SVM light API service on Docker only using docker-compose command.
docker run -it -d -p 5001:3001 mhirano/svm-light-api
If you need change port number, please change 5001 to what you want.
After cloning this repository,
UP
docker-compose up -d
Down
docker-compose down
Initially, this is binded on #5001 port.
So, please access to http://localhost:5001/
If you need, please change the port number from 5001 on docker-compose.yml.
Post to /svm-light/v1/learn with your example file.
Your post file name haa to be "example_file" and its MIME type haa to be "text/plain."
If you need use any option on SVM light, please post "option=option" together. Option details is shown on SVM light.
Some options like "-h" are forbidden because you can not use results showing only on the terminal usually.
After submittion and calaculation on docker, you will get the result of svm model.
If you need example file, you can downlod SVM light official example files. More example are on [the SVM light official site] (http://svmlight.joachims.org/).
curl -X POST http://localhost:5001/svm-light/v1/learn -F "example_file=@train.dat;type=text/plain" -F "option=-f 0" > model
Post to /svm-light/v1/classify with your example file (test data file) and model file.
Your post file names have to be "example_file" and "model_file", and these MIME types have to be "text/plain."
If you need use any option on SVM light, please post "option=option" together. Option details is shown on SVM light.
Some options like "-h" are forbidden because you can not use results showing only on the terminal usually.
After submittion and calaculation on docker, you will get the result of svm predictions.
curl -X POST http://localhost:5001/svm-light/v1/classify -F "example_file=@test.dat;type=text/plain" -F "model_file=@model;type=text/plain" > predictions
Copyright © 2018 · All rights reserved. · Masanori HIRANO