MLOS Documentation

MLOS is a project to enable autotuning </autoapi/mlos_core/index.html> with mlos_core for systems via automated benchmarking with mlos_bench including managing the storage and visualization of the results via mlos_viz.

See below for additional documentation sections.

Here is some documentation pulled from the markdown files in the MLOS source tree:

Source Tree Documentation

• MLOS
  • Contents
  • Overview
  • Organization
  • Contributing
  • Getting Started
    • conda activation
    • Usage Examples
      • mlos-core
      • mlos-bench
      • mlos-viz
  • Installation
  • See Also
    • Examples
    • Publications
• mlos-core
  • Description
  • Features
• mlos-bench
  • Table of Contents
  • Description
  • Features
  • Quickstart
    • 1. Install and activate the conda environment
    • 2. Make sure that you have Azure CLI tool installed and working
    • 3. Generate access tokens to interact with Azure resources
    • 4. Create a JSON config with DB credentials (Optional)
    • 5. Create a top-level configuration file for your MLOS setup
    • 5. Create another config file for the parameters specific to your experiment
      • Importance of the Experiment ID config
    • 6. Run the benchmark
  • Optimization
    • Resuming interrupted experiments
  • Analyzing Results
• mlos-viz

Here is some documentation pulled from the Python docstrings in the MLOS source tree:

API Reference

• mlos_core
  • mlos_core.optimizers
    • mlos_core.optimizers.bayesian_optimizers
      • mlos_core.optimizers.bayesian_optimizers.bayesian_optimizer
      • mlos_core.optimizers.bayesian_optimizers.smac_optimizer
    • mlos_core.optimizers.flaml_optimizer
    • mlos_core.optimizers.optimizer
    • mlos_core.optimizers.random_optimizer
  • mlos_core.spaces
    • mlos_core.spaces.adapters
      • mlos_core.spaces.adapters.adapter
      • mlos_core.spaces.adapters.identity_adapter
      • mlos_core.spaces.adapters.llamatune
    • mlos_core.spaces.converters
      • mlos_core.spaces.converters.flaml
      • mlos_core.spaces.converters.util
  • mlos_core.util
  • mlos_core.version
• mlos_bench
  • mlos_bench.config
    • mlos_bench.config.schemas
      • mlos_bench.config.schemas.config_schemas
  • mlos_bench.dict_templater
  • mlos_bench.environments
    • mlos_bench.environments.base_environment
    • mlos_bench.environments.composite_env
    • mlos_bench.environments.local
      • mlos_bench.environments.local.local_env
      • mlos_bench.environments.local.local_fileshare_env
    • mlos_bench.environments.mock_env
    • mlos_bench.environments.remote
      • mlos_bench.environments.remote.host_env
      • mlos_bench.environments.remote.network_env
      • mlos_bench.environments.remote.os_env
      • mlos_bench.environments.remote.remote_env
      • mlos_bench.environments.remote.saas_env
      • mlos_bench.environments.remote.vm_env
    • mlos_bench.environments.script_env
    • mlos_bench.environments.status
  • mlos_bench.event_loop_context
  • mlos_bench.launcher
  • mlos_bench.optimizers
    • mlos_bench.optimizers.base_optimizer
    • mlos_bench.optimizers.convert_configspace
    • mlos_bench.optimizers.grid_search_optimizer
    • mlos_bench.optimizers.manual_optimizer
    • mlos_bench.optimizers.mlos_core_optimizer
    • mlos_bench.optimizers.mock_optimizer
    • mlos_bench.optimizers.one_shot_optimizer
    • mlos_bench.optimizers.track_best_optimizer
  • mlos_bench.os_environ
  • mlos_bench.run
  • mlos_bench.schedulers
    • mlos_bench.schedulers.base_scheduler
    • mlos_bench.schedulers.sync_scheduler
  • mlos_bench.services
    • mlos_bench.services.base_fileshare
    • mlos_bench.services.base_service
    • mlos_bench.services.config_persistence
    • mlos_bench.services.local
      • mlos_bench.services.local.local_exec
      • mlos_bench.services.local.temp_dir_context
    • mlos_bench.services.remote
      • mlos_bench.services.remote.azure
      • mlos_bench.services.remote.ssh
    • mlos_bench.services.types
      • mlos_bench.services.types.authenticator_type
      • mlos_bench.services.types.config_loader_type
      • mlos_bench.services.types.fileshare_type
      • mlos_bench.services.types.host_ops_type
      • mlos_bench.services.types.host_provisioner_type
      • mlos_bench.services.types.local_exec_type
      • mlos_bench.services.types.network_provisioner_type
      • mlos_bench.services.types.os_ops_type
      • mlos_bench.services.types.remote_config_type
      • mlos_bench.services.types.remote_exec_type
      • mlos_bench.services.types.vm_provisioner_type
  • mlos_bench.storage
    • mlos_bench.storage.base_experiment_data
    • mlos_bench.storage.base_storage
    • mlos_bench.storage.base_trial_data
    • mlos_bench.storage.base_tunable_config_data
    • mlos_bench.storage.base_tunable_config_trial_group_data
    • mlos_bench.storage.sql
      • mlos_bench.storage.sql.common
      • mlos_bench.storage.sql.experiment
      • mlos_bench.storage.sql.experiment_data
      • mlos_bench.storage.sql.schema
      • mlos_bench.storage.sql.storage
      • mlos_bench.storage.sql.trial
      • mlos_bench.storage.sql.trial_data
      • mlos_bench.storage.sql.tunable_config_data
      • mlos_bench.storage.sql.tunable_config_trial_group_data
    • mlos_bench.storage.storage_factory
    • mlos_bench.storage.util
  • mlos_bench.tunables
    • mlos_bench.tunables.covariant_group
    • mlos_bench.tunables.tunable
    • mlos_bench.tunables.tunable_groups
  • mlos_bench.util
  • mlos_bench.version
• mlos_viz
  • mlos_viz.base
  • mlos_viz.dabl
  • mlos_viz.util
  • mlos_viz.version

mlos_bench CLI usage

• mlos_bench CLI usage

References

• Github Source Tree