How to set random seed for the model in YAML file

[PeterPirog]

How to set random seed in YAML file to achieve repeatability of the model?

I found the code https://ludwig-ai.github.io/ludwig-docs/api/LudwigModel/#create_model:

create_model(
  config,
  random_seed=42
)

but I don't know how to set it in YAML file.

Typical configuration in raw code is:

os.environ['PYTHONHASHSEED']=str(SEED)
os.environ['TF_CUDNN_DETERMINISTIC'] = '1'  # new flag present in tf 2.0+
random.seed(SEED)
np.random.seed(SEED)
tf.set_random_seed(SEED)

[tgaddair]

Hey @PeterPirog, good observation. Currently the random seed isn't configurable in the YAML, it is passed in as a command-line parameter, like:

ludwig train ... --random_seed 42

It can also be passed in to the LudwigModel programmatically:

model = LudwigModel(...)
model.train(..., random_seed=42)

It shouldn't be too difficult for us to add this to the config in the training section. What do you think @w4nderlust ?

[PeterPirog]

@tgaddair Thank You for the answer