[535f03]: / docs / AMI.md

Download this file

56 lines (34 with data), 2.0 kB

Using Amazon AMI

You can easily set up OpenSim Reinforcement learning environmen on AWS Cloud using our AMI.
Just create a machine with ami-971be8ee and log in using ssh ubuntu@[IP_OF_THE_MACHINE].
Remember to turn off your machine if you are not using it. Amazon will charge you for every hour of usage.

Screencast

Log in and choose EC2

1

Click "Launch instance"

2

Choose "Community AMI" and seacrh "ami-971be8ee"

3

Choose instance type according to your needs

Attention: They greatly vary in price, check Amazon Pricelist and Terms!
4

Confirm and launch your instance

5

Create a keypair and download it

6

You need to add this key to your keychain or add it to your SSH client. On Linux/MacOS type ssh-add [path-to-key.pem] on windows add it to your ssh client (e.g. putty)

See the confirmation, go to your instances

7

Click on your instance and check IPv4 Public IP

8

Login with ssh client to ubuntu@[IP]

9

Activate conda opensim environment with

source activate opensim-rl

and you are ready to use our osim-rl package. For the example script, navigate to

cd ~/osim-rl/scripts

and run

python example.py --train --model sample

once your model is trained you can run

python example.py --test --model sample --token [YOUR_CROWD_AI_TOKEN]

to submit the solution.