开源软件名称：

开源软件地址：

开源软件介绍：

GraphPipe

Machine Learning Model Deployment Made Simple

What is it?

GraphPipe is a protocol and collection of software designed to simplify machinelearning model deployment and decouple it from framework-specific modelimplementations.

The existing solutions for model serving are inconsistent and/or inefficient.There is no consistent protocol for communicating with these model servers soit is often necessary to build custom clients for each workload. GraphPipesolves these problems by standardizing on an efficient communication protocoland providing simple model servers for the major ML frameworks.

We hope that open sourcing GraphPipe makes the model serving landscape afriendlier place. See more about why we built ithere.

Or browse the rest of the documentation.

Features

• A minimalist machine learning transport specification based on flatbuffers
• Simple, efficient reference model servers for Tensorflow, Caffe2, and ONNX.
• Efficient client implementations in Go, Python, and Java.

What is in this repo?

This repo contains documentation as well as the flatubuffer definition filesthat are used by other language specific repos. If you are looking forGraphPipe clients, servers, and example code, check out our other GraphPiperepos:

• https://github.com/oracle/graphpipe-go - the GraphPipe go client library
• https://github.com/oracle/graphpipe-go/tree/master/cmd/graphpipe-tf -Go implementation of a GraphPipe TensorFlow model server
• https://github.com/oracle/graphpipe-go/tree/master/cmd/graphpipe-onnx -a Go implementation a GraphPipe ONNX/Caffe2 model server
• https://github.com/oracle/graphpipe-py - the GraphPipe client library forPython
• https://github.com/oracle/graphpipe-tf-py - a Python implementation ofa remote operation client for TensorFlow, as well as some example serverimplementations

Building flatbuffer definitions

If you've got flatc installed you can just make all but if you don't wantto install it, you can export USE_DOCKER=1 and then make all. (Remember,make needs vars exported, not just on the command-line where you run make).

This will produce the go, c, and python libraries, which can then be copiedinto their projects graphpipe-go, graphpipe-tf-py, and graphpipe-py,respectively.

Contributing

All of the GraphPipe projects are open source. To find out how to contributesee CONTRIBUTING.md

You can also chat us up on our Slack Channel.