https://github.com/lianxmfor/kfserving-inference-client

a client for kfserving inference protocol - v2 version
https://github.com/lianxmfor/kfserving-inference-client

Last synced: 6 days ago
JSON representation

a client for kfserving inference protocol - v2 version

• Host: GitHub
• URL: https://github.com/lianxmfor/kfserving-inference-client
• Owner: lianxmfor
• Created: 2021-09-02T18:10:57.000Z (about 3 years ago)
• Default Branch: master
• Last Pushed: 2021-11-04T07:52:08.000Z (about 3 years ago)
• Last Synced: 2024-08-02T16:43:26.216Z (3 months ago)
• Language: Go
• Homepage:
• Size: 842 KB
• Stars: 4
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# KFServing Inference Client

A Go re-implementation of [seldon-batch-processor](https://github.com/SeldonIO/seldon-core/blob/master/python/seldon_core/batch_processor.py). See [docs](https://docs.seldon.io/projects/seldon-core/en/stable/servers/batch.html) to understand its usage.

The main reason why we choose to re-implement is because the original implementation follow Seldon protocol, while we need KFServing V2 Protocol in order to use [MLServer](https://github.com/SeldonIO/MLServer) as the inference backed.

## man 

```sh

$ ./kfserving-inference-client -h

Usage of ./kfserving-inference-client:

  -host string

    	The hostname for the seldon model to send the request to, which can be the ingress of the Seldon model or the service itself

  -i string

    	The local filestore path where the input file with the data to process is located

  -m string

    	model name

  -o string

    	The local filestore path where the output file should be written with the outputs of the batch processing

  -u int

    	Batch size greater than 1 can be used to group multiple predictions into a single request. (default 100)

  -w int

    	The number of parallel request processor workers to run for parallel processing (default 100)

```

## Build Docker Image

$ make build