clearml-agent

ClearML Agent - MLOps/LLMOps made easy
MLOps/LLMOps scheduler & orchestration solution supporting Linux, macOS and Windows

🌟 ClearML is open-source - Leave a star to support the project! 🌟

---

ClearML-Agent

Formerly known as Trains Agent

• Run jobs (experiments) on any local or cloud based resource
• Implement optimized resource utilization policies
• Deploy execution environments with either virtualenv or fully docker containerized with zero effort
• Launch-and-Forget service containers
• Cloud autoscaling
• Customizable cleanup
• Advanced pipeline building and execution

It is a zero configuration fire-and-forget execution agent, providing a full ML/DL cluster solution.

Full Automation in 5 steps

1. ClearML Server self-hosted or free tier hosting
2. pip install clearml-agent (install the ClearML Agent on any GPU machine: on-premises / cloud / ...)
3. Create a job or add ClearML to your code with just 2 lines of code
4. Change the parameters in the UI & schedule for execution (or automate with an AutoML pipeline)
5. :chart_with_downwards_trend: :chart_with_upwards_trend: :eyes: :beer:

"All the Deep/Machine-Learning DevOps your research needs, and then some... Because ain't nobody got time for that"

Try ClearML now Self Hosted or Free tier Hosting

Simple, Flexible Experiment Orchestration

The ClearML Agent was built to address the DL/ML R&D DevOps needs:

• Easily add & remove machines from the cluster
• Reuse machines without the need for any dedicated containers or images
• Combine GPU resources across any cloud and on-prem
• No need for yaml / json / template configuration of any kind
• User friendly UI
• Manageable resource allocation that can be used by researchers and engineers
• Flexible and controllable scheduler with priority support
• Automatic instance spinning in the cloud

Using the ClearML Agent, you can now set up a dynamic cluster with *epsilon DevOps

*epsilon - Because we are :triangular_ruler: and nothing is really zero work

Kubernetes Integration (Optional)

We think Kubernetes is awesome, but it is not a must to get started with remote execution agents and cluster management. We designed clearml-agent so you can run both bare-metal and on top of Kubernetes, in any combination that fits your environment.

You can find the Dockerfiles in the docker folder and the helm Chart in https://github.com/allegroai/clearml-helm-charts

Benefits of integrating existing Kubernetes cluster with ClearML

• ClearML-Agent adds the missing scheduling capabilities to your Kubernetes cluster
• Users do not need to have direct Kubernetes access!
• Easy learning curve with UI and CLI requiring no DevOps knowledge from end users
• Unlike other solutions, ClearML-Agents work in tandem with other customers of your Kubernetes cluster
• Allows for more flexible automation from code, building pipelines and visibility
• A programmatic interface for easy CI/CD workflows, enabling GitOps to trigger jobs inside your cluster
• Seamless integration with the ClearML ML/DL/GenAI experiment manager
• Web UI for customization, scheduling & prioritization of jobs
• Enterprise Features: RBAC, vault, multi-tenancy, scheduler, quota management, fractional GPU support

Run the agent in Kubernetes Glue mode an map ClearML jobs directly to K8s jobs:

• Use the ClearML Agent Helm Chart to spin an agent pod acting as a controller
  • Or run the clearml-k8s glue on a Kubernetes cpu node
• The clearml-k8s glue pulls jobs from the ClearML job execution queue and prepares a Kubernetes job (based on provided yaml template)
• Inside each pod the clearml-agent will install the job (experiment) environment and spin and monitor the experiment's process, fully visible in the clearml UI
• Benefits: Kubernetes full view of all running jobs in the system
• Enterprise Features
  • Full scheduler features added on Top of Kubernetes, with quota/over-quota management, priorities and order.
  • Fractional GPU support, allowing multiple isolated containers sharing the same GPU with memory/compute limit per container

SLURM (Optional)

Yes! Slurm integration is available, check the documentation for further details

Using the ClearML Agent

Full scale HPC with a click of a button

The ClearML Agent is a job scheduler that listens on job queue(s), pulls jobs, sets the job environments, executes the job and monitors its progress.

Any 'Draft' experiment can be scheduled for execution by a ClearML agent.

