Lambda CLI

Lambda CLI is a command-line tool for interacting with the Lambda Labs cloud GPU API.

Features

• Validate the API key
• List all available GPU instances
• Start a GPU instance with the specified SSH key
• Stop a specified GPU instance
• List all running GPU instances
• Continuously find and start a GPU instance when it becomes available

Installation

To use Lambda CLI, you need to have Rust and Cargo installed. You can install Rust and Cargo by following the instructions here.

Clone the repository and navigate to the project directory:

git clone https://github.com/cybrly/lambda_cli.git
cd lambda_cli
cargo build --release

Usage

Before using Lambda CLI, set your Lambda API key as an environment variable:

export LAMBDA_API_KEY=your_api_key

Run the CLI tool:

./target/release/lambda_cli [COMMAND]

Commands

• list: List all available GPU instances.
• start --gpu <GPU_TYPE> --ssh <SSH_KEY>: Start a GPU instance with the specified SSH key.
• stop --gpu <GPU_INSTANCE_ID>: Stop a specified GPU instance.
• running: List all running GPU instances.
• find --gpu <GPU_TYPE> [--ssh <SSH_KEY>] [--sec ]: Continuously find and start a GPU instance when it becomes available.

Examples

Validate the API key:

./target/release/lambda_cli

List all available GPU instances:

./target/release/lambda_cli list

Start a GPU instance with the specified SSH key:

./target/release/lambda_cli start --gpu "gpu_1x_a10" --ssh "Chris"

Stop a specified GPU instance:

./target/release/lambda_cli stop --gpu "instance_id"

List all running GPU instances:

./target/release/lambda_cli running

Continuously find and start a GPU instance when it becomes available:

./target/release/lambda_cli find --gpu "8x_h100" --ssh "SSH_KEY_NAME" --sec 30