Role of rng_seed in evaluation of submission

[dvsaisurya]

In documentionation (external rule set), given an opt list consisting of 5 hyperparameters, are they repeated in 5 studies with different rng_seed or with the same seed? Is there random partitioning for submissions with optlists i.e., a total of 25 trials (5 hparams x 5 studies) are randomly partitioned into 5 studies?

[priyakasimbeg]

are they repeated in 5 studies with different rng_seed or with the same seed.

They will be repeated with different rng_seeds.

is there random partitioning for submissions with optlists ..

For opt-list submissions we expect a 5 hyper-parameter opt list and will run 5 studies (each with a different RNG) per point in the opt-list.