Tutorials&Help

GUI

Share

Statistics

v2.0.3534

Light Mode Dark Mode

❌

config.toml, config.yaml, config.json

What are TOML, JSON and YAML files? What are they good for?

TOML, JSON and YAML are file interchange formats that allow you to save data structures on your disk. OmniOpt2 allows you to load parameters via --config_toml path/to/your/configuration.toml, --config_json path/to/your/configuration.json or --config_yaml path/to/your/configuration.yaml.

Basically, any parameter that can be given to the CLI can also be given over one of those two options. You can also merge CLI parameters with config-files, while CLI parameters take precedence.

Only one of these can be used. You cannot, for example, use --config_toml and --config_json together.

Example files

TOML

live_share = true
partition = "alpha"
experiment_name = "__main__tests__"
mem_gb = 4
time = 60
worker_timeout = 5
max_eval = 2
num_parallel_jobs = 20
gpus = 0
run_program = "./.tests/optimization_example --int_param='%(int_param)' --float_param='%(float_param)' --choice_param='%(choice_param)' --int_param_two='%(int_param_two)'"

[[parameter]]
param = ["int_param", "range", "-100", "10", "int"]

[[parameter]]
param = ["float_param", "range", "-100", "10", "float"]

[[parameter]]
param = ["choice_param", "choice", "1,2,4,8,16,hallo"]

[[parameter]]
param = ["int_param_two", "range", "-100", "10", "int"]

num_random_steps = 1
model = "BOTORCH_MODULAR"
auto_exclude_defective_hosts = true
hide_ascii_plots = true

YAML

live_share: true
partition: alpha
experiment_name: __main__tests__
mem_gb: 4
time: 60
worker_timeout: 5
max_eval: 2
num_parallel_jobs: 20
gpus: 0
run_program: >-
  ./.tests/optimization_example --int_param='%(int_param)'
  --float_param='%(float_param)' --choice_param='%(choice_param)'
  --int_param_two='%(int_param_two)'
parameter:
  - param:
      - int_param
      - range
      - '-100'
      - '10'
      - int
  - param:
      - float_param
      - range
      - '-100'
      - '10'
      - float
  - param:
      - choice_param
      - choice
      - '1,2,4,8,16,hallo'
  - param:
      - int_param_two
      - range
      - '-100'
      - '10'
      - int
    num_random_steps: 1
    model: BOTORCH_MODULAR
    auto_exclude_defective_hosts: true
    hide_ascii_plots: true

JSON

{
  "experiment_name": "__main__tests__",
  "gpus": 0,
  "live_share": true,
  "max_eval": 2,
  "mem_gb": 4,
  "num_parallel_jobs": 20,
  "partition": "alpha",
  "run_program": "./.tests/optimization_example --int_param='%(int_param)' --float_param='%(float_param)' --choice_param='%(choice_param)' --int_param_two='%(int_param_two)'",
  "time": 60,
  "worker_timeout": 5,
  "parameter": [
    {
      "param": [
        "int_param",
        "range",
        "-100",
        "10",
        "int"
      ]
    },
    {
      "param": [
        "float_param",
        "range",
        "-100",
        "10",
        "float"
      ]
    },
    {
      "param": [
        "choice_param",
        "choice",
        "1,2,4,8,16,hallo"
      ]
    },
    {
      "auto_exclude_defective_hosts": true,
      "hide_ascii_plots": true,
      "model": "BOTORCH_MODULAR",
      "num_random_steps": 1,
      "param": [
        "int_param_two",
        "range",
        "-100",
        "10",
        "int"
      ]
    }
  ]
}