A previously run experiment can be put into 'Draft' state by either of two methods:

• Using the 'Reset' action from the experiment right-click context menu in the ClearML UI - This will clear any results and artifacts the previous run had created.
• Using the 'Clone' action from the experiment right-click context menu in the ClearML UI - This will create a new 'Draft' experiment with the same configuration as the original experiment.

An experiment is scheduled for execution using the 'Enqueue' action from the experiment right-click context menu in the ClearML UI and selecting the execution queue.

See creating an experiment and enqueuing it for execution.

Once an experiment is enqueued, it will be picked up and executed by a ClearML Agent monitoring this queue.

The ClearML UI Workers & Queues page provides ongoing execution information:

• Workers Tab: Monitor you cluster
  • Review available resources
  • Monitor machines statistics (CPU / GPU / Disk / Network)
• Queues Tab:
  • Control the scheduling order of jobs
  • Cancel or abort job execution
  • Move jobs between execution queues

What The ClearML Agent Actually Does

The ClearML Agent executes experiments using the following process:

• Create a new virtual environment (or launch the selected docker image)
• Clone the code into the virtual-environment (or inside the docker)
• Install python packages based on the package requirements listed for the experiment
  • Special note for PyTorch: The ClearML Agent will automatically select the torch packages based on the CUDA_VERSION environment variable of the machine
• Execute the code, while monitoring the process
• Log all stdout/stderr in the ClearML UI, including the cloning and installation process, for easy debugging
• Monitor the execution and allow you to manually abort the job using the ClearML UI (or, in the unfortunate case of a code crash, catch the error and signal the experiment has failed)

System Design & Flow

Installing the ClearML Agent

pip install clearml-agent

ClearML Agent Usage Examples

Full Interface and capabilities are available with

clearml-agent --help
clearml-agent daemon --help

Configuring the ClearML Agent

clearml-agent init

Note: The ClearML Agent uses a cache folder to cache pip packages, apt packages and cloned repositories. The default ClearML Agent cache folder is ~/.clearml.

See full details in your configuration file at ~/clearml.conf.

Note: The ClearML Agent extends the ClearML configuration file ~/clearml.conf. They are designed to share the same configuration file, see example here

Running the ClearML Agent

For debug and experimentation, start the ClearML agent in foreground mode, where all the output is printed to screen:

clearml-agent daemon --queue default --foreground

For actual service mode, all the stdout will be stored automatically into a temporary file (no need to pipe). Notice: with --detached flag, the clearml-agent will be running in the background

clearml-agent daemon --detached --queue default

GPU allocation is controlled via the standard OS environment NVIDIA_VISIBLE_DEVICES or --gpus flag (or disabled with --cpu-only).

If no flag is set, and NVIDIA_VISIBLE_DEVICES variable doesn't exist, all GPUs will be allocated for the clearml-agent.
If --cpu-only flag is set, or NVIDIA_VISIBLE_DEVICES="none", no gpu will be allocated for the clearml-agent.

Example: spin two agents, one per GPU on the same machine:

Notice: with --detached flag, the clearml-agent will run in the background

clearml-agent daemon --detached --gpus 0 --queue default
clearml-agent daemon --detached --gpus 1 --queue default

Example: spin two agents, pulling from dedicated dual_gpu queue, two GPUs per agent

clearml-agent daemon --detached --gpus 0,1 --queue dual_gpu
clearml-agent daemon --detached --gpus 2,3 --queue dual_gpu

Starting the ClearML Agent in docker mode

For debug and experimentation, start the ClearML agent in foreground mode, where all the output is printed to screen

clearml-agent daemon --queue default --docker --foreground

For actual service mode, all the stdout will be stored automatically into a file (no need to pipe). Notice: with --detached flag, the clearml-agent will run in the background

clearml-agent daemon --detached --queue default --docker

Example: spin two agents, one per gpu on the same machine, with default nvidia/cuda:11.0.3-cudnn8-runtime-ubuntu20.04 docker:

clearml-agent daemon --detached --gpus 0 --queue default --docker nvidia/cuda:11.0.3-cudnn8-runtime-ubuntu20.04
clearml-agent daemon --detached --gpus 1 --queue default --docker nvidia/cuda:11.0.3-cudnn8-runtime-ubuntu20.04

