LLM驱动的 A/H/美股智能分析：多数据源行情 + 实时新闻 + LLM决策仪表盘 + 多渠道推送，零成本定时运行，纯白嫖. LLM-powered stock analysis system for A/H/US markets.

Agent libOS provides a runtime substrate for self-evolving LLM agents with process-like execution, tool management, and security boundaries implemented through explicit capabilities and runtime pri…

An open-source long-horizon SuperAgent harness that researches, codes, and creates. With the help of sandboxes, memories, tools, skill, subagents and message gateway, it handles different levels of…

taac2025

🚀 Efficient implementations for emerging model architectures

Official pytorch implementation of "MUSE: A Simple Yet Effective Multimodal Search-Based Framework for Lifelong User Interest Modeling"

A high-throughput and memory-efficient inference and serving engine for LLMs

Efficient Triton Kernels for LLM Training

TradingAgents: Multi-Agents LLM Financial Trading Framework

分享AI Infra知识&代码练习：PyTorch、vLLM/SGLang、slime/vime框架入门⚡️、性能加速🚀、大模型基础🧠、AI软硬件🔧等

Rk1 solution on MLSys26 competition for the best MOE NKI kernels

Algorithm powering the For You feed on X

LMCache: Supercharge Your LLM with the Fastest KV Cache Layer

Application of RL in trading

[ICLR2025, ICML2025, NeurIPS2025 Spotlight] Quantized Attention achieves speedup of 2-5x compared to FlashAttention, without losing end-to-end metrics across language, image, and video models.

Toolkit for Universal Retrieval, such as text retrieval, item recommendation, image retrieval, etc.

历年ICLR论文和开源项目合集，包含ICLR2021、ICLR2022、ICLR2023、ICLR2024、ICLR2025.

使用符号回归在中国股市与加密市场上进行高效因子挖掘。/ By the way, Leverage the novel features and advanced financial mathematics introduced in Uniswap V4 to effectively mitigate just-in-time (JIT) liquidity provision issues.

Fast and memory-efficient exact attention

Examples for Recommenders - easy to train and deploy on accelerated infrastructure.

TensorDict is a pytorch dedicated tensor container.

MixFormer models for recommender systerm

Triton language and compiler for Ascend NPU

ai agents for trading

Open-source swarm intelligence engine — simulate hundreds of AI agents to predict outcomes. Includes MCP server for Claude.

Official Implementation of KDD‘26 ADS：The Best of Both Worlds: Harmonized Semantic and Hash IDs for Sequential Recommendation