of utilities that supercharge productivity and customization on Windows

SpikingJelly is an open-source deep learning framework for Spiking Neural Network (SNN) based on PyTorch.

Train Models Contrastively in Pytorch

Windows inside a Docker container.