rai-project/rai

Stars
35
Rank 750,836 (Top 15 %)
Language
Go
License
University of Ill...
Created almost 8 years ago
Updated over 5 years ago

rai-project/rai

rai-project

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

The RAI client allows one to interact with a cluster of machine to submit and evaluate code. RAI is a scalable job submission system designed for diverse workloads. RAI’s design addresses challenges of scalability, configurability, security, and cost in delivering a flexible programming environments.

mlmodelscope

MLModelScope is an open source, extensible, and customizable platform to facilitate evaluation and measurement of ML models within AI pipelines.

go-tensorrt

Go binding to TensorRT C API to do inference with pre-trained model in Go

tensorflow-go-examples

Examples of using TensorFlow Go API for different Deep Learning models

go-pytorch

go-fasttext

go-mxnet

Go binding to MXNet C API to do inference with pre-trained model in Go

go-cupti

Use NVIDIA CUPTI from within GO

mxnet

MXNet agent for MLModelScope

raid

The RAI sever evaluates build requests within a container environment. RAI is a scalable job submission system designed for diverse workloads. RAI’s design addresses challenges of scalability, configurability, security, and cost in delivering a flexible programming environments.

pytorch

nvidia-smi

NVIDIA-SMI abstractions. Used to detect the GPU hardware on the system

tensorflow

TensorFlow agent for MLModelScope

go-caffe2

Caffe2 bindings for MLModelScope

cupti_samples

CUPTI samples from NVIDIA

dlframework

Common code for all deep learning predictors used by MLModelScope

go-caffe

Go binding to Caffe C API to do inference with pre-trained model in Go

store

tensorrt

TensorRT agent for MLModelScope

config

Configuration entry point for both RAI and CarML

rai-srcmanager

rai-srcmanager is the utility for managing the rai repositories.

evaluation

Evaluation tools for model performance / accuracy for MLModelScope

plini

caffe2

caffe2 agent used by CarML

dockerfile-builder

The aim of D4P is to enrich the PowerPC container ecosystem by providing both a platform for developers to create docker containers, and for PowerPC community to find docker images. We have already built and published over 200 docker images using this platform, and they are available both in the D4P's image catalog or accessible through the C3SR center's Docker Hub link.

dlplugin

release-tools

bumpversion

profile

vipertags

model

Database models used by RAI and MLModelScope

pipeline

cmd

Tools for the command line for both RAI and CarML

ldcache

tegra

Tools specific to the tegra systems (TX1 and TX2)

web

A collection of web uilities

performance

cpu

cgo

registry

acl

parallel

Parallel tools used by RAI and CarML

caffe

Caffe agent for MLModelScope

inle

tracer

Common tracer abstraction

serializer

cntk

CNTK agent for CarML

dldataset

Builtin deep learning datasets used by MLModelScope

context

database

Database abstraction used by RAI and CarML

scratch

broker

carml-models

utils

Common utility functions

coreml

loadbalance

linguist

archive

sys

low-level system config (rlimits and stat)

synthetic_load

build_tools

health

Dockerfiles

downloadmanager

Utility functions for downloading data from the web

uuid