- Linux (Ubuntu 14.04 LTS)
- 4-core (8 thread) CPU or more
- 12 GB of RAM or more
- 5 GB or more of free hard disk space, as much as possible recommended to train large models (big translators, using over one million sentence pairs)
6-step (+1) procedure, it could take 1-2h depending on hardware specifications:
- Download mtradumatica
$ git clone https://github.com/tradumatica/mtradumatica.git
$ cd mtradumatica
- Ensure that all prerequisites are installed
$ sudo ./scripts/run-as-root.sh
- Download and install local dependencies
$ ./scripts/install.sh
- Create the system database database
$ source venv/bin/activate
$ python db_create.py
$ deactivate
- Start service
$ ./scripts/startup.sh
- Go to "http://localhost:8080" with a web browser
- Enjoy!
Before the startup, edit the file conf/gunicorn.conf
and locat the following line
bind = '0.0.0.0:8080'
Here, replace 8080 by another port number (<= 65535).