Skip to content
@toyaix

ToyAIX

Exploring toy-scale AI systems — simple experiments, serious insights.
README.md

Exploring toy-scale AI systems — simple experiments, serious insights.

This page is accessible via 🔗toyaix.top

🚀 TritonLLM Chinese documentation

TritonLLM v0.1.0: 基于 Triton 的大模型推理初步探索(eb16e6a)

Triton Kernel 优先：全新 LLM 推理方式(47e9dcb)

5090显卡+Triton，轻松玩转GPT-OSS-20B！(6bb4b91)

🔧 Triton Runner Chinese user documentation

Triton Runner：项目介绍及展望

Triton Runner：多层级执行

Triton Runner：多层级执行实战

Triton Runner：多层级dump，快乐对精度

Triton Runner：benchmark

🔧 Triton Runner Chinese develop documentation

Triton Runner v0.2.6 : Python调试，Gluon支持(8eebaaa)

Triton Runner v0.2.0 : 支持调试，多版本支持(4b85c7a)

Triton Runner v0.1.5：支持缓存机制，Benchmark更友好(9c28df1)

Triton Runner v0.1.1：多层级 runner 工具(795ff3d)

Triton Runner v0.0.0：cubin Runner(539d549)

💡 Triton OpenCL Chinese documentation

Triton SPIR-V 后端开发：矩阵乘实现验证(953bff6)

Triton SPIR-V 后端开发：向量加实现验证(f66f77c)

Triton SPIR-V 后端开发：PyBind绑定(457f0aa)

Triton SPIR-V 后端开发：新增Pass(08b0e35)

Triton SPIR-V 后端开发：backend 初始化(02ff396)

📖 OpenAI/Triton blog in Chinese

浅析 Triton 执行流程

深度剖析 Triton编译器 MatMul优化（三）—— TMA

深度剖析 Triton编译器 MatMul优化（二）—— MMA

深度剖析 Triton编译器 MatMul优化（一）—— FMA

Triton 社区首贡献：Bug 修复实录

CUDA优化黑魔法：假装CUTLASS库(Triton PR7298)

triton是否会冲击cuda生态？BobHuang的回答

Organizations I Established

I created and maintain the following organizations:

Pinned Loading

  1. TritonLLM TritonLLM Public

    LLM Inference via Triton (Flexible & Modular): Focused on Kernel Optimization using CUBIN binaries, Starting from gpt-oss Model

    Python 56 2

  2. triton-runner triton-runner Public

    Multi-Level Triton Runner supporting Python, IR, PTX, and cubin.

    Python 76 1

Repositories

Showing 6 of 6 repositories

People

This organization has no public members. You must be a member to see who’s a part of this organization.

Top languages

Loading…

Most used topics

Loading…