2026TAAC腾讯广告算法大赛-KDDCUP方案，best score：0.832321，rank：51

UniRank: A Ranking Model Benchmark for Unified Sequential Modeling and Feature Interaction

TAAC2025初赛第十四名O_o队伍代码

Accelerating MoE with IO and Tile-aware Optimizations

CODA: Rewriting Transformer Blocks as GEMM-Epilogue Programs

TokenSpeed is a speed-of-light LLM inference engine.

A self-hosted ML coding practice platform. 68 problems from ReLU to flow matching — attention, training, RLHF, diffusion, and more. Instant feedback in the browser.

A hyperparameter optimization framework

CUDA kernels for linear attention variants, written in CuTe DSL and CUTLASS C++.

AI agents running research on single-GPU nanochat training automatically

Pytorch domain library for recommendation systems

🎓从0开始训练一个大模型Minimind项目的超详细解析，包括但不限于用到的架构，算法，以及大模型面试经验

The code repository for the KDD 2026 paper "Q-Regularized Generative Auto-Bidding: From Suboptimal Trajectories to Optimal Policies"

Implementation of "FlashPreill: Instantaneous Pattern Discovery and Thresholding for Ultra-Fast Long-Context Prefilling"

flash attention tutorial written in python, triton, cuda, cutlass

Offers a toolset for comprehensive, multi-faceted large-scale data analysis and optimizations

Inference Speed Benchmark for Learning to (Learn at Test Time): RNNs with Expressive Hidden States

IntelliFold: A Controllable Foundation Model for General and Specialized Biomolecular Structure Prediction.

Wan: Open and Advanced Large-Scale Video Generative Models

Research code accompanying AlphaGenome

【Accepted by WWW 2026 🎉🎉】Generative Regression Based Watch Time Prediction for Short-Video Recommendation

High Performance LLM Inference Operator Library

RePo: Language Models with Context Re-Positioning

Kernels, of the mega variety :)