This repository is a collection of tools to use with lambda labs gpu cloud compute infrastructure.
The tools depend on a config.json file at the root of the project directory. This is an example of the contents.
{
"api_key": "xxx",
"check_interval": 5,
"wait_interval": 1
}
api_key: api key for lambda labs, you can create by going to this link.
check_interval: number of minutes to wait when running check.py, defaults to 5
wait_interval: number of minutes to wait when running wait.py, defaults to 1
Contiously checks for availability and prints the currently available instances and the locations they are available at to the terminal. The check is run every 5 minutes. python wait.py [gpu_name]
starts a new instance assuming the gpu and location is available.
Usage: python start.py [gpu_name] [region_name]
Continously checks for availability of a particular gpu and launches an instance at the first oppertunity. The terminal will display the ip address of the started machine after it has launched.
Usage: python wait.py [gpu_name]