Update Suggest API

There is already a function in the Trial class for suggesting a discrete uniformly distributed float (Trial.suggest_discrete_uniform). It would be nice to have the same option that returns integers instead of floats.

My use case is for network representation learning - I want to suggest the embedding dimension in intervals of 50 between 50-350 (e.g. 50, 100, 150, … , 300, 350).\

Currently, I can accomplish this by dividing out my interval then suggesting an integer and multiplying it like:

embeddings_dim = 50 * trial.suggest_int('embeddings_dim', 1, 7)

But this loses the power of tracking the value within optuna. Alternatively I could use the following, but the the data type is tracked incorrectly.

embeddings_dim = int(trial.suggest_discrete_uniform('embeddings_dim', 50, 350, q=7))

[nzw0301]

@sile Thank you for your reply! I understood.

I’ll create a PR to support log argument of suggest_int since @himkt seems too busy.

[sile]

Through the following PRs, the update of the Suggest API has been completed. All of the features will be available by the next major release (v2.0.0) planned for July. Many thanks to all the contributors involved in this issue! Great work!

• Added step to suggest_int: https://github.com/optuna/optuna/pull/910
• Introduced suggest_float: https://github.com/optuna/optuna/pull/1021, https://github.com/optuna/optuna/pull/1292
• Added step to suggest_float: https://github.com/optuna/optuna/pull/1081, https://github.com/optuna/optuna/pull/1205
• Added log to suggest_int: https://github.com/optuna/optuna/pull/1201, https://github.com/optuna/optuna/pull/1275, https://github.com/optuna/optuna/pull/1277, https://github.com/optuna/optuna/pull/1329