Skip to main content

What are Jobs?

Jobs are tasks that can be run and automatically closed after finishing. CGC is designed to enhance the development of various AI tasks by allowing you to incorporate your own images into your namespace. Its purpose is to assist you in scenarios where computational resources are needed only for a defined period of time. This feature is especially useful for users who need to run a specific application or job once, without the need to maintain a running instance of the application or job.

What is the main difference between Jobs and Applications?​

The main difference between Jobs and Applications is that Jobs are designed to run once and automatically close after finishing. Applications are designed to run continuously and can be accessed at any time.

Private Image Repository​

We recognize the need for privacy, hence most of your images can be kept private. We provide an option for you to input a secret to your private repository, ensuring your code and data are securely used.

Initialization Command​

The most efficient way to input a start-up command is by implementing Docker's ENTRYPOINT into your image. However, if you need to modify it or pass it dynamically, we provide a variable that allows you to make changes during creation.

Application Code​

The application code can be integrated into your image. Additionally, CGC volumes can be utilized to enhance accessibility during development. When combined with the full mount path and the startup command, custom code can be directly delivered and executed from your volumes.

Custom Environment Variables​

You can pass custom environment variables to your image. This is done by using the -e flag in the cgc job create command. The variables will be available in your container and can be used to configure your application dynamically.

cgc job create -n <name> -c <cpu_cores> -m <RAM GiB> -g <gpu_count> -gt <gpu_type> --image <image_name> -e <env_var_name>=<env_var_value>

Custom arguments​

You can pass custom arguments to your application. This is done by using the -- flag in the cgc job create command. The arguments will be passed to your application as command-line arguments.

cgc job create -n <name> -c <cpu_cores> -m <RAM GiB> -g <gpu_count> -gt <gpu_type> --image <image_name> -- <arg1> <arg2> ...

Custom App Token​

You can set a custom app token for your application. This token will be used to authenticate requests to your application. You can set it using the -ce flag in the cgc job create command.

cgc job create -n <name> -c <cpu_cores> -m <RAM GiB> -g <gpu_count> -gt <gpu_type> --image <image_name> -ce app_token=<app_token_value>