Skip to main content

OmniOpt2 - Hyperparameter Optimizer for SLURM-based Systems

OmniOpt2 is a tool designed to assist researchers, engineers, and data scientists with hyperparameter optimization on SLURM-based clusters, even though it works without it as well. It simplifies large-scale optimization tasks with built-in fault tolerance and flexibility. A graphical user interface (GUI) is available for command creation, accessible at OmniOpt2 GUI. For tutorials on configuration, exit codes, and debugging, visit OmniOpt2 Tutorials.

Main program

omniopt --partition=alpha --experiment_name=example --mem_gb=1 --time=60 \
    --worker_timeout=60 --max_eval=500 --num_parallel_jobs=500 --gpus=1 \
    --follow --run_program=$(echo 'echo "RESULT: %(param)"' | base64 -w0) \
    --parameter param range 0 1000 float

This command initiates OmniOpt2 and installs dependencies if not already installed. The parameter --run_program uses a Base64-encoded string to specify commands. It is recommended to use the GUI, though.

Plot Results

Generates visualizations, such as scatter and hex scatter plots. --min and --max adjust the plotted result value range.

Or, with --min and --max:

omniopt_plot --run_dir runs/example/0
omniopt_plot --run_dir runs/example/0 --min 0 --max 100

Use Optuna directly

Every Optuna sampler (TPE, CMA-ES, GP, Random, Grid, NSGA-II, NSGA-III, MOTPE, ...) is reachable as a special case of OmniOpt through the omniopt_optuna entry point:

omniopt_optuna --max-eval=20 --parameter 'x range -10 10 int' \
    --run-program 'echo "RESULT: $x"'
omniopt_optuna --model=OPTUNA_NSGAII --max-eval=30 \
    --result-names RESULT1=min RESULT2=max \
    --parameter 'x range 0 5 float'

Under the hood this is just ./omniopt --model=OPTUNA_TPE ...; the entry point adds sensible defaults for every --optuna_* flag and forwards the rest verbatim. See the Optuna integration tutorial for the full surface (pruners, persistent storage, multi-objective, remote control via omniopt_optuna study ...).

Using live-share

Use --live_share (also enablable via GUI) to automatically share the job. You will get a URL where your job data is hosted publically for 30 days, meaning everyone can access your results, and you can see all kinds of visualizations and export them.

Run Tests (Developer Use Only)

The test suite simulates various scenarios, including handling faulty jobs and ensuring program resilience.

./tests/main

See the automated tests tutorial page for more details.

Install from pypi

This may not use the bleeding-edge version, but if you get the version from here it means, the test suite has completely tested it properly.

pip3 install omniopt2

Install from repo (bleeding edge, may contain untested changes)

pip3 install -e git+https://github.com/NormanTUD/OmniOpt.git#egg=OmniOpt2

Alternatively, it can be executed directly, as OmniOpt2 will install its dependencies automatically if required.

Error Codes

For common issues and exit codes, see the exit codes tutorial-page.

Autocompletions

Autocomplete files for zsh and bash are in .shells. Run

bash .shells/install

to install them.

Contributions

I'd be glad to see your contributions!

Issues

If you experience any problems, please write issues at my Github Issues page.

Old OmniOpt

The old OmniOpt version, based on HyperOpt, is not supported anymore. It is still available, though, at https://github.com/NormanTUD/LegacyOmniOpt.

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

omniopt2-9808.tar.gz (230.1 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

omniopt2-9808-py3-none-any.whl (513.0 kB view details)

Uploaded Python 3

File details

Details for the file omniopt2-9808.tar.gz.

File metadata

  • Download URL: omniopt2-9808.tar.gz
  • Upload date:
  • Size: 230.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.9.25

File hashes

Hashes for omniopt2-9808.tar.gz
Algorithm Hash digest
SHA256 4418175514d4907e7ca082993c50b58f74e08600488a00e641e83b29ccf66faa
MD5 bd140ca4d1a79a886aaa1a53e41f47ed
BLAKE2b-256 ada9bd5ad6f4e01e418154384edd06d6ef262f528d10e50fc64688ebfd3c27d6

See more details on using hashes here.

File details

Details for the file omniopt2-9808-py3-none-any.whl.

File metadata

  • Download URL: omniopt2-9808-py3-none-any.whl
  • Upload date:
  • Size: 513.0 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.9.25

File hashes

Hashes for omniopt2-9808-py3-none-any.whl
Algorithm Hash digest
SHA256 8233fb811ae50947c504a2a0c0e5c6e646f9c5294a330eb163aa468aa8728030
MD5 5fd4f8c83aea473537efaf471a31b442
BLAKE2b-256 ab80c07eb78c904a39839c47a3a62b35c71bb749d476b08ce6cce6066bf9a105

See more details on using hashes here.

Release history Release notifications | RSS feed

This release

9808

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page