Superjomn

Paddle

PArallel Distributed Deep LEarning

cuda-from-scratch

matshow

WIP: A visualizer for demonstrate algorithms for inter or intra-matrices.

Superjomn

Config files for my GitHub profile.

superjomn.github.io

TiPS

A distributed training framework for DNN with support of both PS and Collective ways.

triton

Development repository for the Triton language and compiler

a_tmp_images

ccgen

Some initial work about a GPU MLIR dialect.

chun-blog

Chun-blog is an emacs plugin for blogging.

CINN

a Compiler Infrastructure for Neural Networks

cutlass

CUDA Templates for Linear Algebra Subroutines

doom-emacs-docker

A docker for minimum emacs env for development

explor

Some misc code for surveying features from tools or libraries.

haskydots

a python expansion of dot

hugo-PaperMod

A fast, clean, responsive Hugo theme.

llm-scratch

Keep record of some experiments.

next-chat

papaya

Minimalist Hugo theme with social buttons and analytics.

pimacs

A programming language and transpiler for Emacs Lisp

pscore

a C++ serving framework

pytorch

Tensors and Dynamic neural networks in Python with strong GPU acceleration

redoc

A DSL to making instruction documents reproducable.

superjomn.github.io2

temp-file

Useless

TensorRT-LLM

TensorRT-LLM provides users with an easy-to-use Python API to define Large Language Models (LLMs) and build TensorRT engines that contain state-of-the-art optimizations to perform inference efficiently on NVIDIA GPUs. TensorRT-LLM also contains components to create Python and C++ runtimes that execute those TensorRT engines.

tensorrtllm_backend

The Triton TensorRT-LLM Backend

the-algorithm

Source code for Twitter's Recommendation Algorithm

torch-orchestrator-prototype

yallm

Some random code and worknotes related to LLM inference