Zehao Shi's starred repositories

segment-anything

The repository provides code for running inference with the SegmentAnything Model (SAM), links for downloading the trained model checkpoints, and example notebooks that show how to use the model.

Language:Jupyter NotebookLicense:Apache-2.0Stargazers:43392Issues:297Issues:606

ColossalAI

Making large AI models cheaper, faster and more accessible

Language:PythonLicense:Apache-2.0Stargazers:37527Issues:377Issues:1541

DeepSpeed

DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.

Language:PythonLicense:Apache-2.0Stargazers:32144Issues:321Issues:2443

mojo

The Mojo Programming Language

qlora

QLoRA: Efficient Finetuning of Quantized LLMs

Language:Jupyter NotebookLicense:MITStargazers:9213Issues:83Issues:240

dinov2

PyTorch code and models for the DINOv2 self-supervised learning method.

Language:Jupyter NotebookLicense:Apache-2.0Stargazers:7614Issues:94Issues:337

DeepSpeedExamples

Example models using DeepSpeed

Language:PythonLicense:Apache-2.0Stargazers:5566Issues:73Issues:506

bitsandbytes

Accessible large language models via k-bit quantization for PyTorch.

Language:PythonLicense:MITStargazers:5207Issues:46Issues:902

alpa

Training and serving large-scale neural networks with auto parallelization.

Language:PythonLicense:Apache-2.0Stargazers:2964Issues:46Issues:295

UniAD

[CVPR 2023 Best Paper] Planning-oriented Autonomous Driving

Language:PythonLicense:Apache-2.0Stargazers:2714Issues:33Issues:154

optimum

🚀 Accelerate training and inference of 🤗 Transformers and 🤗 Diffusers with easy to use hardware optimization tools

Language:PythonLicense:Apache-2.0Stargazers:2068Issues:54Issues:617

lion-pytorch

🦁 Lion, new optimizer discovered by Google Brain using genetic algorithms that is purportedly better than Adam(w), in Pytorch

Language:PythonLicense:MITStargazers:1889Issues:14Issues:23

SparK

[ICLR'23 Spotlight🔥] The first successful BERT/MAE-style pretraining on any convolutional network; Pytorch impl. of "Designing BERT for Convolutional Networks: Sparse and Hierarchical Masked Modeling"

Language:PythonLicense:MITStargazers:1374Issues:26Issues:78

aliyun-oss-python-sdk

Aliyun OSS SDK for Python

Language:PythonLicense:MITStargazers:905Issues:33Issues:169

eRPC

Efficient RPCs for datacenter networks

Language:C++License:NOASSERTIONStargazers:816Issues:33Issues:96

PatrickStar

PatrickStar enables Larger, Faster, Greener Pretrained Models for NLP and democratizes AI for everyone.

Language:PythonLicense:BSD-3-ClauseStargazers:735Issues:16Issues:56

Fast-BEV

Fast-BEV: A Fast and Strong Bird’s-Eye View Perception Baseline

Language:PythonLicense:NOASSERTIONStargazers:511Issues:13Issues:76

BMTrain

Efficient Training (including pre-training and fine-tuning) for Big Models

Language:PythonLicense:Apache-2.0Stargazers:488Issues:11Issues:83

MS-AMP

Microsoft Automatic Mixed Precision Library

Language:PythonLicense:MITStargazers:442Issues:10Issues:53

dino-vit-features

Official implementation for the paper "Deep ViT Features as Dense Visual Descriptors".

Language:PythonLicense:MITStargazers:318Issues:4Issues:21

algorithmic-efficiency

MLCommons Algorithmic Efficiency is a benchmark and competition measuring neural network training speedups due to algorithmic improvements in both training algorithms and models.

Language:PythonLicense:Apache-2.0Stargazers:280Issues:23Issues:197

Pytorch-PCGrad

Pytorch reimplementation for "Gradient Surgery for Multi-Task Learning"

Language:PythonLicense:BSD-3-ClauseStargazers:267Issues:5Issues:17
Language:PythonLicense:Apache-2.0Stargazers:174Issues:9Issues:4

infinity

A lightweight C++ RDMA library for InfiniBand networks.

Language:C++License:MITStargazers:166Issues:6Issues:8

sagemaker-debugger

Amazon SageMaker Debugger provides functionality to save tensors during training of machine learning jobs and analyze those tensors

Language:PythonLicense:Apache-2.0Stargazers:157Issues:25Issues:92

mtm

MTM Masked Trajectory Models for Prediction, Representation, and Control.

Language:PythonLicense:MITStargazers:135Issues:11Issues:2

slapo

A schedule language for large model training

Language:PythonLicense:Apache-2.0Stargazers:126Issues:13Issues:17

rotograd

Official Pytorch's implementation of RotoGrad

SHARK-Turbine

Unified compiler/runtime for interfacing with PyTorch Dynamo.

Language:PythonLicense:Apache-2.0Stargazers:64Issues:24Issues:306

rdmapp

C++ interfaces for RDMA access

Language:C++License:Apache-2.0Stargazers:39Issues:3Issues:1