Example: spin two agents, pulling from dedicated dual_gpu queue, two GPUs per agent, with default nvidia/cuda:11.0.3-cudnn8-runtime-ubuntu20.04 docker:

clearml-agent daemon --detached --gpus 0,1 --queue dual_gpu --docker nvidia/cuda:11.0.3-cudnn8-runtime-ubuntu20.04
clearml-agent daemon --detached --gpus 2,3 --queue dual_gpu --docker nvidia/cuda:11.0.3-cudnn8-runtime-ubuntu20.04

Starting the ClearML Agent - Priority Queues

Priority Queues are also supported, example use case:

High priority queue: important_jobs, low priority queue: default

clearml-agent daemon --queue important_jobs default

The ClearML Agent will first try to pull jobs from the important_jobs queue, and only if it is empty, the agent will try to pull from the default queue.

Adding queues, managing job order within a queue, and moving jobs between queues, is available using the Web UI, see example on our free server

Stopping the ClearML Agent

To stop a ClearML Agent running in the background, run the same command line used to start the agent with --stop appended. For example, to stop the first of the above shown same machine, single gpu agents:

clearml-agent daemon --detached --gpus 0 --queue default --docker nvidia/cuda:11.0.3-cudnn8-runtime-ubuntu20.04 --stop

How do I create an experiment on the ClearML Server?

• Integrate ClearML with your code

• Execute the code on your machine (Manually / PyCharm / Jupyter Notebook)

• As your code is running, ClearML creates an experiment logging all the necessary execution information:

  • Git repository link and commit ID (or an entire jupyter notebook)
  • Git diff (we’re not saying you never commit and push, but still...)
  • Python packages used by your code (including specific versions used)
  • Hyperparameters
  • Input artifacts

  You now have a 'template' of your experiment with everything required for automated execution

• In the ClearML UI, right-click on the experiment and select 'clone'. A copy of your experiment will be created.

• You now have a new draft experiment cloned from your original experiment, feel free to edit it

  • Change the hyperparameters
  • Switch to the latest code base of the repository
  • Update package versions
  • Select a specific docker image to run in (see docker execution mode section)
  • Or simply change nothing to run the same experiment again...

• Schedule the newly created experiment for execution: right-click the experiment and select 'enqueue'

ClearML-Agent Services Mode

ClearML-Agent Services is a special mode of ClearML-Agent that provides the ability to launch long-lasting jobs that previously had to be executed on local / dedicated machines. It allows a single agent to launch multiple dockers (Tasks) for different use cases:

• Auto-scaler service (spinning instances when the need arises and the budget allows)
• Controllers (Implementing pipelines and more sophisticated DevOps logic)
• Optimizer (such as Hyperparameter Optimization or sweeping)
• Application (such as interactive Bokeh apps for increased data transparency)

ClearML-Agent Services mode will spin any task enqueued into the specified queue. Every task launched by ClearML-Agent Services will be registered as a new node in the system, providing tracking and transparency capabilities. Currently, clearml-agent in services-mode supports CPU only configuration. ClearML-Agent services mode can be launched alongside GPU agents.

clearml-agent daemon --services-mode --detached --queue services --create-queue --docker ubuntu:18.04 --cpu-only

Note: It is the user's responsibility to make sure the proper tasks are pushed into the specified queue.

AutoML and Orchestration Pipelines

The ClearML Agent can also be used to implement AutoML orchestration and Experiment Pipelines in conjunction with the ClearML package.

Sample AutoML & Orchestration examples can be found in the ClearML example/automation folder.

AutoML examples:

• Toy Keras training experiment
  • In order to create an experiment-template in the system, this code must be executed once manually
• Random Search over the above Keras experiment-template
  • This example will create multiple copies of the Keras experiment-template, with different hyperparameter combinations

Experiment Pipeline examples:

• First step experiment
  • This example will "process data", and once done, will launch a copy of the 'second step' experiment-template
• Second step experiment
  • In order to create an experiment-template in the system, this code must be executed once manually

License

Apache License, Version 2.0 (see the LICENSE for more information)