Bitdeer AI Cloud CLI

Bitdeer AI Cloud CLI is a command-line interface (CLI) tool designed to interact with the Bitdeer AI Cloud. This tool allows users to manage and monitor their cloud resources with ease.

Installation

To install bitdeer-ai, follow these steps:

Download the binary

For Windows:
- amd64 link to Windows binary
- arm64 link to Windows binary
For macOS:
- amd64 link to macOS binary
- arm64 link to macOS binary
For Linux:
- amd64 link to Linux binary
- arm64 link to Linux binary

Install the binary:

Move the downloaded binary to a directory included in your system's PATH.

Verify the installation:

bitdeer-ai --version

Usage

bitdeer-ai provides various commands to manage your cloud resources. Below is the basic usage and the available commands.

Basic Usage

BitdeerAI CLI is a CLI tool for BitdeerAI Cloud Platform.

Usage:
  bitdeer-ai [flags]
  bitdeer-ai [command]

Configuration
  login       Login to BitdeerAI Cloud
  logout      Logout from BitdeerAI Cloud

Managements
  training    Training cmd is to manage training jobs on BitdeerAI Cloud.

Additional Commands:
  help        Help about any command

Flags:
  -h, --help      help for bitdeer-ai
  -v, --version   version for bitdeer-ai

Use "bitdeer-ai [command] --help" for more information about a command.

Configuration

Before you start using BitdeerAI CLI, you need to login to the BitdeerAI Cloud. You can login using the following command:

bitdeer-ai login --token "your-api-key"

Training Commands

Create a training job:

Usage

bitdeer-ai training create [flags]

Examples

bitdeer-ai training create -p <project-id> -n <job-name> -j <job-type> -w <worker-spec> -c <num-workers> -i <worker-image> -r <region-id> -z <zone-id>

Flags

-p, --project_id string: Project ID (required)
-n, --job_name string: Job Name (required)
-j, --job_type string: Job Type (required)
-w, --worker_spec string: Worker Spec (required)
-c, --num_workers int: Number of Workers (required)
-i, --worker_image string: Worker Image (required)
-r, --region_id string: Region ID (required) (default "sg")
-z, --zone_id string: Zone ID (required) (default "sg-sg-1")
--args stringArray: Arguments
--cmds stringArray: Commands
--envs stringArray: Environment Variables
--working_dir string: Working Directory
--volume_name string: Volume Name
--volume_mount_path string: Volume Mount Path
-h, --help: Help for create

List all training jobs:

Usage

bitdeer-ai training list [flags]

Examples

bitdeer-ai training list

+--------+--------+-----------+-----------+---------+------------+----------------------+----------------+------------+
| JOB_ID |  NAME  |  STATUS   |   TYPE    | WORKERS | GPUS/WORKER|    CREATED_TIME      |ATTACHED_PROJECT|   REGION   |
+--------+--------+-----------+-----------+---------+------------+----------------------+----------------+------------+
| tj-03  | job-03 | COMPLETED | TORCH_JOB |       2 | 4x H100    | 2024-07-05T04:00:15Z | pj-01          | sg|sg-sg-1 |
| tj-02  | job-02 | SUSPENDED | TORCH_JOB |       4 | 3x RTX3090 | 2024-07-05T03:30:45Z | pj-02          | sg|sg-sg-1 |
| tj-01  | job-01 | FAILED    | TORCH_JOB |       6 | 2x RTX3090 | 2024-07-05T02:10:30Z | pj-01          | sg|sg-sg-1 |
+--------+--------+-----------+-----------+---------+------------+----------------------+----------------+------------+
|                                                                  TOTAL:  3                                          |
+--------+--------+-----------+-----------+---------+------------+----------------------+----------------+------------+

Display a training job:

Usage

bitdeer-ai training get [flags] <job-id>

Examples

bitdeer-ai training get tj-03

+-----------------+--------------------------------+
| FIELD_NAME      | VALUE                          |
+-----------------+--------------------------------+
| TrainingJobID   | tj-03                          |
| JobName         | job-03                         |
| ProjectID       | pj-01                          |
| ProjectName     | LLM Project                    |
| Region          | sg|sg-sg-1                     |
| JobType         | TORCH_JOB                      |
| JobStatus       | COMPLETED                      |
| NumberOfWorkers | 2                              |
| SpecOfWorker    | 4x H100                        |
| WorkerImage     | btdr/pytorch-dist-mnist:latest |
| CreatedAt       | 2024-07-05T04:00:15Z           |
+-----------------+--------------------------------+
+----------------+-----------+
| WORKER_NAME    |   STATUS  |
+----------------+-----------+
| tj-03-master-0 | Succeeded |
| tj-03-worker-0 | Succeeded |
+----------------+-----------+

List workers for a training job:

Usage

bitdeer-ai training workers [flags] <job-id>

examples

bitdeer-ai training workers tj-03

+----------------+-----------+
| WORKER_NAME    |   STATUS  |
+----------------+-----------+
| tj-03-master-0 | Succeeded |
| tj-03-worker-0 | Succeeded |
+----------------+-----------+

Print logs for a worker in a training job:

Usage

bitdeer-ai training logs [flags] <job-id> <worker-name>

Examples

bitdeer-ai training logs -f tj-03 tj-03-master-0

2024-07-05T04:00:15Z: Starting worker tj-03-master-0
2024-07-05T04:00:15Z: Worker tj-03-master-0 is running
2024-07-05T04:00:15Z: Worker tj-03-master-0 is done
...

Delete a training job:

Usage

bitdeer-ai training delete [flags] <job-id>

Examples

bitdeer-ai training delete tj-03

Shutdown a training job:

Usage

bitdeer-ai training suspend [flags] <job-id>

Examples

bitdeer-ai training suspend tj-03

Restart a training job:

Usage

bitdeer-ai training resume [flags] <job-id>

Examples

bitdeer-ai training resume tj-03

License

Bitdeer AI Cloud CLI is licensed under the MIT License.

Support

If you encounter any issues or have questions, please contact our support team at support email.

Bitdeer AI Cloud CLI

Bitdeer AI Cloud CLI

On this page