Automatic hyperparameter optimizer based on Ax/Botorch
Project description
OmniOpt2
A hyperparameter optimizer for SLURM-based systems. GUI for creating commands is available at https://imageseg.scads.de/omniax/ , where also additional help can be found.
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=ZWNobyAiUkVTVUxUOiAlKHBhcmFtKSI= \
--parameter param range 0 1000 float
This will automatically install all dependencies. Internally, it calls a python-script.
Show results
./omniopt_evaluate
Plot results
./plot --run_dir runs/example/0
Or, with --min and --max:
./plot --run_dir runs/example/0 --min 0 --max 100
Run tests
Runs the main test suite. Runs an optimization, continues it, tries to continue one that doesn't exit, and runs a job with many different faulty jobs that fail in all sorts of ways (to test how OmniOpt2 reacts to it).
./tests/main_tests
Install from repo
pip3 install -e git+https://github.com/NormanTUD/OmniOpt2.git#egg=OmniOpt2
Install from pypi
pip3 install omniopt2
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
File details
Details for the file omniopt2-2024.10.27.post2.tar.gz
.
File metadata
- Download URL: omniopt2-2024.10.27.post2.tar.gz
- Upload date:
- Size: 111.1 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/5.1.1 CPython/3.9.20
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 8d057645af7c41e8ffdd97649c9f307a27875a5d6d56ca1569c378d72984a674 |
|
MD5 | 6c31749812bd672416f9ceb4f0bb8d00 |
|
BLAKE2b-256 | 74dd5a20d8512b87eb6504ac4fe865aef6a6987a9b9d66ee78caa4b421a5dc65 |
File details
Details for the file omniopt2-2024.10.27.post2-py3-none-any.whl
.
File metadata
- Download URL: omniopt2-2024.10.27.post2-py3-none-any.whl
- Upload date:
- Size: 251.1 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/5.1.1 CPython/3.9.20
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | ca512206dad64c5f2f6d7d0bd892bcc53ea69d1ec99b1377204da8e053b5612e |
|
MD5 | ea227b4f77549be2a6d5129d64374901 |
|
BLAKE2b-256 | 05b8d2d9f7a97901a09fea50c931cf546515a38b6dc869fe60c0e3dbbb11ae1a |