Saturn is opensource Machine Learning Framework.
- Support multi-language:
python
,node
,markdown
- Distribution your work using
remote
- Unit test available, even though asynchronous loop
- Mailing on event (error, finish)
- saturn
- Web based GUI framework
- saturn-core
- Core for running and compile
- will be update language module for supporting more languages
- saturn-docker
- Docker for easy install
-
Installation & Run
docker pull proin/saturn docker run -it --name saturn -v ~/workspace/saturn:/workspace -p 3000:3000 proin/saturn
-
Run as Daemon
docker run -d --name saturn -v ~/workspace/saturn:/workspace -p 3000:3000 proin/saturn docker kill saturn #if you want to kill process
-
Restart
docker restart saturn
-
Attach Running Container
docker exec -i -t saturn /bin/bash
-
Clone & Build
npm install -g lwot git clone https://github.com/proin/saturn cd saturn bower install cp ./controller/express/config-sample.json ./controller/express/config.json lwot install lwot build lwot express run # then started at your browser, localhost:3000
-
Run as Daemon
lwot express forever start # after launched, press Ctrl+C lwot express forever stop # stop daemon lwot express forever log # show logs
-
Update
- you can easily update saturn in web. just click update.
- or run
sh update.sh
in Unix systems
-
you can change configuration editing by
./controller/express/config.json
file{ "port": 3000, // server port to bind "home": "./saturn-project", // workspace home, start from USER HOME "dev": true, // display express log "log": true, // display thread log "hostname": "http://localhost", // direct url "smtp": { // emailjs configuration object "host": "smtp.your-site.net", "user": "[email protected]", "password": "password", "ssl": true }, "mailingList": "[email protected], [email protected]", // who are received? "mailingOn": { "error": true, // mail when error occured "finish": true // mail when work finished }, "readonly": true, // readable project for whom not signin users "user": "admin", // admin user "password": "admin123", // admin password "session": { "secret": "sAtuRN-seSsIon-sEcreT", "resave": false, "saveUninitialized": false }, "MAX_LOG": 500, // display log size "encrypt": "sAtuRn" }
- you can use machine learning libraries in
saturn
- now
saturn
supportspython (including tensorflow)
,R
,torch
- in
saturn-docker
installation, all frameworks are embedded in there.
saturn
support markdown editor in project- you can write down code more effectively
- you can receive email when the work is finished or occured warning or error
- you no longer have to wait for the job to finish
- you can use graph by making structure & calling function.
- for more usage, refer chart demo
- Distribute some work using
remote
features - you can easily distribute in
[project > remote]
- You can use system terminal where
saturn
is installed - This allow to only
GRANTALL
user
- Support Machine Learning Frameworks if it is possible
Torch
Caffe
R